METHODIKA: Jurnal Teknik Informatika dan Sistem Informasi

ANALYZING LECTURER PERFORMANCE FACTORS FROM COURSE EVALUATION SURVEYS USING K-MEANS CLUSTERING AND C4.5 CLASSIFICATION

2025-12-04T16:09:18+07:00

One of the quality of education of a college can be seen from the quality of the performance of the lecturers in the higher education Tridharma namely education, research, and development and community service. This study aims to analyze the lecturer performance factors based on the course evaluation survey on the Tridharma of higher education in the implementation of academic evaluations and decision making for lecturers in the Computer Science study program, Pakuan University. The research method applies a combination of two data mining methods, namely k-means clustering and C4.5 classification used in the assessment of the performance of lecturers, especially in the process of education and college teaching which includes pedagogical, professional, personality and social competencies. The results of the K-means clustering mining process were assessed by learning, namely 5 sufficient lecturer clusters, 16 good cluster lecturers and 14 excellent cluster lecturers. C 4.5 classification is used to see the connectedness of factors such as learning, publication, education, PKM, support. This study shows that publication criteria are the most influential factors in the performance assessment of lecturers. Testing the level of accuracy using the K-fold Cross Validation method with 5-fold Cross Validation is 80.00% and 7-fold Cross Validation which is 82.86%.

IMPLEMENTASI K-MEANS RFM DAN HOLT-WINTERS EXPONENTIAL SMOOTHING ADDITIVE DALAM SISTEM BUSINESS INTELLIGENCE UNTUK STRATEGI PENGELOLAAN PELANGGAN PADA PERUSAHAAN TRANSPORTASI

2025-12-04T16:24:00+07:00

The growth of customer data in the transportation industry drives the need for analytical systems capable of segmenting customers objectively and strategically. This study aims to apply the K-Means Clustering method based on the Recency, Frequency, and Monetary (RFM) model for customer segmentation and utilize the Holt-Winters Exponential Smoothing Additive method to forecast passenger numbers. The dataset comprises 10,440 customer transactions from PT XYZ during the 2022–2024 period. RFM values were calculated, normalized, and processed using the K-Means algorithm to produce three customer clusters: Loyal, Regular, and Passive. Subsequently, the Holt-Winters method was employed to forecast passenger numbers, achieving the smallest Mean Absolute Percentage Error (MAPE) of 6.88%, indicating a high level of accuracy. The results were visualized through an interactive dashboard using Tableau, enabling management to make data-driven decisions. This research demonstrates that integrating segmentation and forecasting methods into a Business Intelligence system can enhance the effectiveness of marketing strategies and the operational efficiency of the company.

RANCANG BANGUN SISTEM INFORMASI KLINIK BERBASIS WEB DENGAN LARAVEL

2025-12-04T16:30:37+07:00

Previously, patient, doctor, and examination data management at Lestari Mandiri Clinic was still done manually, which often caused delays in recording and data errors. To overcome these problems, a web-based clinic information system was developed using the Laravel framework and MySQL database. The system development went through several stages, namely needs analysis, system design, implementation, testing, and maintenance. This system is designed with a simple and easy-to-use interface to suit the needs of clinic staff. The test results showed that the system can run well and fulfill the main functions such as recording patient data, managing doctors, and examination results. In addition, this system can also be further developed by adding features such as automatic reports, control schedule reminders, and drug stock management. With this information system, the clinic administration process becomes more efficient, accurate, and supports better service to patients.

MODEL PENGUKURAN KINERJA RANTAI PASOK BERBASIS GREEN SCOR DAN FUZZY AHP: STUDI KASUS PT. ARTERIA DAYA MULIA

2025-12-04T16:37:29+07:00

Supply chain performance measurement plays a crucial role in supporting operational continuity and corporate competitiveness, especially in meeting the demands for efficiency, effectiveness, and environmental sustainability. Imbalances in supply chain management can lead to resource waste, environmental pollution, and decreased customer satisfaction. PT. Arteria Daya Mulia, a rope manufacturing company, currently lacks a supply chain performance measurement system that fully incorporates sustainability aspects. This study aims to design a performance measurement model based on the Green SCOR framework and the Fuzzy AHP method as a strategic decision-making tool that considers sustainability dimensions. Performance indicators were determined according to the five main Green SCOR processes (Plan, Source, Make, Deliver, and Return), comprising 14 KPIs developed through literature review and field validation. Data were collected through observations, interviews, and questionnaires, then processed using the Fuzzy AHP method to obtain the priority weight of each indicator. The results show that the total supply chain performance score is 88, calculated by combining the weights with the Snorm de Boer values. Several indicators demonstrated excellent performance with a maximum Snorm value (100). However, one critical indicator was identified with the lowest Snorm value—% Error-free Return Shipped in the Return process—scoring 0.02 with a final SCM score of 0.0008, indicating the need for immediate improvement. The developed information system also generates automatic improvement recommendations based on the measurement results. This model is expected to assist the company in monitoring, evaluating, and continuously improving supply chain performance.

