Ransomware Detection Using Machine Learning Algorithms

Abstract

This research paper delves into the realm of cybersecurity, exploring the synergy between machine learning algorithms and malware detection. The primary objective is to enhance the detection of ransomware cyber attacks through a meticulous analysis of binary file attributes. Leveraging a diverse dataset comprising legitimate binaries and malware files, we employ a comprehensive methodology that spans dataset selection, exploration, segmentation, visualization, cleaning, feature selection, and model building. The study uses three machine learning models; Random Forest Classifier, Logistic Regression, and a Neural Network to analyze and categorize binary files. The Random Forest Classifier shows strong performance with the highest accuracy, indicating its effectiveness with tangled datasets. Logistic Regression, acting as a baseline, demonstrates moderate effectiveness, offering insights into the dataset's complexity. The Neural Network, built with TensorFlow and Keras showcased its ability to understand complicated patterns.the paper provides valuable insights into the capacities of different machine learning models for malware detection, emphasizing the critical role of model selection in addressing the challenges posed by ransomware cyberattacks.