https://ejurnal.methodist.ac.id/index.php/methomika/issue/feedMETHOMIKA: Jurnal Manajemen Informatika & Komputerisasi Akuntansi2026-10-31T00:00:00+07:00Jamaluddinjamaluddin@methodist.ac.idOpen Journal Systems<p data-pm-slice="1 1 []"><strong>METHOMIKA: Jurnal Manajemen Informatika & Komputerisasi Akuntansi</strong> is a peer-reviewed, open-access scientific journal published by Universitas Methodist Indonesia. The journal provides an academic platform for researchers, scholars, and practitioners to disseminate original research and innovative contributions in Informatics Management, Information Systems, Computer Science, Information Technology, Artificial Intelligence, and Computerized Accounting. METHOMIKA is committed to advancing scientific knowledge, technological innovation, and practical applications through the publication of high-quality scholarly articles.</p> <p data-pm-slice="1 1 []">According to the official <a href="https://sinta.kemdiktisaintek.go.id/journals/profile/4219" target="_blank" rel="noopener"><strong>SINTA Profile</strong></a>, METHOMIKA holds a <strong>SINTA 4 Accreditation Status </strong>based on <strong>Nationally Accredited Scientific Journal Decree No. 10/C/C3/DT.05.00/2025</strong></p>https://ejurnal.methodist.ac.id/index.php/methomika/article/view/5778Implementasi Filter Foto Cerdas Berbasis CNN pada Sistem Pelaporan Jalan Rusak Berbasis Web2026-06-12T11:25:16+07:00Kadek Okhrisiana Dwipayana Putraokrisiana13@gmail.comNi Gusti Ayu Putu Harry Saptariniayu.harry@pnb.ac.idMade Pasek Agus Ariawanpasekagus@pnb.ac.id<p><em>Road infrastructure damage threatens user safety, but the effectiveness of public reporting is often hindered by irrelevant reports (spam) that overburden manual verification. This study proposes an update to smart public services by integrating a Convolutional Neural Network (CNN) as an automatic filter and a Geographic Information System (GIS) for spatial mapping. Evaluation using the MobileNetV2 architecture on 500 images yielded 95% accuracy, 89% recall, 94% F1-Score, and a perfect 100% precision. This 100% precision highlights the main scientific contribution: the model's ability to reduce false positives to zero, actively blocking all spam reports at the initial gateway in real-time. Validated reports are automatically visualized on the WebGIS map. This integration is proven to digitize and automate initial report filtering, offering significant bureaucratic efficiency for government agencies.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Kadek Okhrisiana Dwipayana Putra, Ni Gusti Ayu Putu Harry Saptarini, Made Pasek Agus Ariawanhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5638Perancangan Sistem Informasi Pajak Pertambahan Nilai Berbasis Bisnis Proses Notasi pada Web Coretax DJP2026-05-25T21:29:06+07:00Ahmad Mustofaahcmadmustofa@gmail.comLely Indah Kurniaahcmadmustofa@gmail.comRizqi Khoirunisaahcmadmustofa@gmail.comBimo Setyawanahcmadmustofa@gmail.com<p><em>Pajak Pertambahan Nilai (PPN) is a type of tax that plays a strategic role in state revenue, but its administrative process is complex and involves various integrated stages. With the implementation of the Core Tax Administration System (Coretax) by the Directorate General of Taxes as part of tax administration reform, an information system is needed that can represent PPN business processes clearly, structured, and in accordance with applicable tax regulations. The purpose of this study is to design a PPN information system based on business process modeling using the Business Process Model and Notation (BPMN) integrated with the Directorate General of Taxes' Coretax website. The method used is applied research with a descriptive qualitative approach. Data collection techniques include literature studies, tax regulation analysis, and document analysis related to the PPN administration process. Business process modeling is conducted using BPMN to illustrate the PPN process flow. The results of this study are expected to produce a design and BPMN that is easily understood by taxpayers.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Ahmad Mustofa, Lely Indah Kurnia, Rizqi Khoirunisa, Bimo Setyawanhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5952Comparative Analysis of K-Means, K-Medoids, and DBSCAN for Spatial Clustering of Positive Cloud-to-Ground Lightning in Tangerang City2026-07-14T08:54:50+07:00Adi Pranata Rantarantaadi4@gmail.comMarzuki Sinambelasinambela.m@gmail.comAgustina Rachmawardaniagustina.rahmawardani@stmkg.ac.idDeni Septiadideni.septiadi@stmkg.ac.id<p><em>Cloud-to-Ground (CG) lightning is one of the most destructive</em> <em>atmospheric phenomena due to its potential to damage </em><em>electrical infrastructure, communication networks,</em> <em>and public facilities.</em> <em>Understanding its spatial distribution</em> <em>is</em> <em>essential </em><em>for</em> <em>disaster</em> <em>mitigation,</em> <em>infrastructure</em> <em>protection,</em> <em>and</em> <em>urban</em> <em>planning.</em> <em>This</em> <em>study</em> <em>compares</em> <em>the</em> <em>performance</em> <em>of</em> <em>the</em> <em>K-</em><em>Means,</em><em> K-Medoids, and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithms in identifying the spatial distribution of Positive Cloud-to-Ground (CG+) lightning strikes in Tangerang City, Indonesia. The analysis utilizes CG+ lightning observations collected by the Tangerang Geophysical Station of the Indonesian Agency for Meteorology, Climatology, and Geophysics (BMKG) during 2025. Prior to clustering, the dataset underwent data cleaning, CG+ event selection, administrative boundary filtering, feature selection, coordinate transformation from WGS84 to UTM, and Min–Max normalization. The clustering performance of each algorithm was assessed using the Silhouette Coefficient and Davies–Bouldin Index (DBI). The evaluation results indicate that both K-Means and K-Medoids achieved the best performance with K = 3. K-Means obtained the highest Silhouette Coefficient (0.445640) and the lowest DBI (0.789620), while K-Medoids produced comparable results with a Silhouette Coefficient of 0.444728 and a DBI of 0.794027. For DBSCAN, the optimal parameter combination (ε = 200 m and MinPts = 6) generated 64 clusters and 170 noise observations, yielding a Silhouette Coefficient of 0.059880 and a DBI of 1.452474. These findings demonstrate that partition-based clustering approaches provide a more reliable representation of the spatial distribution of CG+ lightning in Tangerang City than the density-based DBSCAN approach.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Adi Pranata Ranta, Marzuki Sinambela, Agustina Rachmawardani, Deni Septiadihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5794Evaluasi Kinerja Pipeline YOLOv8 dan EasyOCR pada Deteksi Kode Aset Kampus2026-06-21T14:03:26+07:00I Made Darma Putradarmaputra2017@gmail.comI Nyoman Eddy Indrayanaeddyindrayana@pnb.ac.idNi Nyoman Harini Puspitamangari@pnb.ac.id<p><em>Manual asset tracking on campus is prone to inefficiencies and human error. This study evaluates the performance of a two-stage architecture that combines the You Only Look Once version 8 (YOLOv8) model for text localization and EasyOCR for optical character recognition on physical inventory labels. The scope of the study is limited to printed asset labels (printed text) tested within the Department of Information Technology. Evaluation was conducted using stress testing methods to assess the system’s resilience to variations in lighting and perspective distortion. The results show that YOLOv8 is highly robust at localizing label areas, achieving a mean average precision (mAP@50-95) of 70.04%. During the character extraction stage, EasyOCR produced the lowest character error rate (CER) of 14.16% under ideal shooting conditions. However, extraction performance declined, with error rates reaching 30.82% under glare conditions and 48.58% at extreme tilt angles. This study concludes that the YOLOv8 pipeline is highly suitable for object localization in campus environments, but the OCR module requires more adaptive image preprocessing optimization to handle variations in environmental conditions. This research provides a practical contribution in the form of an adaptive computer vision architecture guideline and an objective benchmark for overcoming image degradation distortions in local-scale inventory management systems.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Made Darma Putra, I Nyoman Eddy Indrayana, Ni Nyoman Harini Puspitahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5981Analisis Harga Rumah Berdasarkan Faktor Fisik dan Lingkungan Menggunakan Metode Regresi Linier2026-08-02T09:11:25+07:00Ade Putra Wisudawan Nainggolanadeputragolan@gmail.comMargaretha Yohannayohanna.na2@gmail.comAlfonsus Situmorangalfonsussitumorang@gmail.com<p><em>This study aims to analyze the influence of physical and environmental factors on house prices using the multiple linear regression method and to implement the analysis results into the Property Analyzer system. The physical factors analyzed include building area, land area, number of bedrooms, number of bathrooms, and number of garages, while environmental factors are represented by accessibility to schools, markets, shopping centers, and toll road entrances. The study utilizes a 2022–2024 house price dataset sourced from Kaggle, comprising 1,010 data points. The research stages include data collection, preprocessing, regression assumption testing, parameter estimation using Ordinary Least Squares (OLS), result analysis, and model implementation into the Property Analyzer system. The results indicate that building area, land area, number of bedrooms, number of bathrooms, and garage capacity significantly influence house prices (p-value < 0.05). Land area exhibits the strongest statistical relationship with house prices, with a t-statistic value of 22.221. The model yielded a coefficient of determination (R²) of 0.6554, indicating that 65.54% of the variation in house prices is explained by the independent variables in the model, while 34.46% is influenced by factors outside the model. The implementation of Property Analyzer successfully applied the multiple linear regression model to facilitate a more objective, rapid, and structured house price estimation process. The findings demonstrate that multiple linear regression serves as a viable approach for analyzing factors related to house prices and supporting property analysis processes.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Ade Putra Wisudawan Nainggolan, Margaretha Yohanna, Alfonsus Situmoranghttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5796Pengembangan Aplikasi Android untuk Klasifikasi Penyakit Daun Kopi Menggunakan MobileNetV3Small2026-06-15T12:07:19+07:00I Made Gede Weda Pradnyanadekda2003@gmail.comPutu Manik Prihatinimanikprihatini@pnb.ac.idIda Bagus Adisimakrisna Pelingadisimakrisna@pnb.ac.id<p><em>Manual identification of coffee leaf diseases remains challenging because it is time-consuming and highly dependent on the user's ability to recognize disease symptoms. In addition, most previous studies on plant disease classification have focused on model evaluation in desktop environments, while the implementation of on-device inference on mobile devices remains limited. This study aims to develop an Android-based coffee leaf disease detection application using a Convolutional Neural Network (CNN) with the MobileNetV3Small architecture. The dataset consists of 840 coffee leaf images categorized into three classes, namely healthy leaves, leaf spot, and leaf rust. The research stages include image preprocessing, data augmentation, model training, model evaluation, and TensorFlow Lite implementation on an Android application. The results show that the proposed model achieved an accuracy of 95% on the test dataset with satisfactory classification performance across all classes. The model was successfully deployed on Android devices using TensorFlow Lite, resulting in a model size of 1.1 MB and an average inference time of 0.37 seconds without requiring an internet connection. Furthermore, black-box testing confirmed that all application functions operated properly. The contribution of this study lies in the development of a MobileNetV3Small-based coffee leaf disease classification application implemented using TensorFlow Lite on Android devices. In addition to employing preprocessing and data augmentation to enhance the model's generalization capability, this study demonstrates that the on-device inference approach enables offline disease detection with low computational requirements, making it suitable for resource-constrained mobile devices.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Made Gede Weda Pradnyana, Putu Manik Prihatini, Ida Bagus Adisimakrisna Pelinghttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5783Deteksi dan Evaluasi Gerakan Angkat Beban Menggunakan Computer Vision dengan MediaPipe Pose Estimation2026-06-13T18:57:26+07:00I Putu Agus Sugiantaragussugi00@gmail.comAgus Adi Putrawanputrawanagusadi@pnb.ac.idI Putu Oka Wisnawaokawisnawa@pnb.ac.id<p><em>Skeletal muscle toning exercises are mandatory for all individuals. One of the effective exercises is compound exercises; these exercises involve more than one joint and more than one muscle. One of the movements is a squat. A squat is very crucial for bone and muscle health for the prevention of sarcopenia because the movement is efficient to train many muscles at once but has a high risk of injury if done with improper posture. The movement evaluation process generally requires direct supervision from a professional trainer, which is costly and time-consuming. This study aims to develop a real-time squat movement evaluation system using computer vision technology through the MediaPipe framework. In contrast to artificial intelligence classification-based detection approaches, this system uses a rule-based method that relies on geometric calculations of angles between joints to validate movements without requiring a training process and external datasets that will make the system lighter. The system works by extracting 33 body landmark points from video captures, then processing the coordinates of the shoulders, hips, knees, and ankles using trigonometric formulas to calculate the angles. The evaluation focused on monitoring hip depth and back posture alignment. Test results indicated that the applied rule-based logic was able to detect postural errors and automatically calculate repetitions with a very small average difference in angular accuracy compared to manual measurements.