IMPLEMENTASI LOGIKA FUZZY MAMDANI DALAM SISTEM PENILAIAN KESEHATAN MAKANAN KEMASAN BERDASARKAN LABEL NUTRITION FACTS

2025-12-04T16:46:05+07:00

The growth of the packaged food industry has increased the need for an easy-to-understand health assessment system for consumers, especially those with limited nutrition literacy. This study develops a Mamdani fuzzy logic-based decision support system to evaluate the healthiness of packaged foods using Nutrition Facts labels. The system processes nutritional parameters such as fat, sugar, salt, fiber, protein, fruit/vegetable/nut content, and calorie content, converting them into linguistic categories like "low," "moderate," and "high" for easier interpretation by lay users. It effectively handles uncertainties and ambiguities in nutrition data, providing classifications like "Unhealthy," "Healthy," or "Very Healthy." Implemented through a web platform using Python and Flask, the system was tested with five food samples, achieving an 80% agreement with the official NutriScore classification. This indicates the potential of the system as a reliable, practical tool to help consumers make quicker and more accurate dietary decisions and improve nutrition awareness.

EVALUATING THE EFFECTIVENESS OF DISTILBERT FOR SENTIMENT ANALYSIS OF PLAYER FEEDBACK IN GAME DEVELOPMENT

2025-12-04T16:51:19+07:00

Real-time sentiment analysis (SA) plays an increasingly vital role in enhancing player experience through emotion-aware game design. By enabling systems such as dynamic difficulty adjustment, adaptive non-playable character (NPC) behavior, and responsive narrative progression, SA allows games to respond intelligently to player emotions. This study investigates the effectiveness of DistilBERT, a lightweight transformer-based language model, for multi-label emotion classification using the GoEmotions dataset, which includes 27 fine-grained emotion categories. The model’s performance was evaluated in terms of classification accuracy and computational efficiency. Experimental results reveal that DistilBERT delivers surprisingly strong performance despite its reduced size, making it a viable candidate for real-time applications in resource-constrained environments. These findings indicate that lightweight transformer models can support emotionally adaptive gameplay without significant trade-offs in latency or accuracy. Future work will focus on integrating DistilBERT into a live game environment to assess its impact on player engagement and real-time system responsiveness.

PENERAPAN METODE FUZZY TIME SERIES CHEN DALAM SISTEM PERAMALAN PRODUKSI UNTUK OPTIMALISASI PENGADAAN BAHAN BAKU PADA PERUSAHAAN MANUFAKTUR

2025-12-04T16:58:00+07:00

Fluctuating and dynamic production demands require manufacturing companies to adopt adaptive, data-driven planning systems. However, in manufacturing companies producing plastic ropes such as twine, nets, and yarn, production planning is still conducted manually without a systematic quantitative approach. This study aims to design a production forecasting system using the Fuzzy Time Series Chen method, which can address uncertainty in time series data. Monthly production data from January 2022 to December 2024 were used for testing. The results show that this method provides good forecasting accuracy, with MAPE values of 30.15% for twine, 17.89% for nets, and 16.91% for yarn. The production estimates for January 2025 are 70,162 units (twine), 50,599 units (nets), and 81,315 units (yarn). These findings indicate that the FTS Chen method can improve efficiency and accuracy in production planning.

PEMANFAATAN GOOGLE EARTH ENGINE DAN ALGORITMA RANDOM FOREST UNTUK PEMETAAN LAHAN PERKEBUNAN JERUK

2025-12-04T17:03:08+07:00

This study employed Google Earth Engine (GEE) and the Random Forest algorithm to map citrus plantations in Silimakuta District, Simalungun Regency, North Sumatra. As a major citrus production center—reaching 840,000 quintals in 2023—the region faces challenges in producing accurate and efficient maps of plantation distribution. By processing Sentinel-2 and Sentinel-1 satellite imagery in GEE, this study provides a more detailed and reliable mapping solution. The Random Forest model achieved a land-cover classification accuracy of 97% and a Kappa coefficient of 96.3%, demonstrating the method’s effectiveness for land mapping. This approach can overcome existing limitations in land data and deliver visual information useful for increasing citrus plantation productivity in the region. Therefore, the combined use of Google Earth Engine and the Random Forest algorithm shows strong potential to support more optimal and sustainable land management.

