Performance Comparison of Support Vector Machine and Random Forest Using Delta TF-IDF on Sentiment Classification of Tiket.com Reviews

Tiket.com is one of the largest Online Travel Agent (OTA) platforms in Indonesia. Tiket.com as a pioneer of online travel agents in indonesia won an achievement during the pandemic, namely won the 2022 Top Brand Award. Tiket.com application can be downloaded on Google Play Store. Google Play Store has a review feature, which can be an important aspect for consumer decisions. Sentiment analysis can be applied to classify positive and negative reviews, which can help companies review their services. This study aims to analyze the sentiment of customer reviews on the Tiket.com app using Support Vector Machine (SVM) and Random Forest (RF) algorithms with Delta TF-IDF term weighting. The dataset used in this study consists of 1500 most relevant reviews that were scraped from the Google Play Store. The best model for SVM and Random Forest is selected by tuning hyperparameters using Gridsearchcv. The results of this study indicate that SVM produces a better classification performance on 94.33% Accuracy, 87% Recall, 95.6% Precision, 91.1% F1-Score and 86.96% Kappa-Statistic.