A Malware Detection Method for Health Sensor Data Based on Machine Learning and Genetic Algorithm
Main Article Content
Abstract
Small modifications in the virus code are easily detected by conventional signature-based malware detection techniques. The majority of malware programmes today are modifications of other programmes. They therefore have various signatures yet share certain similar patterns. Instead than just noticing slight changes, it's important to recognise the virus pattern in order to protect sensor data. However, we suggest a quick detection strategy to find patterns in the code using machine learning-based approaches in order to quickly discover these health sensor data in malware programmes. To evaluate the code using health sensor data, XGBoost, LightGBM, and Random Forests will be specifically utilised. The codes are either supplied into them as single bytes or tokens or as sequences of bytes or tokens (e.g. 1-, 2-, 3-, or 4- grams). Terabytes of labelled programmes, both virus and benign ones, have been gathered. Choosing and obtaining the features, modifying the three models to train and test the dataset, which comprises of health sensor data, and evaluating the features and models are the challenges of this assignment. When a malware programme is discovered by one model, its pattern is broadcast to the other models, effectively thwarting the infiltration of the malware programme.