Twitter Sentiment Analysis in Indonesian Language using Naive Bayes Classification Method

Social media is a very large source of data and has continuous and very short changes, the time span of which changes also has considerable relevance for various fields. The purpose of this study as a sentiment analysis aimed at twitter status discussing daily topics to determine the attitude of opinion from tweets posted on a trending topic online using the Naive Bayes Classifier method. With various series of steps carried out, namely data collection from Twitter, preprocessing, classification process with the Naives Bayes Classifier algorithm, and evaluation of the results determined by this method. The results at the time of testing, namely in the form of the highest accuracy value obtained in testing 40 tweet data by 82% with a time of 19.92 seconds, for the highest precision value in testing 10, 20, and 50 tweet data by 100% with a time of 14, 85 seconds, for the highest recall value in testing 40 tweet data of 81.82% with a time of 16.09 seconds, for the highest F1-Score value in testing 40 tweet data of 88.52% with a time of 16.38 seconds.