A New Machine Learning Method for Identifying Android Malware by Using Features That Co-Exist

Abstract: The mission investigates Android malware detection utilising various datasets, along with Drebin,Malgenome, and CIC_MALDROID2020. Those datasets offer a massive repository of API and Permission statistics, for thorough examineDiversegadget studying fashions had been hired for classification, inclusive of Logistic Regression,“Support Vector Machine, k-Nearest Neighbors (KNN), Random Forest, Decision Tree”, and a Stacking Classifier that merges Random Forest and Multi-Layer Perceptron (MLP) with LightGBM. The mission entails comprehensive dataset preprocessing, model education, and stringent performance assessment. This thorough technique seeks to create very efficient models for the detection of Android malware. This assignment advances cellular security through turning in successful version outcomes, hence enhancing our capability to hit upon and neutralize Android malware attacks efficiently. The venture's are sizable for mobile security professionals and researchers. This paintings offers a Stacking Classifier that integrates the blessings of Random forest, "Multi-Layer Perceptron (MLP)", and LightGBM models to enhance function extraction. This ensemble learning method markedly enhances predictive accuracy. furthermore, we created an intuitive Flask framework linked with SQLite for secure registration, login, and consumer checking out. This optimized device helps input submission and prediction acquisition, improving the project's robustness and practicality for powerful person engagement.“Index terms - co-existence, FP-growth, machinelearning, malware”.