Paper Key : IRJ************514
Author: Gurpreet Kaur
Date Published: 02 Apr 2024
Abstract
People widely use social media applications to spend free time with their friends and relatives over long distances, and users generate vast amounts of data on social media applications where abusive speech has become a significant issue, so it is essential to identify the solution to hate speech detection. Researchers regularly work hard to build models that automatically detect hate speech context. In this research, we focus on abusive words posted on Twitter comment sessions by users. First, we will download the two datasets from Kaggle, cyber-bullying and Twitter hate speech. Then, we will divide this dataset into training and testing and combine both datasets. Further, we performed data preprocessing and feature extraction. Moreover, we have also used different machine learning techniques such as logistic regression, KNN, nave Byes, decision tree, and random forest for detecting abusive speech. Algorithms were shown different accuracies. We got the best accuracy from a random forest. It was 98.93%, rather than other algorithms. This research work will go ahead if we change the parameters of the datasets
DOI Requested