Twitter User Sentiment Analysis Of TIX ID Applications Using Support Vector Machine Algorithm

Abstract

The cinema is a place to watch movies using the big screen. Various comments on the TIX ID application service can be used to reference the company's evaluation material in assessing the level of service quality that has been provided, so that later the TIX ID application can be used optimally by application users and the company, as for the purpose of this study to find out responses and find out the sentiment analysis stage on twitter social media using the vector machine support algorithm for the TIX ID application. This algorithm is commonly used for text mining by going through the data collection stage, cleaning and labelling data stage, training and testing data sharing stage with 3 comparison scenarios, namely 70:30, 80:20, and 90:10 using 3 kernels, namely dot, radial, and polynomial, then through the text preprocessing stage, the TF-IDF word weighting stage, the data modeling stage, and the evaluation stage. The preprocessing stage consists of transform case, tokenize, and stopwords filters. The result of this study is that the support vector machine algorithm has an accuracy value of 74.17%. The research concludes that the support vector machine algorithm with a ratio of 80:20 training and testing data ratio scenario produces the highest accuracy.