Sentiment Analysis of Customers on Utilizing Online Motorcycle Taxi Service at Twitter with the Support Vector Machine
Jajam Haerul Jaman, Rasdi Abdulrohman · 2019
Online motorcycle taxi is one of the latest public transportation trends to Indonesian people. Although it is still new, its existence is able to change the behavior of Indonesian people. Some people feel that they are being helped by its existence, yet some people do not. Various types of complaints or appreciation about the online motorcycle taxi are frequently discussed on popular social media, namely; twitter. Sentiment Analysis is a part of science in text mining that is able to predict someone's feeling or feeling as outlined in text. This study uses the support of vector machine method with TF-IDF feature selection. The data set were 1183 that were taken from twitter with the keyword gojekindonesia and grabID through the data crawling process. The data set are divided into 3 classes: positive sentiment class, negative sentiment class and neutral sentiment class. The classification process was done by using five scenarios that were training comparison and Testing 50:50, 60:40, 70:30, 80:20 and 90:10. In addition, the classification process also used four kernels such as linear, rbf, sigmoid and polynomial. The highest accuracy results are in the comparison of 90% as training data and 10% as testing data while using a linear and sigmoid kernel are 0.8060 or about 80%.