A Hyper-Heterogeneous Graph Framework for Detecting Malicious Twitter Bots

Abstract

The rapid growth of social media platforms has significantly increased the presence of malicious Twitter bots that spread spam, fake news, phishing links, misinformation, and harmful automated activities, creating major security and trust issues in online communication systems. To address this challenge, this project titled “A Hyper-Heterogeneous Graph Framework for Detecting Malicious Twitter Bots” presents an intelligent machine learning-based framework for identifying malicious bot accounts from Twitter tweet data using advanced text analysis and classification techniques. The proposed system utilizes a dataset containing Twitter tweet information and performs preprocessing operations such as null value handling, text cleaning, and label encoding to prepare the data for efficient model training. TF-IDF (Term Frequency-Inverse Document Frequency) feature extraction is applied to convert textual tweet content into meaningful numerical vectors for machine learning analysis. Multiple classification algorithms including Logistic Regression, Random Forest, Decision Tree, and Naive Bayes are trained and evaluated to identify the most accurate prediction model.