Pengembangan Aplikasi Analisis Sentimen Anies Baswedan pada Twitter

Abstract

The main objective of this research is to develop a sentiment analysis application about Anies Baswedan. The data was obtained from Twitter on 1 April until 28 April 2020. The tweet data was classified into three classes, which are positive, neutral, and negative. The feature extraction method used is Term Frequency – Inverse Document Frequency (TF-IDF). While the classification method used is Multinomial Naïve Bayes. The software development method used in this research is an incremental model, which is divided into two parts. Those are incremental part one that perform text preprocessing, term weighting using TF-IDF, and classification of the data using Multinomial Naïve Bayes algorithm. While the incremental part two will develop the user interface of sentiment analysis application. The accuracy result of the model obtained in this research is 0.71. The average evaluation scores of the model are 0.71 for precision, 0,71 for recall and 0.71 for f-measure.