Detecting Racist and Bad Words Using Text Mining in Social Media

The use of social media has grown rapidly in the last decades. Many of these social media platforms, i.e., Twitter, do not screen or remove potentially offensive/ harmful tweets. It can cause discomfort for the users, especially if minors encounter these words while they are surfing the media. We came up with a profanity detection model with text mining that uses bag of words algorithm. Then we applied four different classifiers to see which one gets the highest accuracy score. The results show a high accuracy performance, with Random Forest classification being the highest with 97.5%.