User Sentiment Analysis of Online Transportation Platforms Using K-Means and K-Nearest Neighbor

Online transportation has been one of the most used type of applications in this day and age, dominated by the platforms Go-Jek and Grab. In order to improve their service and user satisfaction, this research aims to determine the user sentiment analysis on both application using the K-Nearest Neighbor (KNN) method, a supervised learning algorithm used to classify data according to their nearest neighbors, and the K-Means method, an unsupervised learning algorithm used to group data into clusters according to their similarities. Both methods will be implemented using RapidMiner. The data used are 199 usernames and ratings on both applications, collected from Kaggle. The KNN method resulted in a 100 percent accuracy and the K-Means method proved that the cluster with the most items is the one which had the ratings above 4.5. This proves that the user sentiment analysis of online transportation platforms, such as Go-Jek and Grab, using K-Nearest Neighbor (KNN) and K-Means reveals that the sentiments are, on average, positive sentiments.