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Putu Agus Sugiantara, Agus Adi Putrawan, I Putu Oka Wisnawahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5884Perancangan Sistem Informasi Manajemen Layanan Riset pada Bappedalitbang Deli Serdang Menggunakan Design Thinking2026-07-13T23:28:18+07:00Imelda Fitriani Pandianganimeldafitrianipandiangan01@gmail.comFince Tinus Waruwuimeldafitrianipandiangan01@gmail.comAnda Yannyimeldafitrianipandiangan01@gmail.com<p><em>Research services at the Regional Development Planning, Research, and Development Agency (Bappedalitbang) of Deli Serdang Regency still face several problems in their management process, including simple administration, unintegrated research data management, and suboptimal information delivery to users. These conditions cause the service process to be less effective and make it difficult for users to obtain information regarding the progress of research submissions. This study aims to design a prototype of a mobile-based Research Service Management Information System by applying the Design Thinking method as a user-oriented approach. The design process is carried out through the stages of empathize, define, ideate, prototype, and test. Data collection was carried out through observation, interviews, and questionnaires, while the system interface design was developed using Figma. Furthermore, the level of usability of the prototype was evaluated using the System Usability Scale (SUS) method with 30 respondents. The evaluation results showed a SUS score of 79.25, which is included in the Acceptable category, obtained a Grade B, and is in the Excellent category. The findings indicate that the resulting prototype has a good level of usability and has the potential to support the management of research services in a more effective, efficient, and structured manner, thereby increasing the ease of users in accessing service information at the Bappedalitbang of Deli Serdang Regency.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Imelda Fitriani Pandiangan, Fince Tinus Waruwu, Anda Yannyhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5803Penerapan Algoritma K-Means Clustering untuk Mengelompokkan Pola Bermain Pemain TheoTown2026-06-24T08:32:31+07:00Putu Christian Juan Yakhin Ofel Moneputuchristianjuan@gmail.comNi Gusti Ayu Putu Harry Saptariniayu.harry@pnb.ac.id Gde Brahupadhya Subiksabrahupadhya@pnb.ac.id<p><em>City-building simulation games like TheoTown trigger diverse player behaviors, yet the segmentation between strategic and satirical players remains unmapped. This study classifies TheoTown player patterns using the Knowledge Discovery in Databases (KDD) framework and k-means clustering on 1,084 respondents. Data were preprocessed with one-hot encoding and min-max scaling to eliminate Euclidean distance bias. Multi-criteria validation, specifically the Silhouette Coefficient, indicated the most rational separation at K=2 (score 0.1758). The novelty of this study lies in the mixed-method approach integrating K-Means with multi-criteria validation and qualitative content analysis in a sandbox ecosystem. Quantitative results were validated qualitatively via word clouds from respondents' essays. Findings reveal two main archetypes: Cluster 0 (Strategic & Welfare Builders, 57.29%), focusing on economic and spatial efficiency, and Cluster 1 (Experimental & Satirical Players, 42.71%), utilizing the game for radical experiments and government critique. This study indicates that simulation games can serve as a medium for players' psychographic expression.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Putu Christian Juan Yakhin Ofel Mone, Ni Gusti Ayu Putu Harry Saptarini, Gde Brahupadhya Subiksahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5797Penerapan Data Mining pada Sistem Rekomendasi Destinasi Wisata Berbasis Term Frequency-Inverse Document Frequency dan Vector Space Model2026-06-15T16:10:42+07:00I Kadek Nova Hary Saputranovabiasaaja27@gmail.comAgus Adi Putrawanputrawanagusadi@pnb.ac.idGde Brahupadhya Subiksabrahupadhya@pnb.ac.id<p><em>Digital tourism transitions cause information overload, complicating tourist decisions with unstructured data. Collaborative filtering systems also encounter cold-start issues for new destinations that lack sufficient popularity metrics. To address this, this research develops a recommendation system for Bali tourist destinations using content-based filtering. The method integrates Term Frequency-Inverse Document Frequency (TF-IDF) to extract text weights and the Vector Space Model (VSM) via cosine similarity to measure user query and destination relevance. Built on a Model-View-Controller (MVC) architecture, the system dynamically links PHP Laravel and Python Flask via a REST API. Text preprocessing involves case folding, tokenizing, filtering, and stemming using the Sastrawi library. Evaluation results without Top-N cutoff restrictions show an average recall rate of 90.20%, an average accuracy of 88.18%, and a mean average precision (MAP) of 72.82%. Although the absence of cutoff limits suppressed average precision to 34.43%, setting a 0.05 relevance threshold efficiently filtered out irrelevant noise. The system comprehensively recommends potential semantic destinations without relying on popularity, while maintaining high-ranking quality.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Kadek Nova Hary Saputra, Agus Adi Putrawan, Gde Brahupadhya Subiksahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5800Dashboard Analisis Sentimen Ulasan Google Maps Perguruan Tinggi di Bali Berbasis IndoBERT2026-06-26T08:37:28+07:00I Made Pradnyananda Kevin Arnawakevin1arnawa@gmail.comPutu Manik Prihatinimanikprihatini@pnb.ac.idYessi Aprilia Waluyokevin1arnawa@gmail.com<p><em>Public opinions shared on the Google Maps platform factually reflect the reputation and service quality of higher education institutions in Bali Province. However, manually analyzing thousands of reviews is time-consuming and potentially subjective. Therefore, this research focuses on developing a web-based system equipped with an interactive visual dashboard to computerize the categorization and comparison of review sentiments. The text classification process is built using the IndoBERT (Indonesian Bidirectional Encoder Representations from Transformers) deep learning architecture, which sorts opinions into positive or negative sentiments before presenting them on the web interface. Algorithm performance evaluation shows that IndoBERT is reliable, achieving an accuracy rate of 82.18% and a sensitivity (recall) of 91.90%. The modeling output is visualized dynamically through a dashboard, enabling institutions and prospective students to monitor campus sentiments more efficiently, objectively, and precisely.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Made Pradnyananda Kevin Arnawa, Putu Manik Prihatini, Yessi Aprilia Waluyohttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5804Sistem Pendeteksi Lampu Lalu Lintas untuk Kendaraan Pintar Menggunakan YOLOv9 Berbasis PyTorch2026-06-16T19:56:31+07:00Achmad DewaAchmaddewa15@gmail.comNi Gusti Putu Ayu Harry Saptariniayu.harry@pnb.ac.idI Putu Astya Prayudhaastya@pnb.ac.id<p><em>The ability to accurately detect traffic lights in real time is a crucial component for the safety and navigation of autonomous vehicles. This study aims to develop and evaluate a traffic light detection system (red, yellow, green) using the YOLOv9 deep learning algorithm based on the PyTorch framework. This study employs a quantitative approach with an experimental method utilizing the S2TLD dataset as well as additional data covering various lighting and weather conditions. The model was evaluated using Precision, Recall, Mean Average Precision (mAP), and inference time metrics. Test results show that the YOLOv9 model is highly effective, achieving a 98.0% mAP@0.5 in standalone testing and 96.5% when tested directly on a dashboard system using a local dataset. Additionally, the system recorded 0% fatal cross-color classification errors; the system recorded 0% cross-color classification errors on the test dataset, demonstrating a high level of reliability to support smart vehicle systems. In terms of processing speed, the system achieves an average inference time of 41.1 ms per frame (±24 FPS) using GPU acceleration. In conclusion, the YOLOv9 architecture has proven to be accurate, safe, and robust for implementation as an autonomous navigation perception system.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Achmad Dewa, Ni Gusti Putu Ayu Harry Saptarini, I Putu Astya Prayudhahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5806Implementasi Convolutional Neural Network (CNN) untuk Deteksi Visual Situs Judi Online2026-06-17T11:21:01+07:00Gusti Ngurah Senja Ghautamasenjaghautama2004@gmail.comAgus Adi Putrawansenjaghautama2004@gmail.comMade Pasek Agus Ariawansenjaghautama2004@gmail.com<p><em>Online gambling poses a serious cyber threat in Indonesia, causing financial, social, and psychological impacts on society. According to PPATK's 2025 annual report, the circulation of online gambling funds reached IDR 286.84 trillion throughout 2025, with 12.3 million depositors recorded, despite declining from the previous year due to government intervention. The rapid emergence of new domains renders blacklist-based blocking ineffective. URL and text-content detection methods are also vulnerable to obfuscation, whereas visual characteristics of web pages remain more consistent and harder to manipulate. This study implements a Convolutional Neural Network (CNN) for detecting online gambling sites based on visual analysis of web pages. The novelty lies in a custom CNN architecture with four convolutional blocks for binary visual classification of gambling pages, without relying on URL patterns or text content. The dataset consists of 2,188 web page screenshots collected via Selenium-based crawling, split 70:15:15 for training, validation, and testing. Preprocessing includes resizing images to 216x384 pixels and data augmentation. The CNN architecture employs four convolutional blocks (Conv2D, Batch Normalization, ReLU, and Max Pooling), followed by GlobalAveragePooling2D, Dense layers, and Dropout. Testing achieved 97.27% accuracy with a precision, recall, and F1-score of 0.97, demonstrating the strong potential of this visual CNN approach for automated online gambling site detection.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Gusti Ngurah Senja Ghautama, Agus Adi Putrawan, Made Pasek Agus Ariawanhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6154Implementasi Metode Analitycal Hierarchy Processs (AHP) – Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) untuk Pemberian Kredit di Koperasi Mandiri2026-07-30T22:05:36+07:00Cindi Fatika Sarifacindi079@gmail.comPriska Choirinapriska_choirina@uniramalang.ac.idPangestuti Prima Darajatpriska_choirina@uniramalang.ac.id<p><em>This paper presents the development of a web-based information system to improve the credit eligibility assessment process at Koperasi Mandiri. The existing manual assessment process has several limitations, including low efficiency, a high risk of calculation errors, and subjective decision-making. The system was developed using the Waterfall software development methodology. Furthermore, the Analytical Hierarchy Process was employed to determine the weights of the evaluation criteria, while the Technique for Order Preference by Similarity to Ideal Solution was applied to rank the alternatives and determine credit eligibility based on the preference value of each alternative. The assessment was conducted using four criteria: membership duration, basic salary, number of dependents, and loan amount. The proposed system evaluated 18 loan applicants using these four assessment criteria. The AHP consistency ratio was 0.0218, indicating acceptable consistency, while TOPSIS successfully ranked all alternatives, with A12 obtaining the highest preference score (0.7693). System functionality was evaluated using Black Box Testing, while White Box Testing was performed to verify the program logic. The results demonstrate that the developed web-based AHP-TOPSIS application is capable of performing computerized, objective, and valid credit eligibility assessments. In addition, the system achieved a 100% success rate in logical testing, indicating its effectiveness in improving operational efficiency and enhancing data transparency at Koperasi Mandiri.