PENGEMBANGAN APLIKASI MANAJEMEN DATA BUKU MENGGUNAKAN APLIKASI VISUAL STUDIO CODE DENGAN FRAMEWORK LARAVEL

2025-12-04T17:08:23+07:00

Book data management in libraries and bookstores often faces challenges related to efficiency, accuracy, and speed in recording and searching for information. Manual systems still used in many places tend to be slow, error-prone, and impractical in data processing. Therefore, a software-based solution is needed that can support the digitization of the book management process. This study aims to develop a book data management application using Visual Studio Code as the development environment and the Laravel framework as the backend basis. The research method used is Research and Development (R&D) with a prototyping approach, which involves the stages of needs analysis, design, implementation, testing, and evaluation and revision. The results show that the application is able to simplify the process of recording, editing, deleting, and searching for book data efficiently. The designed interface is responsive and easy to use, so it can be operated by users with diverse technical backgrounds. System testing achieved a success rate of up to 95% in terms of functionality and data accuracy, confirming that the use of Laravel in book management application development makes a significant contribution to increasing the productivity, security, and effectiveness of book data management.

FEW-SHOT LEARNING FOR AML CELL CLASSIFICATION USING PROTOTYPICAL

2025-12-04T17:19:54+07:00

Accurate blood cell classification is crucial for diagnosing Acute Myeloid Leukemia (AML) but limited medical data poses challenges for traditional machine learning models. This study presents a Few-Shot Learning (FSL) framework utilizing a Prototypical Network architecture with a ResNet-34 backbone to classify AML blood cell types from microscopic images. In this study, we utilize datasets consisting of 15 morphologically distinct cell classes. A 15-way, 5-shot, 5-query episodic setup was adopted to simulate data-scarce conditions. Evaluation via 5-fold cross-validation yielded strong performance, with an average accuracy of 97.76%, precision of 98.78%, recall of 96.55%, and F1-score of 97.76%. FSL training times were consistent (4.22–4.26 minutes per fold), and t-SNE along with confusion matrices confirmed the model’s ability to distinguish similar cell types. To validate the approach, its performance was compared with a conventional supervised CNN using the same ResNet-34 backbone. The FSL model outperformed the CNN across all metrics such as accuracy (98.32% vs. 77.25%), precision (98.55% vs. 76.87%), recall (98.31% vs. 78.66%), and F1-score (98.33% vs. 75.26%), while also requiring far less training time (~4.24 min/fold vs. ~420 min total). These results highlight the promise of FSL based methods for accurate, efficient, and scalable hematologic diagnostics in data limited settings.

RANCANG BANGUN APLIKASI HELP DESK BERBASIS WEB DI PT TRI STAR SURYA GEMILANG PURWOKERTO MENGGUNAKAN METODE DESIGN THINKING

2025-12-04T17:26:53+07:00

This research aims to design and develop a web-based Help Desk application at PT Tristar Surya Gemilang Purwokerto using the Design Thinking method. The application is developed to improve efficiency and transparency in managing complaints, while accelerating problem resolution through an integrated system. The research follows four main stages: Empathize, Define, Prototype, and Testing, with a focus on addressing user needs to develop an effective solution. The results of Blackbox Testing show that the application meets the expected functionality for both admins and users, with features aligned with actual user needs. Key features tested include a secure login page, clear ticket data display, and an informative dashboard and reporting system. Figures 3 to 7 in the results show an intuitive user interface that ensures an efficient workflow. The system is expected to enhance the IT team’s performance in handling complaints and technical issues more efficiently, reduce communication errors between departments, and speed up problem resolution. The implementation of this system aims to improve productivity, operational effectiveness, and user satisfaction through better-managed complaint handling.

PERANCANGAN INTERFACE APLIKASI INFORMASI RUANG KELAS KOSONG BERBASIS MOBILE DENGAN METODE USER-CENTERED DESIGN (UCD)

2025-12-04T17:31:51+07:00

The rapid advancement of information technology in higher education necessitates greater efficiency in managing academic resources, including the provision of information on empty classrooms. Students and lecturers often encounter difficulties in locating available rooms due to manual processes and non–user-friendly interfaces. This study aims to design a web-based application interface for empty classroom information using the User-Centered Design (UCD) approach to enhance User Experience (UX). A qualitative research method was employed, incorporating observation and interviews with prospective users to understand their needs, expectations, and usage contexts. The design outcome is a prototype interface comprising the splash screen, login, main page, and user profile. The design emphasizes ease of navigation, clarity of information, and continuous user involvement throughout the design process. By implementing UCD, the application is expected to provide an effective solution for real-time discovery of empty classrooms.

PREDIKSI KEHADIRAN PESERTA RAKORNAS APTIKOM MENGGUNAKAN METODE LEAST SQUARE

2025-12-04T17:36:46+07:00

APTIKOM (Asosiasi Pendidikan Tinggi Informatika dan Komputer) is an association that brings together Indonesian Universities offering Computer Science and Information Technology programs, playing a role in curriculum development, educational standards, and professional certification in the field of Information Technology (IT). Predicting the number of participants at the APTIKOM National Conference can provide an estimate of the number of participants for the following year, helping the conference organizing committee plan activities more effectively anda data-driven using the Least Square method. The prediction results are analyzed and categorized into activity levels based on participant attendance frequency. The model evaluation results indicate that the Least Square method can be effectively used to predict participation patterns and generate useful category analyses as a basis for decision-making by APTIKOM.