A Comparative Analysis of Support Vector Machines and Logistic Regression for Analyzing User Sentiment Regarding App UI/UX

Abstract

Advances in digital technology have driven an increase in the use of mobile and web applications, making the quality of the User Interface (UI) and User Experience (UX) crucial factors for user satisfaction. User comments on app stores can be leveraged to assess UI/UX quality through sentiment analysis. This study aims to evaluate the comparative performance of the Support Vector Machine (SVM) and Logistic Regression algorithms in classifying user sentiment regarding app UI/UX. Research data was obtained from the Sentiment Analysis Dataset—App Reviews available on Kaggle. The research process included data cleaning, feature extraction using the Term Frequency–Inverse Document Frequency (TF-IDF) method, model training, and model testing with an 80:20 training-to-test data split. Performance was measured using accuracy, precision, recall, and F1-score metrics. The findings show that SVM yields higher results compared to Logistic Regression, with an accuracy of 88.26%, precision of 91.13%, recall of 86.93%, and an F1-score of 88.98%. Meanwhile, Logistic Regression achieved an accuracy of 86.75% and an F1-score of 87.54%. These findings indicate that SVM is more effective for classifying user sentiment regarding the UI/UX of text-based applications. Advances in digital technology have driven an increase in the use of mobile and web applications, making the quality of the User Interface (UI) and User Experience (UX) crucial factors for user satisfaction. User comments on app stores can be leveraged to assess UI/UX quality through sentiment analysis. This study aims to evaluate the comparative performance of the Support Vector Machine (SVM) and Logistic Regression algorithms in classifying user sentiment regarding app UI/UX. Research data was obtained from the Sentiment Analysis Dataset—App Reviews available on Kaggle. The research process included data cleaning, feature extraction using the Term Frequency–Inverse Document Frequency (TF-IDF) method, model training, and model testing with an 80:20 training-to-test data split. Performance was measured using accuracy, precision, recall, and F1-score metrics. The findings show that SVM yields higher results compared to Logistic Regression, with an accuracy of 88.26%, precision of 91.13%, recall of 86.93%, and an F1-score of 88.98%. Meanwhile, Logistic Regression achieved an accuracy of 86.75% and an F1-score of 87.54%. These findings indicate that SVM is more effective for classifying user sentiment regarding the UI/UX of text-based applications.