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Cindi Fatika Sari, Priska Choirina, Pangestuti Prima Darajathttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5900Sistem Ekstraksi Informasi Lirik Lagu Daerah Menggunakan Term Frequency-Inverse Document Frequency (TF-IDF) dan Cosine Similarity2026-07-21T22:40:37+07:00Nur Ifdayantinurifdayanti710@gmail.comFajar Ratnawatifajar@polbeng.ac.id<p><em>This study develops an information extraction system for Indonesian regional folk-song lyrics by combining Term Frequency-Inverse Document Frequency (TF-IDF), Cosine Similarity, and Part-of-Speech (POS) Tagging with rule-based extraction. A preliminary survey of 33 respondents indicated limited familiarity with regional songs and difficulty understanding lyric meanings, motivating a web-based system developed using Rapid Application Development (RAD). The system processes lyric queries through case folding, tokenization, stopword removal, stemming, TF-IDF weighting, and Cosine Similarity ranking to retrieve the most relevant song. The selected lyrics are then analyzed to extract nouns, verbs, adjectives, pronouns, predicates, actions, subjects, objects, and related semantic elements. Black-box testing showed that all main functions operated successfully. Retrieval evaluation on ten test queries achieved 100% Top-1 Accuracy, while linguistic extraction against manual annotation produced Precision 0.74, Recall 0.76, and F1-Score 0.75. These results indicate that integrating statistical weighting and linguistic analysis can effectively support cultural documentation and regional-language learning.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Nur Ifdayanti, Fajar Ratnawatihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5950Perancangan Aplikasi Website Booking Barbershop pada Reza Barbershop Menggunakan Metode User Centered Design2026-07-13T23:26:43+07:00I Made Sandi Putrawansandiputrawan@gmail.comAgus Adi Putrawanputrawanagusadi@pnb.ac.idLuh Gede Putri Suardaniputrisuardani@pnb.ac.id<p><em>Reza Barbershop is a men's hair care service located in Padang Sambian, Denpasar, Bali, which still relies on a manual booking process. Customers are required to visit the barbershop directly or contact staff through personal phone numbers to make reservations, resulting in inefficient service, longer waiting times, and potential customer dissatisfaction. This study aims to develop a web-based booking system using the User Centered Design (UCD) approach to improve the booking process and enhance service quality. The development process consisted of identifying user requirements, designing the user interface, implementing the system using the Laravel framework and MySQL database, and evaluating the developed application. Functional testing was conducted using Black Box Testing, while usability evaluation was performed using the System Usability Scale (SUS) involving 30 respondents. The results indicate that all system functionalities operated successfully according to user requirements. Furthermore, the usability evaluation achieved an average SUS score of 82.67, corresponding to Grade A, an Excellent adjective rating, and an Acceptable level of usability. These findings demonstrate that the proposed web-based booking system provides a high level of usability and effectively supports customers in making online reservations while improving operational efficiency and service management at Reza Barbershop.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Made Sandi Putrawan, Agus Adi Putrawan, Luh Gede Putri Suardanihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5832Pengembangan Model Prediksi Harga Kos di Wilayah Jimbaran Menggunakan Metode Decision Tree Regression2026-06-21T13:55:19+07:00I Putu Krishna Paradiptakrishnaparadipta@gmail.comPutu Manik Prihatinimanikprihatini@pnb.ac.idNi Nyoman Harini Puspitamangari@pnb.ac.id<p><em>The demand for temporary housing in Jimbaran continues to increase, making it more complex to determine boarding house prices due to various physical characteristics, location, and facilities. Previous studies have limitations in modeling nonlinear relationships among variables and have not integrated feature engineering and hyperparameter tuning to improve prediction performance. This study aims to develop a model for predicting boarding house prices using the decision tree regression algorithm. The model was developed through data preprocessing, feature engineering, and hyperparameter tuning using GridSearchCV with stratified K-fold cross-validation. The optimal hyperparameter combination was max_depth = 8, min_samples_split = 5, min_samples_leaf = 2, and max_leaf_nodes = 80. The model achieved R² values of 0.8871 for the training data and 0.8326 for the testing data, with an MAE of IDR 225,438 and an RMSE of IDR 303,080. The R² difference of 0.0546 indicates satisfactory generalization without significant overfitting. In addition to producing an accurate prediction model, this study contributes through the implementation of data preprocessing, feature engineering, hyperparameter tuning, and feature importance analysis to identify the factors that most significantly affect boarding house prices.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Putu Krishna Paradipta, Putu Manik Prihatini, Ni Nyoman Harini Puspitahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5826Pemanfaatan Ekstensi PRF WebAuthn untuk Pembungkusan Kunci Privat Sisi Klien pada Penandatanganan Transaksi Offline Hyperledger Fabric2026-06-28T17:32:36+07:00I Made Prajna Paramahansa Satriagamaprajnaparamahansa@gmail.comI Nyoman Gede Arya Astawaarya_kmg@pnb.ac.idAgus Adi Putrawanputrawanagusadi@pnb.ac.id<p><em>Private key management remains a practical barrier to adopting consortium blockchains such as Hyperledger Fabric, where transactions must be signed with the ECDSA private key associated with an X.509 certificate. Server custody simplifies implementation but creates a single point of compromise, whereas delegating key files to users reduces usability and raises key-loss risk. This study proposes and evaluates a non-custodial prototype that uses the WebAuthn Pseudo-Random Function (PRF) extension to derive a client-side key-wrapping key. Unlike existing Fabric key management (file-based wallets, server custody, or HSM/PKCS#11) and prior WebAuthn PRF uses that mainly encrypt user data, this work combines PRF-derived key wrapping, envelope encryption, and Fabric offline signing to protect rather than replace the ECDSA signing key. The ECDSA P-256 key is generated in the browser and protected by envelope encryption: it is encrypted with a random Data Encryption Key (DEK, AES-256-GCM), while the DEK is wrapped with a Key Encryption Key derived from PRF output via HKDF-SHA256. The server stores only ciphertexts and the user certificate. During submission, the key is briefly decrypted in browser memory for client-side offline signing before being sent to the Fabric network. Evaluation with the asset-transfer-basic chaincode shows that browser-signed transactions are accepted and validated without exposing the plaintext private key to the server. Average cryptographic overhead is 0.248 ms (SD 0.022 ms). It remains feasible for interactive use, although PRF support remains uneven across authenticators. This study is a proof of concept for technical feasibility, not a comprehensive security analysis.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Made Prajna Paramahansa Satriagama, I Nyoman Gede Arya Astawa, Agus Adi Putrawanhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5802Implementasi Sistem Peramalan Harga Saham Indeks LQ45 Menggunakan Algoritma Prophet dan Fitur Moving Average2026-06-15T12:10:26+07:00I Made Adi Anugrah Anugrahimadeadianugrah@gmail.comNi Gusti Ayu Putu Harry Saptariniayu.harry@pnb.ac.idIda Bagus Adisimakrisna Pelingadisimakrisna@pnb.ac.id<p><em>Volatile stock price movements in the LQ45 Index often make it difficult for investors to make investment decisions. Previous research on stock price forecasting has generally focused on static analysis of a single issuer in isolation, making it less efficient for broad market monitoring. This study develops a web-based multi-issuer forecasting system using the Streamlit framework to centrally project the price movements of the 45 constituents of the LQ45 Index. This forecasting model integrates technical indicators—moving averages (MA-5, MA-10, and MA-20)—as external regressors in the Facebook Prophet algorithm to capture momentum signals not detected by the base model. The dataset used includes daily trading history covering a maximum period of the last 5 years, adjusted according to the availability of historical data for each issuer. A strict chronological split and a rolling cross-validation mechanism were used to keep the testing honest and stop data from leaking during model validation. Test results show an average MAPE of 7.54% for a 7-day forecast horizon, 7.47% for 14 days, and 16.63% for 30 days. Overall, the system recorded an average MAPE of 10.57%, which was converted into a forecasting performance percentage of approximately 89.43% to make it easier for investors to interpret as an investment decision-making tool.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Made Adi Anugrah Anugrah, Ni Gusti Ayu Putu Harry Saptarini, Ida Bagus Adisimakrisna Pelinghttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5840Model Prediksi Risiko Kredit Macet Menggunakan XGBoost Dengan AI Explainer Berbasis SHAP2026-06-27T11:25:32+07:00Alvian Maulana Mansyurplayinghd10@gmail.comPutu Manik Prihatinimanikprihatini@pnb.ac.idGde Brahupadhya Subiksabrahupadhya@pnb.ac.id<p><em>Credit is a financial instrument that carries the risk of default when borrowers are unable to fulfill their repayment obligations. This risk may lead to financial losses for lending institutions, making accurate credit risk prediction essential. This study aims to develop a credit risk prediction model using Extreme Gradient Boosting (XGBoost) and point out its predictions using an AI explainer based on SHapley Additive exPlanations (SHAP). The dataset was obtained from the internal database of a fintech company in Indonesia and consisted of 116,250 records with 19 attributes. The research stages included Exploratory Data Analysis (EDA), missing value and outlier handling, data splitting, model training, evaluation, and result interpretation. The XGBoost model achieved an accuracy of 92%. For the Less Current class, the model obtained a precision of 0.91, a recall of 0.97, and an F1-score of 0.94. For the current class, the model achieved a precision of 0.86, a recall of 0.62, and an F1-score of 0.72. For the Default class, the model obtained a precision of 0.97, a recall of 0.91, and an F1-score of 0.94. SHAP analysis as an AI explainer revealed that the income and side_income features were the most influential factors in determining credit risk.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Alvian Maulana Mansyur, Putu Manik Prihatini, Gde Brahupadhya Subiksahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5810Sistem Pakar Penilaian Resiko Joint Closure pada Infrastruktur Fiber Optik Menggunakan Forward Chaining2026-06-16T17:38:40+07:00Komang Wahyudita Eka Pramanakomangwahyuu87@gmail.comPutu Indah Ciptayanikomangwahyuu87@gmail.comYessi Aprilia Waluyokomangwahyuu87@gmail.com<p><em>The manual management of Fiber optic infrastructure data at PT Aplikanusa Lintasarta has the potential to cause data duplication, recording errors, difficulty in information retrieval, and inconsistencies in Joint Closure risk assessments. These problems can affect the effectiveness of decision-making related to network infrastructure maintenance. This study aims to design and build a Fiber Optic Infrastructure Information System equipped with a Joint Closure Risk Assessment Expert System feature using the Forward Chaining method. This method is used to determine the risk level based on location conditions, installation age, and disruption history represented in the form of expert knowledge-based rules. The Forward Chaining method is implemented using 26 rules that represent expert knowledge in determining the Joint Closure risk level. System evaluation is carried out through Black Box testing, System Usability Scale (SUS), and expert validation. The results of the study indicate that the system runs according to its function with a Black Box test success rate of 100%, obtaining a SUS score of 71.83 which is included in the Good and Acceptable categories, and a level of conformity with expert validation results of 100%. Based on these results, the system built is able to support infrastructure data management and Joint Closure risk assessment in an effective, consistent, and structured manner.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Komang Wahyudita Eka Pramana, Putu Indah Ciptayani, Yessi Aprilia Waluyohttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5787Pengembangan Sistem Informasi Manajemen Kinerja Pegawai Berbasis Dynamic KPI pada PT Sanata System2026-06-21T13:35:31+07:00Dian Aulia Safitridianauliasafitri0@gmail.comI Wayan Candra Winetracandrawinetra@pnb.ac.idMade Pasek Agus Ariawanpasekagus@pnb.ac.id<p><em>The employee performance evaluation process at PT Sanata System currently relies on manual monthly and quarterly calculations by the head of division using Microsoft Excel and Word. This conventional process triggers a heavy administrative burden, a high risk of human error, and a static (hard-coded) evaluation structure that is difficult to adapt to dynamic divisional performance assessment needs. This study aims to develop an employee performance management information system based on dynamic KPIs to automate and improve the flexibility of performance evaluation processes. The research adopts the Design Science Research (DSR) approach, while the system is developed using the Waterfall software development model consisting of analysis, design, implementation, and testing phases. The resulting system is a web-based application developed using the Laravel framework and MySQL database. The main feature developed is a dynamic KPI module based on a self-referencing relationship database architecture, integrated with a master-detail form that facilitates simultaneous input of main indicators (parent) and sub-indicators (child). The system is also equipped with auto-validation functionality to ensure a total KPI weight of 100% and snapshot logic to preserve historical evaluation data integrity. Functional testing was conducted using black-box testing, including boundary value analysis on the KPI weighting feature, which achieved a 100% pass rate. Meanwhile, User Acceptance Testing (UAT) obtained an average score of 4.76/5.00 (95.2%). The results indicate that the proposed system is capable of supporting flexible KPI management while maintaining the consistency of historical performance evaluation.