SENTIMENT ANALYSIS USING NAIVE BAYES CLASSIFIER WITH TF-IDF AND N-GRAM

• Main Author: Fakhri Alhafizh, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/81406
• Institution: Institut Teknologi Bandung

Machine learning can be used to solve text classification problems in this final project. Sentiment analysis is the process of understanding opinions towards a particular subject. This final project focuses on sentiment analysis of product reviews on an e-commerce platform using the Naive Bayes Classifier algorithm combined with TF-IDF (Term Frequency-Inverse Document Frequency) and N-Gram techniques. The aim of this final project is to develop a Naive Bayes Classifier model and classify sentiments as positive or negative in product reviews, which is useful for providing deeper insights into customer perceptions of a product. The methods used include collecting product review data from e-commerce sites, text preprocessing to remove noise, and feature extraction with TF-IDF and N-Gram to numerically model the text. Subsequently, the Naive Bayes Classifier algorithm is applied for sentiment classification. The results of the final project show that the use of TF-IDF provides the best performance in sentiment classification compared to other combination methods, with an accuracy of 90.38%. The resulting model demonstrates high accuracy in predicting the sentiment of product reviews.