ANALISIS SENTIMEN TWITTER TERHADAP WACANA PENUNDAAN PEMILU DENGAN METODE SUPPORT VECTOR MACHINE

Abstract

Machine learning plays an important role in managing important issues to classify and predict information that develops ahead of the General Election in Indonesia. Especially in knowing public sentiment on the discourse on the postponement of the 2024 election through Twitter social media. So it is necessary to analyze the discourse by categorizing it as positive or negative. The Support Vector Machine (SVM) model is used for analysis and classification. The sample data used were 100 tweets data which were then scraped in the period January 2022 – May 2022. The processing was done using Python programming and Jupyter Lab tools. Before doing the analysis, do preprocess to eliminate unnecessary words and information so that the level of accuracy of the results of this Twitter sentiment classification can provide a closer picture of reality. As for the results of the grouping carried out positive sentiment in as many as 40 data tweets and negative sentiment in as many as 60 data tweets. The classification test results on tweets data with a good level of accuracy of 92%. These results are expected to be a reference for future researchers who want to improve accuracy or analysis results.