Online buying and selling transactions are increasing in Indonesia due to the ease of using marketplace platforms, and online shopping saves more time than offline shopping. Each marketplace has advantages and disadvantages, this affects customer sentiment who have made transactions on the marketplace platform. This research uses customer opinion from tweet data based on positive or negative sentiments to compare the Naive Bayes (NB) and Support Vector Machine (SVM) classification algorithms with the aim of finding out the best classification algorithm based on the accuracy value for sentiment analysis using the marketplace platform. The tweet data in this study was taken from October 18 to November 11, 2022. To test the performance of the NB and SVM classification algorithms using the Cross Validation method and from the results of the comparison test that the SVM algorithm has the best accuracy value compared to the NB algorithm. Where the accuracy value of Tokopedia uses the NB algorithm is 85.34%, and the accuracy value uses SVM 86.82%, the accuracy value for Shopee uses NB is 80.04%, and the accuracy value uses SVM 80.91%. and Lazada which uses the NB algorithm has an accuracy value of 83.52%, while the accuracy value uses SVM 88.93%, which means that the use of the SVM algorithm has the best level of accuracy.
Copyrights © 2023