Mobile malware detecton using RNN-LSTM through opcode

Azman, Ahmad Razin (2021) Mobile malware detecton using RNN-LSTM through opcode. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Mobile malware detecton using RNN-LSTM through opcode.pdf - Submitted Version Download (726kB)
	Text (Full Text) Mobile malware detecton using RNN-LSTM through opcode.pdf - Submitted Version Restricted to Repository staff only Download (3MB)

Abstract

The popularity of Android and the development of third-party app stores have led to Android malware growing in recent years. Emerging Android malware families are progressively implementing advanced detection avoidance tactics, necessitating more effective Android malware detection methodologies. Hence, in this project analyse an opcode features-based framework to identifying and categorizing Android malware using RNN-LSTM. This method allows for automatic feature discovery without the need for previous expert or subject knowledge for pre-defined features. Identify mobile malware using opcode is the aim of this paper. Not only that, in this research create and analyse RNN-LSTM models for mobile malware detection through opcode. In this research, synthesis all material that have related to the mobile malware detection from any journal. Summarizing the material, analyse, interpret and make implications for researcher in order to properly draw a conclusion to provide a solution. After that, in this project experimental setup is required, providing an isolation environment to prevent malware harmful PC host. The activities in the isolation environment for doing static analysis on malware samples using the jadx-gui tool involve Android package extraction and code disassembly. In this isolation environment, 1000 malware samples and 1000 benign samples will use the most recent version of Python to extract opcode. Google Colaboratory RNN-LSTM design is the way to train all data sets (80% training 20% testing). This research aims to analyse and evaluate the output of a dataset to obtain the value of the True Positive rates (TPR) and False Positive Rates (FPR).

Actions (login required)

View Item