Ransomware Detection on Android Devices Using Machine Learning

In recent years, the rapid increase in Android-based ransomware attacks has posed serious threats to mobile security and user data privacy. Detecting such malicious behavior through efficient network traffic analysis has become a critical area of research. This paper presents a hybrid approach for network traffic classification using Artificial Bee Colony (ABC) optimization for feature selection and Random Forest as the classification algorithm. The primary goal is to improve detection accuracy by eliminating redundant or irrelevant features, thereby reducing the complexity of the model. The dataset used comprises labeled Android network traffic data, which includes both benign and ransomware samples. After preprocessing the data, ABC is applied to identify the most significant features. These selected features are then used to train a Random Forest classifier. The model is evaluated using metrics such as accuracy, precision, recall, and F1-score. Additionally, a user-friendly interface is developed using Gradio, enabling real-time prediction and enhancing usability. Experimental results show that our proposed model achieves high classification performance while significantly reducing the feature set, making it suitable for real-time threat detection scenarios. Keywords: Artificial Bee Colony (ABC), Feature Selection, Random Forest, Android Ransomware, Network Traffic Classification, Gradio, Machine Learning