PERBANDINGAN METODE KLASIFIKASI SUPPORT VECTOR MACHINE DAN NAÏVE BAYES PADA ANALISIS SENTIMEN KENDARAAN LISTRIK

Abstract

Electric vehicles are one of the solutions that can be used to deal with the problem of greenhouse gas emissions. Switching to electric vehicles can be an effective solution because electric vehicles have many advantages. However, the acceptance of electric vehicles in Indonesia depends on the opinions or sentiments given by the community. Sentiment analysis can provide a specific picture of how the sentiment or opinion given by the Indonesian people towards electric vehicles. The sentiment analysis process is carried out using the Python programming language. This research compares SVM and Naïve Bayes methods in sentiment analysis in terms of accuracy and time efficiency. A total of 717 data were used as test data and SVM correctly classified 150 negative data, 152 neutral data, and 277 positive data. Meanwhile, Naïve Bayes correctly classified 166 negative data, 143 neutral data, and 282 positive data. training time required for the SVM method is 37.42 seconds while Naïve Bayes is 0.10 seconds. Naïve Bayes is the best method in this study because of its high accuracy and fast training time.