Analisis Sentimen Pengguna Twitter Menggunakan Support Vector Machine Pada Kasus Kenaikan Harga BBM

Abstract

Twitter is one of the social media with the most active users, which is 24 million active users. Information published on twitter contains comments from users on an object. Sentiment analysis is used to determine whether the data includes negative comments or positive comments because the comments taken on twitter are textual data. The method used in this sentiment analysis is Support Vector Machine (SVM) about public comments on fuel price increases on twitter. The comment data used was 258 data on September 4, 2022 because on that date it was exactly the day after the fuel price increase. First, preprocessing is done to remove unnecessary words or information. Then the data is divided into training data by 80% and testing data by 20%. The accuracy rate is 82.69%, sensitivity is 100%, and specificity is 79.07%. Then from the results of testing 52 data obtained the results of 43 negative comments and 9 positive comments so that it can be concluded that more people disagree with the increase in fuel prices.

Keywords

• comment
• twitter
• preprocessing
• sentiment
• support vector machine
• fuel price