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Dian Aulia Safitri, I Wayan Candra Winetra, Made Pasek Agus Ariawanhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5876Implementasi Metode Hybrid Deep Neural Network dan Naive Bayes dalam Klasifikasi Kanker Payudara2026-07-13T23:39:13+07:00Angelika Gloria Sinagaangelikagloriass@gmail.comMargaretha Yohannayohanna.na2@gmail.comArina Prima Silalahiprimaarinasilalahi@gmail.com<p><em>Breast cancer is one of the leading causes of death among women, especially if it is not detected early. Therefore, accurate and efficient classification methods are needed to aid the diagnosis process based on clinical data. This study aims to classify breast cancer using a hybrid approach that combines Deep Neural Networks (DNN) and Naive Bayes on the Wisconsin Breast Cancer Dataset (WBCD). DNN is used as a feature extraction method to generate a more informative feature representation from the original data, while Naive Bayes is used as the final classification algorithm. The dataset consists of 569 data points with 30 numerical features and two classes: benign and malignant. The research stages include data preprocessing, training the DNN using backpropagation, feature extraction from the hidden layer, and classification using Naive Bayes. The test results show that the model achieved an accuracy of 92.11%, precision of 92.31%, recall of 85.71%, and an F1-score of 88.89%. These results indicate that the proposed hybrid approach is capable of improving the quality of data representation and delivering fairly good classification performance. However, the lower recall value suggests that there are still some positive cases that have not been optimally detected.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Angelika Gloria Sinaga, Margaretha Yohanna, Arina Prima Silalahihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5977Pengembangan Sistem Pendaftaran Magang Berbasis Website Menggunakan Metode Simple Additive Weighting pada PT Foxbyte Global Inovasi2026-07-13T23:42:59+07:00Gede Agus Aditya Dharma Serna Putragedeagusadityadharma@gmail.comAgus Adi Putrawanputrawanagusadi@pnb.ac.idI Putu Bagus Arya Pradnyanabagusarya12@pnb.ac.id<p><em>PT Foxbyte Global Inovasi (FGI) is a technology company in Bali that offers internship Programs for vocational high school and university students. However, the registration and selection process for interns at this company is still conducted manually using physical forms and Microsoft Excel. This conventional system causes various obstacles, such as slow data processing, a high risk of document loss, and subjectivity or bias in candidate evaluation. This study aims to develop a web-based internship registration information system named FoxPath to improve the efficiency, accuracy, and transparency of the intern recruitment process. This system applies the Simple Additive Weighting (</em>SAW<em>) method as a decision support system to objectively rank participants based on academic scores (40%), interview results (30%), and Curriculum Vitae (CV) evaluation (30%). The system development followed the Waterfall Model, integrating the ReactJS library on the Frontend, the Laravel Framework on the Backend, and MySQL as the database management system. System testing was conducted using Black box Testing to ensure that feature functionalities run according to requirement specifications. The results show that the FoxPath system successfully processes registration data, normalizes criteria matrices, and displays ranking results accurately in accordance with manual calculations. This system successfully transforms the internship administration process into a digital platform, accelerates decision-making for HRD, and automatically dispatches graduation notifications via email.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Gede Agus Aditya Dharma Serna Putra, Agus Adi Putrawan, I Putu Bagus Arya Pradnyanahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6116Rancang Bangun Aplikasi Mobile Marketplace Barang Bekas dan Pindahan dengan Fitur Geolokasi2026-08-07T08:38:45+07:00Octania Mey Anti Togatoropoctaniatogatorop21@gmail.comRoni J. Simamoraronimor4@gmail.comResianta Perangin-anginresianta88@gmail.comJamaluddin Jamaluddinjamaluddin@methodist.ac.id<p><em>A used goods and moving marketplace is one of the media that can help people conduct transactions for buying and selling items that are still suitable for use. However, the process of buying and selling used goods is still often carried out manually through social media or direct communication, resulting in several problems such as difficulties in finding products based on location, lack of transaction security, and poor transaction data management. This study aims to develop an Android-based marketplace application for used and moving goods to facilitate buying and selling transactions in a safer, easier, and more organized manner. The system development method used in this research is the Waterfall method, which consists of requirements analysis, system design, implementation, testing, and maintenance stages. The backend system was developed using the Laravel framework, while the mobile frontend application was developed using the Java programming language on the Android platform. The system provides several main features such as product management, geolocation-based product search, chat between sellers and buyers, transactions, refunds, disputes, and a fund release mechanism after the buyer confirms receipt of the goods. The results of the study indicate that the application was successfully developed and all main features functioned according to the functional requirements based on black box testing results.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Octania Mey Anti Togatorop, Roni J. Simamora, Resianta Perangin-angin, Jamaluddin Jamaluddinhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6073Analisis Perbandingan Thresholding Otsu dan Sauvola untuk Binarisasi Dokumen Semi-Terstruktur2026-08-03T00:20:17+07:00Zulfikar Yusya Mubarakzulfikaryusya@gmail.comImam Riadizulfikaryusya@gmail.comAbdul Fadlilzulfikaryusya@gmail.com<p><em>Semi-structured transaction documents acquired by scanners and cameras exhibit different degradation patterns. This study evaluates Otsu and Sauvola as baseline binarization methods on 100 self-collected document images, balanced between scanner and camera acquisitions (50 images each). Each image was converted to grayscale, enhanced with CLAHE, processed using six limited parameter configurations, and compared with manually validated pixel-level binary ground truth. Performance was assessed using MSE, PSNR, SSIM, precision, recall, F1-score, foreground ratio, and DRD, with F1-score as the primary criterion. Otsu_T10 achieved the best scanner-domain performance (F1-score = 0.917; DRD = 0.166). For camera images, Otsu_T11 produced the highest F1-score (0.172), while Sauvola_k03 yielded the lowest DRD (0.136). Visual failure analysis showed that scanner errors were dominated by broken thin strokes or character thickening, whereas camera errors included disconnected characters, missing small digits and table lines, and thick foreground artifacts caused by shadows and dark borders. These results confirm that the tested methods are baseline approaches rather than a complete camera-document processing solution. Perspective-aware cropping or correction, illumination normalization, and shadow reduction are mandatory before thresholding and downstream OCR on camera images. The conclusions are limited to the tested dataset and parameter ranges.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Zulfikar Yusya Mubarak, Imam Riadi, Abdul Fadlilhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5971Rancangan Alat Sortir Telur Ayam Otomatis Berbasis Load Cell2026-07-28T11:52:03+07:00Kadek Linda Spa Aprilianiwork.kadeklinda@gmail.comNi Wayan Wisswaniwisswani@pnb.ac.idMade Pasek Agus Ariawanpasekagus@pnb.ac.id<p><em>Weight-based egg grading is a routine yet essential task for laying hen farms seeking consistent product quality and efficient operations, but the conventional practice of visual estimation remains time consuming and error prone. This study presents the design and implementation of an automated weight-based egg sorting system that combines a load cell sensor with an HX711 amplifier module and a web application for monitoring. The research process covered needs assessment, system design, hardware and software development, and comprehensive testing across 100 egg samples through three schemes, namely sensor accuracy, processing speed, and Black Box Testing. Findings indicate that the device automatically weighs eggs and sorts them into three classes, namely small, medium, and large, while recording and presenting the results through the web application. The load cell achieved 99% accuracy with an average error margin of 1%, and the system processed each egg in roughly 3.28 seconds, markedly outperforming the manual method's 5 to 10 second pace. All features also passed Black Box Testing. The resulting device proves capable of making the egg sorting workflow faster, more organized, and computerized relative to the manual approach previously relied upon.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Kadek Linda Spa Apriliani, Ni Wayan Wisswani, Made Pasek Agus Ariawanhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6091Perancangan Prototype UI/UX Sistem E-Counter Kehadiran Jemaat Berbasis Figma Menggunakan Metode Design Thinking2026-07-21T22:45:12+07:00Frida Enjelina Pandianganfridapandiangan18@gmail.comEferoni Ndrurufridapandiangan18@gmail.comArini Vika Sarifridapandiangan18@gmail.com<p><em>The congregation attendance recording at the HKBP Sentosa Tanjung Morawa Church is still done manually using paper, causing the data recording and recapitulation process to take quite a long time, potentially causing recording errors, and making it difficult to search for attendance data in a certain period. This study aims to design a UI/UX prototype for a congregation attendance e-counter system based on figma using the design thinking method. The design thinking stages applied include empathize, define, ideate, prototype, and test. The results of the study are a prototype of a congregation attendance e-counter system that has login, dashboard, congregation attendance recording, attendance statistics, attendance history, report sending, and logout features. The prototype was then tested using the System Usability Scale (SUS) method on 20 respondents. The test results showed an average SUS score of 80.625 which is included in the Acceptable category, obtained a Grade A, and is in the Excellent category.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Frida Enjelina Pandiangan, Eferoni Ndruru, Arini Vika Sarihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5815Implementasi XGBoost dengan Hyperparameter Tuning Optuna untuk Estimasi Harga Akomodasi di Bali Berbasis Mobile2026-06-20T09:42:58+07:00I Gede Igo Wisnu Wardanaigowisnuwardana@gmail.comI Wayan Candra Winetracandrawinetra@pnb.ac.idIda Bagus Adisimakrisna Pelingadisimakrisna@pnb.ac.idNi Gusti Ayu Putu Harry Saptarini ayu.harry@pnb.ac.id<p><em>The recovering tourism industry in Bali has triggered a surge in accommodation demand, but this also creates the problem of daily rental price inconsistency. Subjective pricing by property managers often leads to price disparities in the market. Previous studies generally used linear regression or Random Forest for property price prediction, but not many have integrated hyperparameter optimization using Optuna and mobile implementation for accommodation price estimation in Bali. Therefore, the novelty and main contribution of this study is the integration of the XGBoost model optimized using Optuna into a mobile application to generate real-time accommodation price estimations. This study aims to implement the XGBoost algorithm to estimate the rental prices of accommodations (hotels, villas, and resorts) in Bali based on location and facility variables. The research data was obtained through web scraping on Online Travel Agent (OTA) platforms, resulting in 4,618 clean data rows. To optimize accuracy and prevent overfitting, a Hyperparameter Tuning technique using the Optuna framework was applied to find the best parameter combination. The model's performance was evaluated using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of Determination (R²) metrics. The results showed the optimized XGBoost model achieved high accuracy with an R² of 0.797, RMSE of 0.3752, and MAE of 0.3007. System testing through a User Acceptance Test (UAT) also recorded an acceptance rate of 89.2% (Very Good). In conclusion, the developed prediction model is highly accurate and applicable as an objective price reference instrument for both property managers and tourists.