Analisis Sentimen Pelanggan terhadap Produk UMKM pada Marketplace Menggunakan Algoritma Naïve Bayes

Abstract

The rapid growth of online marketplaces has provided significant opportunities for Micro, Small, and Medium Enterprises (MSMEs) to market their products to a broader audience. Customer reviews posted on marketplaces contain valuable information regarding customer satisfaction, experiences, and perceptions of the products and services offered. However, the increasing volume of reviews makes manual analysis inefficient and time-consuming. Therefore, an automated method is needed to process and analyze review data effectively. This study aims to analyze the sentiment of MSME reviews on marketplaces using the Naïve Bayes algorithm and Natural Language Processing (NLP) techniques. This research employs a text mining approach consisting of data collection, preprocessing stages including case folding, tokenization, stopword removal, and stemming, followed by term weighting using the Term Frequency-Inverse Document Frequency (TF-IDF) method. The processed data are then classified into three sentiment categories: positive, negative, and neutral, using the Naïve Bayes algorithm. Model performance is evaluated using a confusion matrix as well as accuracy, precision, recall, and F1-score metrics. The results indicate that the Naïve Bayes model performs well in classifying sentiment from marketplace reviews. Based on the confusion matrix, the model correctly classified 176 out of 200 testing data instances. The evaluation results show an accuracy of 88.00%, precision of 87.25%, recall of 86.80%, and F1-score of 87.02%. The best performance was achieved in the positive sentiment class, with a precision of 92.96%, recall of 94.29%, and F1-score of 93.62%. These findings demonstrate that the Naïve Bayes method combined with NLP techniques is effective for conducting sentiment analysis of MSME reviews on marketplaces. This study is expected to assist MSME owners in understanding customer opinions more quickly and accurately, thereby providing valuable insights for improving product quality and service performance.