Android Malware Detection Using Permission-Based Static Analysis and Machine Learning

The Rapid Growth Of Android Applications Has Significantly Increased The Risk Of Malicious Software Targeting Mobile Users. Traditional Signature-Based Detection Methods Fail To Identify Newly Emerging And Obfuscated Malware. This Paper Presents A Permission-Based Static Analysis Approach For An- Droid Malware Detection Using Machine Learning Techniques. The System Extracts Declared Permissions From Apk Files Using The Androguard Framework And Converts Them Into Binary Feature Vectors. A Random Forest Classifier Is Trained Using An 80–20 Train-Test Split To Classify Applications As Benign Or Malicious. Performance Metrics Including Accuracy, Precision, Recall, F1- Score, And Confusion Matrix Are Evaluated. The System Also Includes A Heuristic Fallback Mechanism Based On Dangerous Permission Thresholds. Experimental Results Demonstrate That Permission-Based Static Analysis Combined With Machine Learning Provides An Efficient And Scalable Approach For Preliminary Android Malware Detection.