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Gede Igo Wisnu Wardana, I Wayan Candra Winetra, Ida Bagus Adisimakrisna Peling, Ni Gusti Ayu Putu Harry Saptarini https://ejurnal.methodist.ac.id/index.php/methomika/article/view/5828Perancangan Sistem Informasi Penjualan dengan Metode Single Exponential Smoothing untuk Perencanaan Produksi2026-06-28T15:04:23+07:00Rolland Roy Hakikirollandroyhakiki@gmail.comChristian Cahyaningtyaschristi@shantibhuana.ac.id<p><em>The development of information technology has encouraged business owners to utilize systems that can support data management and decision-making processes more effectively. One of the common problems faced by Micro, Small, and Medium Enterprises (MSMEs) is that sales recording and production planning are still carried out manually, making it difficult to determine production quantities that match market demand. This condition also occurs at Mamis Bakery Bengkayang, where sales data recording has not been managed systematically and production quantities are determined based solely on estimation without a clear calculation method. This study aims to design and develop a web-based sales information system integrated with the Single Exponential Smoothing (SES) method to support production planning at Mamis Bakery Bengkayang. The research method employed is software engineering with a quantitative descriptive approach. The developed system is capable of managing product data, customer data, sales transactions, customer orders, and forecasting processes based on historical sales data. The SES method was selected due to its simplicity and effectiveness in predicting fluctuating sales data. System testing was conducted using the black box testing method to ensure that all features function according to user requirements. The results indicate that the developed system is able to improve the management of sales data, accelerate report generation, and produce sales forecasts that can be used as a basis for determining production quantities in subsequent periods, thereby supporting more effective and efficient decision-making processes.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Rolland Roy Hakiki, Christian Cahyaningtyashttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5889Pengembangan Chatbot Informasi Pariwisata di Bali Menggunakan Metode Hybrid RAG 2026-07-03T10:15:27+07:00I Kadek Dwi Pratamadwipratamaikadek@gmail.comNi Wayan Wisswaniwisswani@pnb.ac.idGde Brahupadhya Subiksabrahupadhya@pnb.ac.id<p><em>The development of artificial intelligence-based chatbots has enabled the automated delivery of information. However, chatbots still face challenges in generating relevant answers to user queries. In addition, Retrieval-Augmented Generation (RAG) systems that rely on a single retrieval method have limitations in identifying the most relevant documents, which may affect the quality of the generated responses. Therefore, this study aims to implement a Hybrid Retrieval-Augmented Generation (Hybrid RAG) approach by combining BM25, Bi-Encoder, and Cross-Encoder models. The dataset used consists of 200 Bali tourism documents collected from web scraping, the Bali Central Bureau of Statistics (BPS), and Kaggle. The system was developed using BM25 as the initial retrieval stage, Bi-Encoder for semantic retrieval, Cross-Encoder for reranking, and Gemini as the answer generation model. The evaluation was conducted using Recall@k, Precision@k, and nDCG@k metrics to measure retrieval performance, as well as response time testing to evaluate system performance. The results show that the Hybrid RAG approach, utilizing BM25, Bi-Encoder, and Cross-Encoder, improves document retrieval quality through keyword-based matching, semantic similarity analysis, and reranking processes. Based on the findings, the Hybrid RAG approach was successfully implemented in a Bali tourism information chatbot and was able to provide relevant context to support the answer generation process.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Kadek Dwi Pratama, Ni Wayan Wisswani, Gde Brahupadhya Subiksahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6473Analisis Fitur Citra Digital pada Membran Selulosa Bakteri–Kitosan dengan Variasi Konsentrasi Centella Asiatica2026-10-12T10:39:38+07:00Claudia Shinta Octa Wibowoclaudia.shinta@untirta.ac.idCzidni Sika Azkiaczidni.sika@untirta.ac.idAyu Puspa Wiraniayu.puspa@untirta.ac.id<p><em>Digital image analysis provides a simple, non-destructive approach for quantifying the visual characteristics of biomaterial membrane surfaces. This study evaluated bacterial cellulose–chitosan membranes impregnated with Centella asiatica extract at concentrations of 0%, 1%, 2%, and 3%, while maintaining a constant chitosan concentration. One full-resolution macroscopic image of each membrane was divided into ten non-overlapping regions of interest (ROIs), yielding a total of 40 ROI observations. Image processing was performed using Python to extract four color and intensity features, namely mean red, mean green, mean blue, and mean grayscale intensity, along with three texture features: contrast, homogeneity, and entropy. The analysis was conducted using a descriptive and exploratory approach, as each concentration was represented by only one membrane. The membrane containing 1% Centella asiatica extract exhibited the lowest mean grayscale intensity and the highest homogeneity, whereas the 2% membrane showed the highest mean grayscale intensity and the lowest homogeneity. The highest contrast value was observed in the 0% membrane, while the 2% and 3% membranes exhibited higher entropy values than the 0% and 1% membranes. These findings indicate that color and texture features provide complementary information for objectively characterizing visual differences and local heterogeneity among membrane formulations.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Claudia Shinta Octa Wibowo, Czidni Sika Azkia, Ayu Puspa Wiranihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5904Aplikasi Donasi dan Kebutuhan Panti Asuhan Al-Fajar Desa Dompas Berbasis Web dengan Metode Scrum2026-07-04T09:10:12+07:00Siti Fatimahfatimmahsiti347@gmail.comFajar Ratnawatifajar@polbeng.ac.id<p><em>Donation management at Al-Fajar Orphanage in Dompas Village is still carried out manually using record books, causing risks of data loss, recording errors, and low transparency in reporting to donors. This study aims to design and develop a web-based donation and needs management application using the Scrum method to improve the effectiveness of data management and service quality for donors. The system was developed through five iterative sprints covering user authentication, admin and donor dashboards, orphanage needs management, donation management, and donation reporting, built with PHP, Laravel, and MySQL. The system was evaluated using Black Box Testing to verify functional correctness and a System Usability Scale (SUS) survey involving 30 respondents to measure usability. Black Box Testing results show that all core modules function according to specification. The SUS evaluation produced an average score of 86.41, categorized as Acceptable with an Excellent adjective rating, indicating that the application interface is easy to understand and efficient to use. These results demonstrate that Scrum is effective for iteratively developing a donation management system that meets user needs.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Siti Fatimah, Fajar Ratnawatihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6309Prediksi Harga Resale Sepatu Adidas Samba OG Menggunakan Metode Random Forest Regression Berdasarkan Variabel E-Commerce Sebagai Referensi Harga Untuk Seller dan Buyer2026-09-03T15:11:00+07:00Milhan Muidmilhanmuid@gmail.comFida Maisa Hanafidamaisa@umkudus.ac.idWidya Cholid Wahyudinwidyacholidwahyudin@umkudus.ac.id<p><em>The growth of e-commerce has expanded fashion product trading, including the Adidas Samba OG resale market, where prices vary across listings. This condition may complicate buyers in assessing price reasonableness and sellers in determining selling prices. This study aims to predict Adidas Samba OG resale prices using Random Forest Regression based on Size, Color, e-commerce Platform, and Units Sold. A total of 999 observations were manually collected from Shopee, Tokopedia, and TikTok Shop. The research stages included preprocessing, an 80:20 dataset split, baseline modeling, hyperparameter tuning using GridSearchCV, evaluation, feature importance analysis, visualization, and new-data prediction. The baseline model achieved an MAE of IDR 29,884.25, RMSE of IDR 111,979.93, and R² of 0.8401. After hyperparameter tuning, the best model obtained an MAE of IDR 32,907.91, RMSE of IDR 104,414.54, and R² of 0.8610. Feature importance analysis identified Units Sold as the most influential variable, with an importance of 58.99%. Testing on five new-data scenarios produced estimated resale prices ranging from IDR 2,138,312 to IDR 2,404,340. The results indicate that the model can serve as supporting information for buyers and sellers in considering Adidas Samba OG resale prices.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Milhan Muid, Fida Maisa Hana, Widya Cholid Wahyudinhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5799Implementasi Program Pengolahan Teks Timestamp pada Gambar Berbasis OCR pada Diskominfos Kota Denpasar2026-06-15T15:47:23+07:00I Kadek Angga Prayogaianggaprayogaa@gmail.comI Wayan Candra Winetraianggaprayogaa@gmail.comLuh Gede Putri Suardani ianggaprayogaa@gmail.com<p><em>The development of information technology encourages government institutions to improve the efficiency of digital field documentation data management. At the Denpasar City Department of Communication, Informatics, and Statistics (Diskominfos), the photo documentation upload process is still accompanied by the manual entry of supporting data, such as location coordinates, date and time, and street names. This study designs and implements a web application using the Flask framework and MySQL database to automate this process. The system enables batch processing of field photographs, automatically extracting timestamp texts using Tesseract Optical Character Recognition (OCR) combined with standard fixed-area cropping preprocessing. Based on evaluations using 50 test images, the system achieves an overall text extraction accuracy of 94.2% with an average processing time of 1.42 seconds per image. Functional evaluations using Black Box testing across 10 distinct operational scenarios demonstrate a 100% success rate, ensuring seamless flow from multi-file ingestion to Excel reporting. This study effectively eliminates manual entry redundancies, reduces human data entry errors, and establishes an integrated pipeline for automated institutional reporting.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Kadek Angga Prayoga, I Wayan Candra Winetra, Luh Gede Putri Suardani https://ejurnal.methodist.ac.id/index.php/methomika/article/view/5809Implementasi Metode AHP pada Sistem Pendukung Keputusan Pemilihan Glamping di Kintamani Berbasis Web Menggunakan Data Online Travel Agent (OTA)2026-06-17T11:27:23+07:00Alit Krisna Dwipayanaalitkrisna48@gmail.comNi Gusti Ayu Putu Harry Saptariniayu.harry@pnb.ac.idI Nyoman Rai Widartha Kesumaraiwidartha@pnb.ac.id<p><em>The post-pandemic shift in tourism trends has triggered a significant increase in the number of glamping accommodations in the Kintamani region. This rapid growth inadvertently induces an information overload phenomenon across Online Travel Agent (OTA) platforms, precipitating customer confusion and decision postponement. Tourists face significant cognitive barriers when objectively comparing dynamic pricing, facility completeness, and scenic quality within inflexible search architectures. This study implements a web-based Decision Support System (DSS) utilizing the Analytical Hierarchy Process (AHP) Rating Model to deliver objective and highly personalized accommodation recommendations. A hybrid data acquisition framework is deployed, amalgamating automated web scraping from Tiket.com with continuous manual verification to eradicate data gaps inherently present in OTAs. This research offers explicit novelty by establishing a highly scalable algorithmic integration that significantly enhances computational efficiency for massive alternatives, effectively mitigating information overload using real-time OTA data. The AHP Rating Model was selected for its dual capacity to mathematically validate user preference consistency and process vast arrays of alternatives simultaneously via absolute measurement, thereby bypassing the computational inefficiencies of pairwise comparisons. Empirical implementation demonstrates that the algorithmic integration successfully converts unstructured OTA data into a standardized mathematical ranking based on subjective user weights. System evaluation confirms that Black Box Testing indicates all system functions operate with a 100% success rate, while the User Acceptance Test (UAT) involving 12 respondents achieved a user acceptance rate of 95.2%. Statistical descriptive analysis further substantiates the consistency of user satisfaction with low variance. The implemented system has been proven to effectively filter complex variables, thus significantly mitigating tourists' cognitive load during the decision-making process.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Alit Krisna Dwipayana, Ni Gusti Ayu Putu Harry Saptarini, I Nyoman Rai Widartha Kesumahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6238Kompleksitas Proses dan Non-Konformitas Sebagai Faktor Keterlambatan Pada Alur Kerja Purchase-To-Pay Lengkap Pendekatan Machine Learning2026-08-07T19:44:48+07:00Lucia Sri Istiyowatilucia.istyowati@perbanas.idPratiwi Rachmadi lucia.istyowati@perbanas.idNani Krisnawaty Tachjarlucia.istyowati@perbanas.id<p><em>Invoice processing is a crucial component of a financial information system, particularly the accounts payable module integrated with an Enterprise Resource Planning (ERP) system. Delays in invoice processing are common, leading to financial penalties. This study employs a data-driven approach by integrating process mining and machine learning using the XGBoost algorithm to analyze the factors contributing to delays in the purchase-to-pay (P2P) workflow. Event log data from the BPI Challenge 2019 public dataset and refined to 3,471 complete cases. Process discovery and performance analysis were performed using the Disco tool, while an XGBoost binary classification model was developed with 16 process-level features extracted from the event logs. The model yielded an AUC-ROC of 0.6267 with a 5-fold cross-validation of 0.6382 and a standard deviation of 0.0099, indicating moderate discriminatory ability. Feature importance analysis identified process complexity (0.191), ideal workflow conformity (0.187), and invoice cancellation frequency (0.122) as the dominant delay factors, while document completeness contributed insignificantly. These findings suggest that in a complete P2P workflow, the cause of delay lies in the characteristics of process execution. The results of this study provide direction for developing accounts payable information systems with a focus on complexity reduction and real-time compliance monitoring.