Public Opinion Sentiment Analysis on Train Transport in Jakarta Using a Hybrid Model Machine Learning

Transportation is a key element in smoothing the wheels of the economy and connecting various regions, especially in big cities like Jakarta which has a high population density. This leads to dense and complex traffic conditions. Improving the quality and facilities of public transportation is important to overcome these problems. However, people are still reluctant to use public transportation for various reasons. Therefore, it is important to understand public sentiment towards public transportation in Jakarta. This research focuses on sentiment analysis of train-based transportation, namely KRL, MRT, and LRT. Sentiment analysis is conducted using a hybrid learning model with a voting model method, which combines SVM, logistic regression, and CNN algorithms. The data used is labeled with InSet sentiment dictionary and extracted features using TF-IDF method. The modeling results show that this hybrid model produces 89% accuracy for the KRL dataset, 88% for the MRT dataset, and 81% for the LRT dataset. However, this model still has difficulty in predicting neutral and positive classes. The results of this study show that hybrid learning with the voting model method can provide quite good results in public transportation sentiment analysis, but there is still room for improvement in the classification of neutral and positive sentiments. The findings provide important insights for the development of strategies to improve the quality of public transportation and encourage people to use the service more.