IMPLEMENTATION OF TEXT CLASSIFICATION ON USER REVIEWS IN DANA APPLICATION USING SUPPORT VECTOR MACHINE (SVM) AND GAUSSIAN NAÏVE BAYES (GNB)
Alfatha Fitrah Insan, Detty Purnamasari · Jurnal Ilmiah Teknologi dan Rekayasa · 2024
Conventional methods or devices are unable to efficiently process large volumes and diverse categories of information, which are collectively referred to as big data. Text mining is a commonly used technique for analyzing big data. This study evaluates the effectiveness of Support Vector Machines (SVM) and Gaussian Naïve Bayes (GNB) in the classification of user reviews from the DANA application, obtained from the Google Play Store. The four fundamental phases of the investigation are data collection, data preparation, data modelling, and evaluation. This study utilized a dataset of 15.451 user reviews, dividing it into three subsets with varying data sizes and each subset having varying training-to-testing ratios. The evaluation will calculate four measurements, which are accuracy, precision, recall, F1-Score, and ROC Curve. The results illustrate that SVM and GNB achieved accuracy rates of at least 75%. SVM achieves an average accuracy of 84%, 88%, and 91%, while GNB achieves an average accuracy of 71%, 81%, and 85%. Based on the implementation results, sentiment analysis is more effective when performed with SVM than with GNB.