E-Marketplace Selection for Indonesian MSMEs: A Sentiment Analysis Approach Combining Lexicon-Based and Support Vector Machine

This study aims to examine e-marketplace performance using Lexicon-based research method and Support Vector Machine based on sentiment analysis to help Micro, Small, and Medium enterprises with the best-fit recommendation. Six popular e-marketplaces in Indonesia, including Tiktokshop, Shopee, Lazada, Tokopedia, Bukalapak, and Blibli.com, were selected as case studies. The findings show that Lazada and Tiktokshop gained the highest positive review value, totaling 96.04% with an accuracy rate of 93%, followed by Tiktokshop with 92% accuracy. Tokopedia ranked third with a review value of 95.47% and an accuracy value of 91%, while Blibli.com attained a total of 94.56% with 95% accuracy. Shopee secured value of 91.46% with 90% accuracy, and Bukalapak concluded with 90.92% and 89% accuracy. This research provides two significant contributions. First, it makes a scientific contribution by employing a combined lexicon-based and SVM model for sentiment analysis of e-marketplaces in Indonesia. Second, while this study is in the Indonesian context, the approach and findings offer a practical contribution for Micro, Small, and Medium enterprises owners selecting suitable e-marketplaces for their marketing strategies.