</em></p> <p> </p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Lucia Sri Istiyowati, Pratiwi Rachmadi , Nani Krisnawaty Tachjarhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6187Web-Based Asset Management Information System at PT Astra Daihatsu Ngurah Rai Tuban2026-07-30T23:23:43+07:00Yohannes Putu Alvin Sutrisnasugasmine11@gmail.comNi Wayan WisswaniWisswani@pnb.ac.idPutu Desiana Wulaning Ayuwulaning_ayu@pnb.ac.id<p><em>Effective asset management plays a crucial role in supporting organizational operations by ensuring the availability, accuracy, and accessibility of asset information. However, the asset recording process at PT Astra Daihatsu Ngurah Rai Tuban was previously conducted using a semi-manual approach, resulting in data redundancy, delays in report generation, and difficulties in obtaining real-time inventory information. This study aims to design and develop a web-based Asset Management Information System to improve the effectiveness and efficiency of asset management within the company. The system was developed using the Waterfall software development methodology, while data were collected through observation, interviews, and documentation. The application was implemented using the Laravel framework with a MySQL database and incorporates Role-Based Access Control (RBAC) to manage user permissions for Administrators, Warehouse Staff, Sales Staff, and Managers. The system provides integrated features for master data management, incoming and outgoing goods transactions, goods return processing, inventory reporting, and report export in PDF and Microsoft Excel formats. System evaluation was conducted using Black Box Testing and User Acceptance Testing (UAT). The testing results indicate that all functional requirements were successfully fulfilled, with all test cases achieving a Passed status. Furthermore, UAT results demonstrate that the developed system satisfies user requirements and effectively supports daily asset management activities. Therefore, the proposed system provides a reliable and efficient solution for improving asset management and supporting managerial decision-making at PT Astra Daihatsu Ngurah Rai Tuban.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Yohannes Putu Alvin Sutrisna, Ni Wayan Wisswani, Putu Desiana Wulaning Ayuhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6127Analisis Prediktif Performa Pemain Piala Dunia FIFA Menggunakan XGBoost untuk Mendukung Pengambilan Keputusan Pelatih2026-07-25T11:27:44+07:00Rena Nainggolanfilipiopat@gmail.comFenina A. T. Tobingfilipiopat@gmail.com<p><em>Machine learning allows player performance analysis based on data to improve the credibility of coaches' decision-making. This study aims to create a player performance prediction model for the FIFA World Cup using data containing 54,600 observations and 75 attributes, covering technical, tactical, and physical statistics, as well as indicators of player performance. Data preprocessing, feature selection, splitting into training and testing sets (80:20), and model building were done using Linear Regression, Decision Tree, Random Forest, Gradient Boosting, and XGBoost. The results show that Gradient Boosting delivered the best performance with an MAE of 1.970177, RMSE of 2, and Coefficient of Determination (R2). In the XGBoost model, Feature Importance and SHAP analysis found that playing minutes, tournament rating, passing accuracy, and expected goals (xG) are the main components affecting player performance.</em> <em>This model can generate player rankings, recommendations for the Starting XI players, and substitutes, and it works as a Decision Support System (DSS). The DSS helps coaches make strategies, assess player performance, and decide the team lineup with accurate data.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Rena Nainggolan, Fenina A. T. Tobinghttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6193Implementasi Algoritma SAW dan Topsis untuk Penentuan Prioritas Program Kerja pada Sistem Informasi Manajemen HMI UNIRA Malang2026-08-08T08:21:03+07:00Lailatu Badriyahlbadriyyah1@gmail.comPriska Choirinapriska_choirina@uniramalang.ac.idUrnika Mudhifatul Jannahlbadriyyah1@gmail.com<p><em>Work program prioritization at HMI Wahid Hasyim Commissariat, UNIRA Malang, has been carried out subjectively through daily meetings (RAHAR) without measurable assessment parameters, potentially causing more strategic programs to be delayed due to poorly managed resource limitations. This study aims to design and build a web-based Management Information System for HMI's administration, equipped with a Decision Support System module using the hybrid SAW-TOPSIS method with five assessment criteria: Human Resources, Budget Requirements, Urgency Level, Schedule Suitability, and Technical Complexity. The SAW method is used to normalize and weight the criteria values, while the TOPSIS method determines the distance of each alternative from the positive and negative ideal solutions. The system was developed using the Waterfall method and implemented with JavaScript, Node.js, Express.js, and PostgreSQL. Black Box Testing showed a 100% success rate, and sensitivity analysis with 20 weight change scenarios yielded an average ranking change percentage of 0%, proving the system produces stable and reliable work program priority recommendations.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Lailatu Badriyah, Priska Choirina, Urnika Mudhifatul Jannahhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6260Perancangan dan Implementasi Sistem Manajemen Status Pengaduan Warga Terintegrasi Notifikasi Firebase Cloud Messaging (FCM)2026-08-09T15:23:28+07:00Ulfiatul Rosidahulfiarosidah52@gmail.comAhmad Zulham Fahamsyah Havyulfiarosidah52@gmail.com<p><em>So far, residents of Tempuran Village have submitted complaints manually through the RT head or by visiting the village office, resulting in less efficient reporting and limited transparency. This research aims to create an Android-based Pantau Desa app that simplifies complaint submission and enables real-time status updates via Firebase Cloud Messaging (FCM). The system development follows the Waterfall approach within the Research and Development (R&D) framework, covering needs analysis, design, implementation, testing, and maintenance. Built with Flutter, the app uses Firebase Firestore for data storage and FCM for notifications. Testing involved Black Box Testing, Integration Testing, and User Acceptance Test (UAT). All test cases in Black Box and Integration Testing were successfully completed. The UAT, conducted with 25 participants (20 residents and 5 village officials), yielded average scores of 4.55 (90.95%) from residents and 4.49 (89.87%) from officials, classified as Very Good. These findings indicate that the Pantau Desa app is suitable for use as an Android-based platform for community complaints, enhancing ease of reporting, improving status management effectiveness, and increasing transparency in village public services.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Ulfiatul Rosidah, Ahmad Zulham Fahamsyah Havyhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6306Klasifikasi Tiga Jenis Macan Tutul Berbasis Pengolahan Citra Digital Menggunakan Kombinasi Fitur RGB, GLCM, PCA dan SVM2026-08-15T12:46:56+07:00Joanne Oliviajoanne.535230004@stu.untar.ac.idAdriano Deshonadriano.535230017@stu.untar.ac.idKhenichi Kuolimpokhenichi.535230142@stu.untar.ac.id<p><em>Classification of leopard species based on digital images is one of the applications of image processing and data mining that can support the automatic identification of wildlife. This study aims to develop a classification model for three leopard species, namely African Leopard, Clouded Leopard, and Snow Leopard, based on the color and texture characteristics of images. The research process includes image pre-processing, color feature extraction using the Red, Green, Blue (RGB) channels, and texture feature extraction using the Gray-Level Co-occurrence Matrix (GLCM), which produces contrast, correlation, energy, and homogeneity features. The extracted features are then reduced using Principal Component Analysis (PCA) while retaining 95% of the information variance, before being classified using Support Vector Machine (SVM) with the multi-class Error-Correcting Output Codes (ECOC) approach. Evaluation is conducted using three data-splitting scenarios for training and testing sets. The results show that the scenario with 70% training data and 30% testing data achieves the highest accuracy of 87.06%, with class accuracies of 87.5% for African Leopard, 79.3% for Clouded Leopard, and 94.6% for Snow Leopard. These results indicate that the combination of RGB, GLCM, PCA, and SVM can classify the three leopard species with reasonably good performance.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Joanne Olivia, Adriano Deshon, Khenichi Kuolimpohttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5000Adaptive Intrusion Detection and Automated Cryptographic Response Framework for Real-Time Cyber Threat Mitigation2026-09-09T12:43:33+07:00Jamaluddin Jamaluddinjamaluddin@methodist.ac.idDarwis Robinson Manalumanaludarwis@gmail.comSyahril Efendisyahril1@usu.ac.idJonson Manurungjhonson.geo@gmail.comRoni Jhonson Simamoraronimor4@gmail.com<p><em>The increasing dynamics of cyberattacks require security systems that are not only capable of detecting threats but also responding rapidly, adaptively, and proportionally. This study proposes the Adaptive Intrusion Detection and Automated Cryptographic Response (AID-ACR) Framework, a closed-loop cybersecurity approach integrating intrusion detection, dynamic threat-risk assessment, automated response, cryptographic mechanisms, and adaptive feedback. The prototype was evaluated through a reproducible simulation study using 32,000 synthetic CICIoT-like traffic samples with 39 numerical features and eight classes: Benign, DDoS, DoS, Recon, Web, Brute Force, Spoofing, and Mirai. Random Forest, XGBoost, LightGBM, and a DNN based on a multilayer perceptron were compared as intrusion detection engines. LightGBM achieved the best overall detection performance with 98.77% accuracy, 98.66% macro-precision, 97.21% macro-recall, and 97.91% macro-F1. The adaptive threat-risk assessment achieved 94.44% agreement with the reference simulation policy and reduced the Unnecessary Response Rate from 0.9749% under static response to 0.5571%, a relative reduction of 42.86%. The prototype produced an end-to-end threat mitigation latency of 0.659 ms per event. These findings indicate that AID-ACR can reduce the detection-to-mitigation gap by combining contextual network controls with cryptographic actions such as session re-keying, key rotation, and key revocation. Validation using the official CICIoT2023 files and a real network testbed remains necessary before operational deployment.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Jamaluddin Jamaluddin, Darwis Robinson Manalu, Syahril Efendi, Jonson Manurung, Roni Jhonson Simamorahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6369Rancang Bangun Game Edukasi Berbasis Android Menggunakan Godot Engine untuk Mengenal Huruf Hijaiyah pada Anak Usia Dini2026-09-15T20:57:29+07:00Lailil Maghfirohsecil4055@gmail.comWalidini Syaihul Hudawalidini@yudharta.ac.id<p><em>Hijaiyah letter recognition is an important part of introducing Qur’anic reading to early childhood learners. Observations and interviews at TK Al-Hamidy showed that learning activities still use conventional media, such as word cards and whiteboards, causing some children to lose focus, feel bored, and have difficulty distinguishing hijaiyah letters with similar shapes. This condition encouraged the development of an Android-based educational game using Godot Engine as an interactive learning medium, with Supabase supporting teacher-managed question data. This research adopted a Research and Development (R&D) approach using the ADDIE development model, which consists of five stages: Analyze, Design, Development, Implementation, and Evaluation. Product evaluation included Black Box Testing, expert validation, pretest-posttest comparison, N-Gain analysis, and user questionnaires. The results showed that all game functions operated according to the design. Material and media validation reached 98.89% and 82.2%, respectively, both categorized as highly valid. The mean score increased from 39.2 in the pretest to 82.3 in the posttest, with an N-Gain of 0.70, categorized as high. The user questionnaire obtained a percentage of 72.31%, categorized as interesting. These results indicate that the game can be used as a supporting medium for hijaiyah letter learning in early childhood.