AN ANALYSIS OF BAYESIAN ALGORITHM IN PREDICTING THE INTERESTS OF HIGH SCHOOL GRADUATES TO STUDY FURTHER INTO UNIVERSITY USING SOCIAL MEDIA TWITTER

Abstract

Sentiment analysis is a very necessary feature in collecting information about the user preferences about a topic on social media. Analyzing the user sentiments is important to find out negative comments and positive comments. This study aims to classify tweet data into 2 sentiments namely positive and negative. In this study, we collect the Indonesian texts dataset from big data of Twitter social media. The collected dataset from tweeter is called as tweets bag of word is used in this study to predict the student preferences about university study. After extracting the datasets, we get 1978 example of tweets data containing both positive and negative opinions which will be classified using a text mining approach of Naïve Bayes algorithm. Before classification, several stages of text processing are carried out such as case folding, normalization, tokenization and stopwords removal. There are 113 negative tweets, 1744 neutral tweets and 116positive tweets. We also test the accuracy of the Naïve Bayes algorithm and get 85% accuracy rate. We implied that the tweet analysis can be used by university decision maker as information for decision maker to build strategy and interest of high school graduates to continue studying to university.  
 
Keywords: Sentiment analysis, Text Mining, Naïve Bayes, classification, twitter.