Applying Machine Learning Techniques for Religious Extremism Detection on Online User Contents

In this research paper, we propose a corpus for the task of detecting religious extremism in social networks and open sources compare various machine learning algorithms binary classification problem using previously created corpus, thereby checking whether it is possible to detect extremist messages Kazakh language. To do this, authors trained models six classic machine-learning such as Support Vector Machine, Decision Tree, Random Forest, K Nearest Neighbors, Naive Bayes, Logistic Regression. increase accuracy texts, used characteristics Statistical Features, TF-IDF, POS, LIWC, applied oversampling undersampling techniques handle imbalanced data. As result, achieved 98% texts collected dataset. Testing developed databases that are often found everyday life “Jokes”, “News”, “Toxic content”, “Spam”, “Advertising” has also shown high rates detection.