The comparation of text mining with Naive Bayes classifier, nearest neighbor, and decision tree to detect Indonesian swear words on Twitter

摘要

Twitter is one of world most famous social media. There are many statement expresed in Twitter like happiness, sadness, public information, etc. Unfortunately, people may got angry to each other and write it down as a tweet on Twitter. Some tweet may contain Indonesian swear words. It's serious problem because many Indonesians may not tolerated swear words. Some Indonesian swear words may have multiple means, not always an Indonesian swear word means insulting. Twitter has provide tweet's data by account, trending topics, and advance keyword. This work try to analyze many tweet about political news, political event, and some Indonesian famous person because the tweet assumed contains many Indonesian swear word. The derived tweets will process in text mining and then analyzed by classification process using Naive Bayes Classifier, Nearest Neighbor, and Decision Tree to detect Indonesian swear word. This work expected to discover the high accurate classification model. It means, the model can differentiate the real meaning of Indonesian swear word contained in tweet.