Malware detection using ensemble method

Ahmad, Amizah Aida (2017) Malware detection using ensemble method. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Malware detection using ensemble method.pdf - Submitted Version Download (693kB)
	Text (Full Text) Malware detection using ensemble method.pdf - Submitted Version Restricted to Registered users only Download (1MB)

Abstract

In today’s technology driven world, the increasing of malware in the cybercriminals that exploiting the internet and always create and distribute harmful malware has become a serious threat. Malware significantly impact computer’s performance and often go unnoticed in our systems and causes several problems to the user. Hence, It’s imperative to take the precautions necessary to detect and prevent malware infections. One of the way to detect malware detection is by using machine learning techniques. Malware detection is detected by looking at its behavioural. Behavioural malware detection is a field where malware is detected by its behaviour and the machine learning will look at the pattern of the behavioural. Then it will be analyzed and a report will be generate from the data. Thus, in this project, the behavioral of malware is analyzed and ensemble method is applied in detecting malware. Firstly, the data is collected by a multiple categories of system log and parser chooses from application. Then from the dataset it will classify it to 5 type of n-gram. Secondly, the best features from each of the n-gram are extracted using three feature selection techniques, namely Information Gain, Symmetrical Uncertainty and Chi-Square. SVM classifier is used to train the feature vectors and create a model for each n-gram. Finally, every model from 1-gram to 5-gram is combined using ensemble method. The significant contribution of this project is the effectiveness and efficiently of malware prediction using the state-of-the art techniques named ensemble method.

Actions (login required)

View Item