Chi-square statistical and support vector machine for android malware detection

Johari, Muhammad Zulhilmi (2017) Chi-square statistical and support vector machine for android malware detection. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Chi-square statistical and support vector machine for android malware detection.pdf - Submitted Version Download (1MB)

Abstract

This project focus on feature selection and classification approach for android malware. Feature selection step is reducing the available features to a set that is optimal or sub-optimal and capable of producing results which are equal or better to that of the original set. This technique is used for descriptors and allows survey response to be put in into meaningful categories in order to have the useful data. The classification of malware based on behaviour of different malware is done by using data mining techniques. Classifier is a supervised learning and requires training data accompanied by labelling the class and the new data is classified based on the training data set. So, this project is significantly to propose combinational method for better android malware detection. The problem statements for this project is there are too many irrelevant Features that make hard to detect android malware behaviour and also difficult to differentiate malware activity more precisely. So, the objective of this project is to obtain android malware activity more accurately using Support Vector Machine and propose Chi-square as Feature Selection method to identify relevant features accurately. In addition, the best possible accuracy and detection rate can be achieved by using feature selection method which is Chi-square. Throughout the experiment, accuracy and detection rate for android malware activity improved by applying Chi Square and Support Vector Machine.

Actions (login required)

View Item