Detection of Android Adwares by using Machine Learning Algorithms
Dinesh C Dobhal1, Purushottam Das2, Kiran Aswal3
1Dinesh C Dobhal, Department of CS&E , Graphic Era University, Dehradun, India.
2Purushottam Das, Department of CS&E , Graphic Era Hill University, Dehradun, India.
3Kiran Aswal, Department of CS&E , Graphic Era Hill University, Dehradun, India.
Manuscript received on 15 June 2019 | Revised Manuscript received on 25 June 2019 | Manuscript Published on 02 July 2019 | PP: 17-21 | Volume-8 Issue-4S, April 2019 | Retrieval Number: D10050484S19/19©BEIESP | DOI: 10.35940/ijeat.D1005.0484S19
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: With the rapid growth in technology, which is improving every day and becoming more pervasive, smartphone users also are increasing. These intelligent devices and gadgets are now being used in automated vehicles, IoT enabled industries, surveillance, education, entertainment, etc. Android, Linux kernel based mobile operating system, with its largest market share is now being used in almost every device that is capable to do some computation. These devices may include smartwatches, digital cameras, smart glasses, smart mirrors, Home Automation System (HAS), Internet of Things (IoT), Internet of Vehicles (IoV), and many more. Parallel to these developments, the android based systems also are being targeted by the cyber attacker by developing more advanced malwares. Such attacks may harm to the system as well as human life. The android malwares are evolving day by day and are capable to escape the traditional security solutions. Therefore, security is the primary issue of android based system, which requires to be re-investigated. In this paper, we analyze the pertinence of machine learning based solutions to detect android malware, particularly Adware. Logistic Regression (LR), Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Classification And Regression Trees (CART), and Naive Bayes (NB) machine-learning algorithms are trained and tested for two scenarios. Scenario A for binary classification and Scenario B is for multi-class classification. The 60% of the dataset is used to train the ML algorithms and the remaining 40% is reserved for the testing. The algorithms are evaluated by using 10-fold cross-validation method.
Keywords: Network Security, Adware, Malware, Anomaly detection, Machine Learning (ML).
Scope of the Article: Machine Learning