Comparison of Classification Algorithms for Sentiment Analysis on Movie Comments

Dian Sa'adillah Maylawati (Teknik Informatika, UIN Sunan Gunung Djati Bandung)
Melani Nur Mudyawati (Teknik Informatika, UIN Sunan Gunung Djati Bandung)
Muhammad Humam Wahisyam (Teknik Informatika, UIN Sunan Gunung Djati Bandung)
Riki Ahmad Maulana (Teknik Informatika, UIN Sunan Gunung Djati Bandung)

Gunung Djati Conference Series
Vol. 3 (2021): Mini Seminar Kelas Data Mining 2020

The film industry is growing rapidly nowadays, various genres and storylines are nicely packaged to convey messages and entertain audiences. Sentiment analysis technology can be used for the advancement of the film industry as well as film recommendations that need to be presented next. This study aims to compare several algorithms used for sentiment analysis of movie reviews or comments. The algorithms used in this study are K-Nearest Neighbor (k-NN), Naïve Bayes Classifier (NBC), and Logistic Regression. The experimental results using 25,000 film comment datasets show that Logistic Regression has the highest accuracy rate with an accuracy of 89%, compared to Naïve Bayes' accuracy of 86%, while k-NN is 65.22%.