Android Malware Detection Using Genetic Algorithm Based Optimized Feature Selection and Machine Learning

Android has the biggest worldwide market share owing to its open-source nature and Google’s support. Since it is the most broadly utilized working framework in the world, it has attracted the attention of cyber criminals who use it to spread malware. Using an evolving genetic algorithm for feature selection, the researchers developed an Android malware detection machine-learning approach that relies on machine learning. Machine-learning classifiers are trained using chosen features from the Genetic algorithm, and their ability to recognize malware is assessed when include choice. According to the trials, the Genetic algorithm gives the most efficient feature subset, enabling the feature dimension to be decreased by half from the original feature set. After feature selection, machine learning–based classifiers retain a classification accuracy of more than 94 percent despite operating on a significantly smaller feature dimension, reducing computing cost of learning classifiers.