Analisis Sentimen Ulasan Aplikasi Gojek Di Google Play Store Menggunakan Multinomial Naïve Bayes

Lundam, Ingrida Suryani
Darmawan, J.B. Budi

Prosiding Seminar Nasional Rekayasa Teknologi Industri dan Informasi ReTII
2024: Prosiding Seminar Nasional ReTII ke-19 (Edisi Penelitian)

The development of technology in Indonesia has facilitated various human activities, including in the field of transportation through online-based services such as Gojek, a decacorn company that provides various transportation services. Gojek has changed the way Indonesian people move around in urban areas. Understanding user sentiment towards the Gojek application, especially through reviews on the Google Play Store, is key to improving and improving the services offered. This study aims to evaluate the performance of sentiment analysis on 4746 Gojek application reviews on the Google Play Store using the Naive Bayes algorithm. The methods used include data translation processes, labeling using Vader, text preprocessing, word weighting with TF-IDF, data balancing with SMOTE, and model testing using K-fold cross validation. Sentiment classification is done using Multinomial Naive Bayes, and model evaluation using a confusion matrix. The research results show that testing with K-fold cross validation k = 11 and alpha = 0.001 produces the highest accuracy of 92.25%, precision of 92.16%, recall of 89.71%, and F1-score of 90.42%.