Artificial Neural Network Based Predictive Model for Android Malware Detection Using a Sine Cosine Algorithm for Feature Engineering

The aim of this research is to accurately classify applications running in the Android environment as either malware or legitimate software using machine learning techniques. This accuracy is achieved through a well-structured machine learning pipeline (MLP), which includes feature engineering involving both feature selection and feature extraction methods. This approach significantly aids in dimensionality reduction, as the novel dataset used for Android malware classification, known as NATICUSdroid, contains an extensive number of features. The dataset includes both native and custom permissions as features. Therefore, the research proposes designing an MLP that focuses on the most representative features to achieve high accuracy with a minimal number of features. Classification is performed using an artificial neural network, with feature engineering carried out through Singular Value Decomposition and Sine Cosine Algorithm. The performance of the predictive model is evaluated using metrics such as accuracy, precision, recall, and F1 Score.