Opinion mining on Twitter microblogging using Support Vector Machine: Public opinion about State Islamic University of Bandung

Abstract

Tweet data on Twitter as microblogging can be processed to be an important and useful information. We propose opinion mining with Support Vector Machine (SVM) algorithm to classify tweet opinion data which is a huge data. This opinion mining will be used to get insight of public opinion about State Islamic University of Sunan Gunung Djati Bandung which is one of large university in Indonesia. We have two classes for opinion classification that is negative and positive opinions. Pre-processing phase before classifying consists of cleaning data, emotion tokenizing, case folding, stop words removal, and stemming process. The result of this research is 0.838 precision value and 0.76 recall for positive class. Then, 0.78 precision value and 0.853 recall for negative class. Opinion classification with SVM of this research has accuracy 78.75%.