Aspect-Based Sentiment Analysis Mengenai Tingkat Kepuasan Pelanggan terhadap Ekspedisi Pengiriman Barang Menggunakan Random Forest

Abstract

Today‘s trend on online shopping through various e-commerce applications has had a positive impact on the develompent of business in Expedition Services in Indonesia. The existence of various brands of goods delivery services allows people to have options in choosing any brand based on the preferences and needs of their own. As a form of brand competition amidst the increasingly specific needs and demands of society, companies need not merely brand awareness but also a good brand reputation. Assessment of a brand‘s reputation can be carried out based on user reviews through various platforms, especially social media Twitter, which is oftenly used by the public as a medium for Ekspresing opinion. This researchs was developed to obtain sentiment conclusions about an expedition service in Indonesia, namely SiCepat Ekspres, which is assessed based on various aspects, namely responsiveness, cost, goods condition, delivery, and time precision. The system was built using the Random Forest Classifier approach, where the data was tested based on the iteration of the Decision Tree method (tree_number). The results obtained through this sentiment analysis system are negative for the time precision aspect, while for other aspects mentioned before are in neutral. The average accuracy for the overall sentiment analysis results obtained is 90.8%.