Implementasi Naïve Bayes dalam Merekomendasikan Pembelian Barang pada Aplikasi E-commerce

Abstract

Online shopping behavior in Indonesia has become a significant cultural and economic trend, supported by the emergence of various e-commerce platforms that simplify the shopping experience. These platforms not only offer convenience through features like promotions and home delivery but also generate consumer reviews that reflect the quality of products and services. This study aims to develop a recommendation system using the Naïve Bayes algorithm to classify customer reviews into positive and negative sentiments. The data used in this study were collected from MSI Official Store reviews across three major e-commerce platforms: Tokopedia, Shopee, and Blibli, totaling 921 cleaned review entries out of 1,287 raw data points. The dataset was processed through a series of text pre-processing stages including case folding, tokenization, stopword removal, and stemming. The dataset was then split into 80% training and 20% testing. The model evaluation showed an accuracy of 82% for training data and 71% for testing data. These results indicate the effectiveness of the Naïve Bayes algorithm in generating sentiment-based product recommendations. The final model was deployed into a web-based application to assist users in determining whether a product is recommended, based on the balance of positive and negative reviews. This study contributes to the development of intelligent recommendation systems that enhance user decision-making processes in digital marketplaces.