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Lailil Maghfiroh, Walidini Syaihul Hudahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6484Sistem Informasi Manajemen Produksi Menggunakan Metode Economic Order Quantity (EOQ) Dan Full Costing Pada Pabrik Bakso Rayhan2026-09-29T12:01:24+07:00Nurhaliza Febryaninurhalizafebryani@gmail.comAninda Muliani Harahapnurhalizafebryani@gmail.com<p><em>At the Rayhan Meatball Factory, workers currently record raw material inventory and calculate production costs manually using logbooks. This practice carries risks such as data loss and clerical errors, and lacks a standardized method for determining raw material order quantities or the Cost of Goods Manufactured (COGM). This study employs the Economic Order Quantity (EOQ) method for inventory management and the Full Costing method to determine production costs. A quantitative research approach was adopted The Waterfall model is a process that has different steps, including looking at what the system needs, designing the system, building it, and checking if it works properly. Integrating the EOQ method into the system facilitates recommendations regarding optimal order quantities and appropriate ordering frequencies. Meanwhile, The Full Costing method is used to figure out the cost of each unit of goods produced by automatically accounting for all production cost components. The developed Laravel and MySQL-based system successfully automates inventory procurement and production cost calculations while generating clear reports, thereby enabling the owner to manage inventory and product manufacturing costs more effectively.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Nurhaliza Febryani, Aninda Muliani Harahaphttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5949Comparative Analysis of Long Short-Term Memory and Random Forest for Soil Temperature Prediction in Greenhouse Microclimate2026-07-28T11:51:05+07:00Muhammad Asyril Ar Rosyadasyrilrosyad@gmail.comMarzuki Sinambelasinambela.m@gmail.comAgustina Rachmawardani agustina.rahmawardani@stmkg.ac.idFadhli Luthfiansyahfadhli.luthfiansyah@stmkg.ac.id<p><em>Soil temperature prediction is an important aspect of greenhouse microclimate management as it affects plant growth and resource efficiency. This study aims to compare the performance of Random Forest (RF) and Long Short-Term Memory (LSTM) algorithms in predicting soil temperature one hour ahead in a greenhouse environment. The dataset was collected from observations conducted from March 23 to June 3, 2026, with a recording interval of 1 minute, including air temperature, air humidity, soil temperature, soil moisture, and light intensity parameters. The data prepossessing stages included data cleaning, missing value interpolation, outlier detection using the Interquartile Range (IQR) method, target variable construction using time shifting, and feature engineering. This study employed four model scenarios: RF with soil moisture, RF without soil moisture, LSTM with soil moisture, and LSTM without soil moisture. Model performance was evaluated using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of Determination (R²). The results showed that all models achieved good performance with R² values above 0.95. The best performance was achieved by the LSTM model without soil moisture, with MAE of 0.3287, RMSE of 0.5881, and R² of 0.9662. These findings indicate that the LSTM model is more effective in capturing the temporal patterns of soil temperature compared to Random Forest. In addition, the soil moisture parameter was found to have a low contribution to prediction accuracy.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Muhammad Asyril Ar Rosyad, Marzuki Sinambela, Agustina Rachmawardani , Fadhli Luthfiansyahhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6421Analisis Sentimen Ulasan Wisatawan terhadap Destinasi Wisata di Sumatera Utara Menggunakan IndoBERT 2026-09-11T20:57:21+07:00Ommi Alfinany.aroen@gmail.comM. Safiim.safii@stikomtunasbangsa.ac.id<p><em>The growth of tourist reviews on travel destinations generates textual information that can potentially be used to understand tourists' perceptions, but sentiment classification faces challenges due to class distribution imbalance.</em> <em>This study aims to analyze the performance of IndoBERT in classifying tourist reviews into positive, neutral, and negative categories. The study uses a fine-tuning approach on IndoBERT with 1,500 tourist reviews and evaluates performance using accuracy, precision, recall, F1-score, and a classification report. The test results showed an accuracy of 72.67%, precision of 52.80%, recall of 72.67%, and an F1-score of 61.16%. However, class-wise analysis showed that the model only did well in identifying the neutral class, with a recall of 100% and an F1-score of 84.17%, while the negative and positive classes both scored 0% for precision, recall, and F1-score. This condition is related to the dominance of the neutral class at 73.07% in the dataset. This finding shows that accuracy alone isn't enough to evaluate a model on imbalanced sentiment data, and strategies to handle class imbalance are needed to improve the model's ability to recognize positive and negative sentiments.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Ommi Alfina, M. Safiihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6115Pengembangan Proof of Concept Aplikasi Matematika Nusantara Adaptif Berbasis AI untuk Siswa SD Kelas V2026-08-09T13:58:53+07:00Oktaviana Bangunoktavianabangun170@gmail.comRafika Sari Br. Sembiringoktavianabangun170@gmail.comHakiki Annisa' Batubaraoktavianabangun170@gmail.com<p><em>Mathematics learning in elementary schools still faces challenges due to differences in students' learning abilities and the limited availability of adaptive learning media that integrate local cultural values. This study aims to develop a proof of concept of an Artificial Intelligence (AI)-based Adaptive Mathematics Nusantara application for fifth-grade elementary school students. The research employed the Research and Development (R&D) method using the ADDIE development model, which includes the stages of analysis, design, development, implementation, and evaluation. The application was designed to provide personalized mathematics learning by adapting learning materials and exercises according to students' abilities while integrating ethnomathematics based on Indonesian local culture. The developed prototype was evaluated through expert validation and limited trials involving teachers and students to assess its functionality and usability. The results indicate that the prototype functions properly and demonstrates the feasibility of AI-based adaptive learning integrated with Mathematics Nusantara. This study contributes to the development of innovative educational technology that supports personalized, contextual, and culturally responsive mathematics learning in elementary schools.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Oktaviana Bangun, Rafika Sari Br. Sembiring, Hakiki Annisa' Batubarahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6447Pengembangan Sistem Informasi Klasifikasi Kesesuaian Kandungan Gizi Menu Terintegrasi dengan Modul Artificial Neural Network2026-09-18T22:12:03+07:00Hen Hen Lukmanahenhenlukmana@unsil.ac.idMuhammad Alhusainialhusaini@unsil.ac.idLuh Desi PuspareniLuhdesipuspareni@unsil.ac.id<p><em>Evaluating the nutritional adequacy of meal menus in nutrition fulfillment programs is generally performed manually, which is time-consuming, requires precision, and is prone to recording errors. This study aims to develop an integrated information system for classifying the nutritional adequacy of meal menus with an Artificial Neural Network (ANN)-based classification module. The system unifies menu data management, automatic nutrient computation, adequacy classification, and analysis history storage within a single platform. The ANN module was trained on a reproducibly generated controlled synthetic dataset of 4,500 samples with three balanced classes (Insufficient, Adequate, and Excessive) and ten nutrient adequacy percentage features. Testing covered model performance using accuracy, precision, recall, F1-score, balanced accuracy, and the confusion matrix; functional black-box testing; and usability testing with the System Usability Scale (SUS). The classification module achieved a test accuracy of 98.81%, a macro F1-score of 98.81%, and a balanced accuracy of 98.81%; all black-box scenarios were valid (100%), and the mean SUS score was 88.50 (grade A+). These results show that the system can automate nutritional adequacy evaluation consistently and in a documented manner. The main contribution lies in integrating menu data management, nutrient computation, and automatic classification within one system, rather than in proposing a new classification method.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Hen Hen Lukmana, Muhammad Alhusaini, Luh Desi Pusparenihttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/5839Pengembangan Sistem Pengaduan Masyarakat Berbasis Citra Digital Menggunakan CNN dan Content-Based Filtering pada Kantor Kelurahan Peguyangan2026-06-23T15:50:22+07:00Putu Muna Saskara Widnyanawidnyanamuna@gmail.comNi Gusti Ayu Putu Harry Saptariniayu.harry@pnb.ac.idPutu Desiana Wulaning Ayuwulaning_ayu@pnb.ac.id<p><em>This study aims to develop a digital image-based road damage reporting system at the Peguyangan Subdistrict Office to assist in the automatic classification of road damage levels and detect duplicate reports. The system was built by integrating a MobileNetV2-based Convolutional Neural Network (CNN) for classifying road damage images and Content-Based Filtering (CBF) for identifying similar reports based on image similarity and proximity of location. Model evaluation was conducted using 5-fold cross-validation on a dataset of 500 images of damaged roads consisting of two classes: severe damage and minor damage. The test results showed that the model achieved an average accuracy of 89.40% with a best F1-score of 0.9194. The duplicate detection mechanism successfully grouped similar reports, thereby aiding in the prioritization of repairs. Additionally, black-box testing across 20 test scenarios showed that all system functions operated as designed with a 100% success rate. The research results indicate that the integration of CNN and CBF can support the management of road damage complaints in a more effective, structured, and responsive manner.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Putu Muna Saskara Widnyana, Ni Gusti Ayu Putu Harry Saptarini, Putu Desiana Wulaning Ayuhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6472Implementasi Logika Fuzzy Mamdani dalam Penentuan Status Gizi pada Balita2026-10-18T21:19:45+07:00Hamid Wijayahamidwijaya35@gmail.comRima Ruktiaririmaruktiarii@gmail.comNur Alisanuralisa2022@gmail.comSunyantisunyantis@gmail.comMuhammad Na’im Al Jum’ahmuhnaimusn@gmail.com<p><em>This study aims to develop a web-based system for determining the nutritional status of toddlers using the Mamdani Fuzzy Logic method at the Mawasangka Village Integrated Service Post (Posyandu). The system uses age, height, and weight as input variables. The study employed a qualitative approach through observation, interviews, documentation, and literature study. Data collected from Posyandu Mawasangka were used to construct membership functions, a rule base, inference processes, and defuzzification. System testing was conducted on 120 toddler records by comparing system outputs with nutritional status validation results from Posyandu. The results show that 104 of 120 records were consistent, resulting in an accuracy of 86.67%. Differences in several records were influenced by numerical rounding, calculation precision, numerical approaches used by the system, and implementation variations in membership functions, aggregation, and defuzzification. These findings indicate that Mamdani Fuzzy Logic can support a computerized process for determining toddler nutritional status at Posyandu Mawasangka.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Hamid Wijaya, Rima Ruktiari, Nur Alisa, Sunyanti, Muhammad Na’im Al Jum’ahhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6252Algoritma FP-Tree Pola Hubungan Kepadatan Penduduk dan Luas Daerah Kecamatan Kota Medan Upaya Peningkatan Pelayanan Publik2026-08-11T21:03:54+07:00Lismardiana Sembiringlismardianasembiring@gmail.comWanra Tariganmarbun2005@gmail.comJaidup Banjarnahorwansibro@yahoo.comDavidson Gintinglismardianasembiring@gmail.com<p><em>Good governance prioritizes excellent service to the community, in Law No. 25 of 2009 concerning public services can be used as a basis for improving public services by state civil servants or the government. To realize good public services, the government places the number and quality of existing staff/employees must be appropriate and have an understanding of good public services, so that public services are on target and the services provided can also bring the bureaucracy closer to the community. Poor service to the community can have a negative impact on public dissatisfaction. If there is a decline in public trust in the government, this will have a negative impact on social and economic development in the Medan City Government environment. Regional expansion can improve public services by bringing the government closer to the community, improving community welfare and creating targeted services according to the bureaucracy to the community. For the solution of this problem, researchers use the FP_Growth Algorithm method. From the results of the output process from the rule mining association above, it states the group results, items with a Support System value of 34% are {A4, A3, B2, C2, C4, D1, E1}; a support system worth 60% is {A1, B1, D2}; a support system that is 17% is {C6, E2, A2, B3, C1, E3, C3} providing recommendations for sub-districts that will be expanded by the Medan city government.