Sentiment Analysis og Goride Services on Twitter Social Media Using Naive Bayes Algorithm

Abstract

Online motorcycle taxi is an application-based transportation technology innovation. Online motorcycles offer relatively low prices and offer discount features. However, the existence of online motorcycles creates congestion problems and conflicts between conventional transports. Various speculations arose in the midst of the public against Goride. So it makes the public opine and wants to judge an object openly through social media, one of which is Twitter. An opinion given by society is a textual opinion that can be analyzed. Sentiment analysis is used to detect opinions in the form of a person's judgment, evaluation, attitude, and emotion. The textual classification algorithm used in this study was Naive Bayes. This research aims to find out the public sentiment towards Goride's service as an online motorcycle taxi in positive and negative categories and to find out the accuracy results of the Naive Bayes algorithm against Goride's service. The data used in this study are secondary data. Data obtained by crawling using an API provided by Twitter developer. Analysis techniques are performed by text preprodeing, data labelling, word weighting, classification, then performance evaluation of classification. The results of the positive category sentiment classification are 698 data, while the negative category sentiment is 517 data. Which means more positive sentiment than negative sentiment. The accuracy performance of the Naive Bayes algorithm results in an accuracy rate of 77.78%.