Sentiment Analysis of Negative Comments (Hate Speech) on Twitter with K-means Clustering Algorithm Using RapidMiner

Abstract

This study discusses sentiment analysis of negative comments (hate speech) on the social media platform Twitter by applying the K-Means Clustering algorithm using RapidMiner software. In today’s digital era, Twitter has become one of the main platforms for the open dissemination of public opinion, including negative comments that may lead to hate speech. To understand the sentiment patterns in these comments, clustering was carried out on a dataset consisting of 27,325 tweets obtained from Kaggle. The research stages included data collection, preprocessing, and the implementation of the K-Means algorithm with three clusters, categorizing the comments into negative, neutral, and positive groups. The results showed that most of the comments fell into the negative cluster, comprising 14,032 entries, followed by 9,924 neutral entries and 3,369 positive entries. These findings demonstrate that the K-Means algorithm is effective in identifying the distribution of hate speech on social media and provides valuable insights for mitigating and monitoring negative content automatically. This study is expected to serve as a foundation for developing more accurate and adaptive sentiment analysis systems in response to the dynamics of digital communication.