Detection of Android Malware using Sensitive APIS’s

Sajikumar, Vivek (2021) Detection of Android Malware using Sensitive APIS’s. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (501kB) | Preview
Preview	PDF (Configuration manual) Download (464kB) | Preview

Abstract

In this modern world filled with technologies, the usage and growth of android apps has gone up in the previous years. In a recent survey conducted it was noted that out of 100 randomly selected people, 87 of them used mobile devices based on android. This shows that android users are of 87%. As the number of users increases the chances of being targeted for malicious activities are also high. A survey states that an approximate of 38,000 malwares are being created by malicious users to exploit android devices. It is high time that, a measure to be implemented to overcome the malwares. As technologies are being improved on the daily basis and by using reverse engineering methods, malicious users are generating malwares which are extremely difficult to detect. Malwares in a device can be broadly classified as code based malware and behaviour based malware.

This paper proposes a malware detection method which is based on behaviour. This method is considered to be effective and agile. The proposed method is based on sensitive API. There are two phases in this method, which are training phase and testing phase. The training phase is responsible for collecting the API of multiple apps in a eigenvector format. This collected data are then used later on to find out the connection between API calls and malware, from which new combination of API are generated. Performance of each algorithm are tested and the best 3 are selected using the help of K-fold cross validation method. The selected models are then embedded to ensemble learning model for the evaluation of performance. Performance of the model is identified by its working efficiency and its outputs accuracy. In this paper, the algorithms used are Neighbour’s, SVM Classifier, Decision Tree, Naïve Bayes, Random Forest, Linear Discriminant Analysis and the top 3 performing algorithm along with ensemble learning model gave out an accuracy rate of 96%.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security Q Science > QA Mathematics > Computer software > Mobile Phone Applications T Technology > T Technology (General) > Information Technology > Computer software > Mobile Phone Applications
Divisions:	School of Computing > Master of Science in Cyber Security
Depositing User:	Tamara Malone
Date Deposited:	29 Dec 2022 15:03
Last Modified:	29 Dec 2022 15:03
URI:	https://norma.ncirl.ie/id/eprint/6051

Actions (login required)

View Item