Paper Title
Comparison of Maximum Entropy and Support Vector Machine Methods for Sentiment Analysis of Pertalite Product through Twitter Social Network

Abstract
Support Vector Machine (SVM) and Maximum Entropy (Maxent) are a method of machine learning that can be used in text data classification. The difference between the two methods lies in its performance in defining certain categories or classes in the text. SVM works by finding the best hyperplane to separate the two data classes. Maxent works by finding the maximum entropy value. From these differences, the researchers wanted to compare the accuracy level of the classification of the two methods by using case study of 1411 tweets taken from Twitter social networking about Pertalite in May, June, and July 2017. From this research it can be seen that there is a change of performance from machine Learning during the months of May, June, and July 2017. This can be seen from the changes in accuracy rates generated by machine learning in each month. In May 2017, machine learning using the Maxent method gave a lower accuracy level than the SVM method with the difference of the accuracy level of both methods of 4,84%. While in June and July 2017, the Maxent method has a higher accuracy rate than SVM with a difference of 2,25% accuracy rate for both methods in June 2017 and 2,99% in July 2017. Keywords - Machine Learning, Support Vector Machine, Maximum Entropy, Twitter, Pertalite, Accuracy.