ANALISIS DAN PERBANDINGAN STOPWORD TERHADAP AKURASI ANALISIS SENTIMEN TEKS DENGAN MENGGUNAKAN TF-IDF STUDI KASUS NLP

Abstrak

In the rapidly evolving digital era, the amount of online text data has significantly increased, encompassing product reviews, social media comments, and news articles. Sentiment analysis is crucial for understanding public opinion. This research aims to develop a more relevant stopword list using the TF-IDF algorithm to enhance text representation in sentiment analysis. Additionally, it evaluates and compares the impact of using stopwords generated by the TF-IDF algorithm on the accuracy of sentiment analysis models, compared to using Sastrawi stopwords. The results show that TF-IDF helps identify less important words, but Sastrawi stopwords are better at recognizing context. Evaluation with different data split ratios (90:10, 80:20, 70:30) showed the highest accuracy of 0.789 at the 80:20 ratio, although there is room for improvement. This study is expected to improve the performance of sentiment analysis models with a more suitable stopword list.