Analisis Performa InSet Lexicon Pada Pelabelan Sentimen Pengungsi Rohingya Menggunakan SVM

Abstract

Sentiment analysis is a study that evaluates a person's opinions, sentiments, judgments, attitudes, and emotions towards certain entities such as issues or events. One common approach is the lexicon-based method, where text is analyzed using a sentiment dictionary such as InSet Lexicon that groups words into positive or negative categories. The dataset used in this research is tweets from application X (Twitter) regarding Rohingya refugees in Indonesia. Data classification is done with the Support Vector Machine (SVM) algorithm. The results of labeling using InSet Lexicon found 886 positive sentiments, 2906 negative sentiments, and 353 neutral sentiments with an average accuracy value generated from all tests reaching 84.9%.