LSTM based Malware Detection Framework for Android Devices
Prof. Kanchan Umavane, Mr. Prathamesh Avhad, Mr. Rutik Tavanoji, Ms. Rupali Pardeshi · Zenodo (CERN European Organization for Nuclear Research) · 2023
Lately, the matter of dangerous malware in devices is spreading speedily, particularly the repackaged android malware. Though understanding Android malware detection via dynamic analysis will give a comprehensive read, and need to also relate between the app’s features and the features that are needed to deliver its category’s functionality. The most popular freely accessible smartphone operating system is Android, yet it lacks virus detection in its permission declaration and access control systems. The matter of investigating such malware presents distinctive challenges thanks to the restricted resources accessible and restricted privileges granted to the user. But, each APK file also offers unique possibilities for the necessary information connected to each programme. By integrating permissions and API callas characteristics to describe malware, as well as using machine learning methods to automatically extract patterns to distinguish between benign and malicious Apps, and objective is to create a solution that effectively reduces the danger of Android malware. This model provide a machine learning-based method for malware detection on Android-powered devices. This model uses an intuitive user interface to efficiently identify, detect, classify, and protect Android mobile devices against harmful applications, preventing any data theft or misuse. This research aims to develop an LSTM rule-based malware detection system based on code behaviour signatures. It will be able to identify malicious code and its variations successfully in runtime and expand malware characteristics data dynamically. For static features and dynamic actions, machine learning techniques are the current approaches to Android malware patterns, and testing results show that the methodology combines a high detection rate and a low rate of false positives and false negatives. Recurrent neural networks are adjusted to create Long Short-Term Memory (LSTM) networks, which facilitate better memory retention for prior information.