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Lismardiana Sembiring, Wanra Tarigan, Jaidup Banjarnahor, Davidson Gintinghttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6508Sistem Pencadangan Basis Data Asinkron Berbasis Task Queue menggunakan RabbitMQ2026-10-18T21:11:22+07:00Ahmad Ihyaulhaq22523278@students.uii.ac.idIzzati Muhimmahizzati@uii.ac.id<p><em>Universitas Terbuka (UT</em><em>), Indonesia's distance-learning university with 39 regional offices, previously ran database backups separately at each office, leaving the central office without visibility into backup status, history, or success rates. Running large backups synchronously inside the main application also posed a potential risk to service performance. This study develops an asynchronous database backup architecture and a centralized monitoring dashboard for SIBU (Sistem Informasi Backup Universitas Terbuka), using a task queue to separate backup execution from the main application. The system was built with the waterfall method (requirement, design, implementation, and testing). RabbitMQ acts as the message broker in a competing consumer pattern: the backend publishes backup jobs to a queue, and a couple independent workers consume them, dumps the database, encrypts the output, and streams it to MinIO object storage. Black-box testing showed that all tested scenarios, including job submission, queueing, status updates, failure handling, and the central reporting dashboard, produced the expected results. The architecture is designed to keep backup workloads outside the main API process and to give the central office a single view of backup status across regional offices. Effects on latency, throughput, and resource usage were not measured and are left for future load testing.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Ahmad Ihyaulhaq, Izzati Muhimmahhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6535Penerapan Transfer Learning dan Explainable AI untuk Pengenalan Tokoh Wayang Berbasis Web2026-10-14T22:39:33+07:00Muhammad Al-Husainialhusaini@unsil.ac.idHen Hen Lukmanahenhenlukmana@unsil.ac.idMuhammad Sidik Asyakyasyaky@unsil.ac.idIchsan Fauzi Rachmanichsanfauzirachman@unsil.ac.id<p><em>Recognizing wayang characters remains difficult because many classes share similar visual traits, capture conditions vary, and labeled data is scarce. This study develops a web-based wayang character recognition system that combines transfer learning with visual explanation for classroom use. Two lightweight backbones, EfficientNetV2-S and DeiT-Tiny, were fine-tuned on ten characters using horizontal flipping, RandAugment, and TrivialAugment, then evaluated through five-fold stratified cross-validation. Performance was measured using accuracy, macro-F1, and the Matthews Correlation Coefficient. EfficientNetV2-S reached a mean accuracy of 86.8%, a macro-F1 of 0.87, and an MCC of 0.85, slightly above DeiT-Tiny at 86.0%, 0.86, and 0.83. Confusion analysis identified recurring errors between visually similar pairs such as Nakula-Sadewa and Gareng-Petruk. Grad-CAM overlays were integrated into a Streamlit interface so that every prediction carries visual evidence that non-experts can inspect. Median inference latency on a mid-range CPU was 28 ms for EfficientNetV2-S and 32 ms for DeiT-Tiny, fast enough for interactive use. The system shows that accuracy, speed, and interpretability can be balanced within the constraints of cultural heritage datasets.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Muhammad Al-Husaini, Hen Hen Lukmana, Muhammad Sidik Asyaky, Ichsan Fauzi Rachmanhttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6511Pengembangan Aplikasi Kakao Assistant Berbasis Flutter dan Large Language Model sebagai Asisten Virtual Petani Kakao di Bali2026-10-18T21:22:12+07:00I Nyoman Rai Widartha Kesumaraiwidartha@pnb.ac.idLuh Gede Putri Suardaniputrisuardani@pnb.ac.idAgus Adi Putrawanputrawanagusadi@pnb.ac.id I Putu Astya Prayudhaastyaprayudha@pnb.ac.id<p><em>Cocoa (Theobroma cacao L.) represents a strategic agricultural commodity supporting rural smallholder livelihoods across Bali Province, notably in Tabanan, Jembrana, Gianyar, and Buleleng regencies. Nevertheless, farm productivity and bean quality remain severely threatened by Cocoa Pod Borer (Conopomorpha cramerella), Helopeltis sp., and black pod rot (Phytophthora palmivora), exacerbated by the limited availability of agricultural extension officers in remote rural areas. This study develops Kakao Assistant, a cross-platform mobile virtual assistant built on Flutter integrating Google Gemini 3.1 Flash Large Language Model (LLM) under a single-role architecture dedicated to Cocoa Farmers. System development follows the Modified Waterfall lifecycle comprising requirements analysis, system architecture design, implementation, and empirical testing. The application provides multimodal agronomic consultations (text inquiries and pod disease image diagnostics), regional agroclimatic contextualization for Bali plantations, Text-to-Speech (TTS) voice synthesis, message bookmarking, and an on-device model evaluation module ensuring transparency of artificial intelligence inferences. Black-box testing validated that 100% of the 11 functional scenarios functioned in strict compliance with system specifications. Prototype testing verified the module's capability to track response latency, network fault resilience, and local FIFO history logs prior to extensive community deployment across Bali cocoa farms.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 I Nyoman Rai Widartha Kesuma, Luh Gede Putri Suardani, Agus Adi Putrawan, I Putu Astya Prayudhahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6536Analisis Sentimen Terkait Pelaksanaan Wifest pada Instagram dengan Metode Naïve Bayes2026-10-15T07:22:43+07:00Muhammad Ismail2141039@wicida.ac.idBartolomius Harpadharpad@wicida.ac.idYunitayunita@wicida.ac.id<p><em>Instagram comments capture public responses to an event, but reading them manually is slow and subjective. This study measures the sentiment tendency toward Wifest and assesses how well Naïve Bayes classifies it. A total of 120 @wifest comments were collected with IGCommentsExport, labeled positive or negative, and preprocessed through data cleaning, case folding, normalization, tokenization, stopword removal, and stemming. The text was then weighted with TF-IDF and classified using Naïve Bayes in RapidMiner Studio with a 70% training and 30% testing split. Of the 120 comments, 84 (70%) were positive and 36 (30%) were negative. The model achieved an accuracy of 86.11%, a precision of 85.71%, a recall of 96.00%, and a macro-average F1-score of 82.01%, exceeding a baseline that labels every comment positive (69.44%). However, negative-class recall was only 63.64%, so the model recognizes positive comments more reliably. It is concluded that Wifest drew mostly positive responses and that Naïve Bayes suits sentiment mapping of comments on this account. Organizers should use the model to estimate sentiment proportions and read negative comments directly, and future studies should add negative data and balance the classes.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Muhammad Ismail, Bartolomius Harpad, Yunitahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6270Redesain Website Security Management Berbasis User Centered Design dengan Fitur Rekomendasi Dokumen Berbasis Konten Diagnosis Usability Awal dan Pengembangan2026-08-20T06:31:51+07:00Stefanus Sapioperstevanusreba@gmail.comNi Gusti Putu Ayu Harry Saptariniayu.harry@pnb.ac.idI Putu Astya Prayudhaastya@pnb.ac.id<p><em>The Security Risk Management (SRM) Department of PT Freeport Indonesia uses an internal website to distribute security policies and Standard Operating Procedures (SOPs), but its layered navigation and the absence of any mechanism for surfacing related documents make retrieval inefficient. This study diagnosed the usability of the existing system, elicited user requirements, and designed and implemented a redesigned system based on User-Centered Design (ISO 9241-210) integrated with a content-based document recommendation feature using TF-IDF and Cosine Similarity. A survey of 40 purposively selected internal users produced a mean System Usability Scale score of 65.25 (SD 14.06; 95% CI [60.75, 69.75]; Cronbach's alpha 0.804), falling in the Marginal category. A 12-item adapted User Experience Questionnaire placed all six dimensions above the neutral point, with Perspicuity numerically lowest, although the confidence intervals of all dimensions overlapped substantially. The most frequently prioritised user requirement was easier navigation and document search. The resulting system passed all ten Black Box Testing scenarios, and the recommendation module was documented at the configuration level to support replication. Post-implementation usability evaluation has not yet been conducted; this article therefore makes no claim regarding the effects of the redesign.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Stefanus Sapioper, Ni Gusti Putu Ayu Harry Saptarini, I Putu Astya Prayudhahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6182Prediksi Penjualan Mobil pada PT Astra Daihatsu Menggunakan Algoritma Multiple Linear Regression2026-07-31T10:39:55+07:00Nyoman Agus Mahardiputraagusputraa9999@gmail.comNi Wayan Wisswaniwisswani@pnb.ac.idI Putu Bagus Arya Pradnyanabagusarya12@pnb.ac.id<p><em>Determining Terios car stock at PT Astra Daihatsu Ngurah Rai Tuban currently relies on manual, intuition-based forecasting, which is susceptible to realization gaps. This study aims to develop a Machine Learning-based computational model to predict sales volume objectively and measurably. This quantitative research adopts the CRISP-DM framework by applying the Multiple Linear Regression (Ordinary Least Squares) algorithm. Observational data includes promotion costs, the number of active sales personnel, and event frequencies from January 2023 to March 2026. The results show that the model satisfies the classical assumption tests (BLUE) and achieves a Coefficient of Determination (R²) of 74.72%. Statistical testing proves that all variables simultaneously have a significant effect on sales. Partially, strategic findings reveal that active sales personnel and event frequency provide a significant positive impact, whereas promotion costs are statistically insignificant. Implementing the model into a prediction calculator system successfully transforms the estimation process into an automated workflow, recommending management to shift marketing strategies from passive promotions to strengthening direct sales forces and exhibitions.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Nyoman Agus Mahardiputra, Ni Wayan Wisswani, I Putu Bagus Arya Pradnyanahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6178Implementasi YOLOv8n dan Deepsort untuk Deteksi Kecelakaan Serta Analisis Lalu Lintas Pada Video CCTV Jalan Raya2026-07-29T17:04:17+07:00Anak Agung Gede Dwitya Nugraha Parardhyaanakagunggede04@gmail.comI Wayan Candra Winetraanakagunggede04@gmail.comI Nyoman Rai Widartha Kesumaanakagunggede04@gmail.com<p><em>The increasing number of motorized vehicles has created challenges in traffic management, particularly in monitoring highway incidents. Conventional CCTV surveillance relies heavily on human observation, making it prone to human error and delayed emergency responses. Therefore, this study proposes a web-based traffic analysis system utilizing Computer Vision technology. The YOLOv8n deep learning model is employed to perform accident detection, speed estimation, and vehicle counting simultaneously. The model was trained using 36 hyperparameter combinations in Google Colab. The best model achieved a mean Average Precision (mAP@50) of 0.862, with an mAP@50 of 0.881 for the accident class. Furthermore, the developed system was integrated with live CCTV streams and achieved an average processing speed of 25 Frames Per Second (FPS). The proposed system provides an integrated solution for real-time traffic analysis and accident detection, supporting more effective traffic monitoring.</em></p>2026-10-31T00:00:00+07:00Copyright (c) 2026 Anak Agung Gede Dwitya Nugraha Parardhya, I Wayan Candra Winetra, I Nyoman Rai Widartha Kesumahttps://ejurnal.methodist.ac.id/index.php/methomika/article/view/6547Implementasi Metode Hybrid Support Vector Machine dan K-Nearest Neighbor untuk Klasifikasi Risiko Kesehatan Maternal pada Ibu Hamil2026-10-28T14:21:26+07:00Mochamad Febry Herdianmochamadfebryherdian2020@gmail.comHadi Zakariadosen00274@unpam.ac.id<p><em>The classification of anemia risk levels in pregnant women faces challenges regarding accuracy and processing speed when handling complex medical record data. This study optimizes the accuracy of anemia risk prediction using a hybrid approach combining Support Vector Machine (SVM) and K-Nearest Neighbor (KNN). Data analysis was conducted on a dataset of 1,500 medical records, which were extracted into five clinical features and standardized using the Z-Score. Initial classification using SVM with a Radial Basis Function (RBF) kernel showed misclassification tendencies at the boundary regions between classes; therefore, it was integrated with KNN using an Ensemble Weighted Voting approach with an equal weight of 0.5 for each algorithm. The test results on 300 evaluation data demonstrate that the standalone SVM model achieved an accuracy rate of 74.0%, while the hybrid SVM-KNN model reached 71.0%. Although the SVM model demonstrated higher accuracy, the hybrid SVM-KNN integration on the 1,500 maternal medical records dataset still offers a complementary classification framework—combining the global generalization of SVM with the local sensitivity of KNN—making it highly adaptable for supporting early anemia risk detection systems and objective clinical decision-making</em></p>2026-11-16T00:00:00+07:00Copyright (c) 2026 Mochamad Febry Herdian, Hadi Zakaria