IRO Journals

Journal of Information Technology and Digital World
Home / Archives / Volume-3 / Issue-1 / Article-5

Volume - 3 | Issue - 1 | march 2021

Design of Associate Content Based Classifier for Malicious URL Prediction by Rule Generation Algorithm
Vivekanadam Balasubramaniam  149  93
Pages: 44-56
Cite this article
Balasubramaniam, V. (2021). Design of Associate Content Based Classifier for Malicious URL Prediction by Rule Generation Algorithm. Journal of Information Technology and Digital World, 3(1), 44-56. doi:10.36548/jitdw.2021.1.005
Published
18 May, 2021
Abstract

Recently, the internet is becoming as the most effective tool to interact with many foreign societies especially during COVID-19 pandemic. Moreover, the digital platform is increasing in many developing countries and at the same time, the chance of fraudulence is also increasing day by day. In the digital world, phishing assaults are emerging as the most common type of social engineering attack. Currently, many websites are targeting to acquire the confidential data, which is stored in websites. Recently, the classification techniques are employed to detect the phishing websites. Many tools are used for anti-phishing purposes; they are blacklist and antivirus software. The confidential data in a fake surrounding has intended the category of leaked data due to the action of attackers. In this scenario, machine learning method is observed as a very effective to classify the phishing and non-phishing web (Uniform Resource Locator) URLs. This classification struggles in classifying the leaked data content-based challenge. Therefore, the proposed algorithm is associated with the content-based classification method along with the rule-based generator algorithm. This research article integrates the content-based classification with a rule-based generator algorithm to improve the overall performance of the system. The updated public online repository called Mendeley dataset is used in the proposed research work. The proposed algorithm is used in 7k phishing and real websites content data for performing feature extraction. The extracted feature is then analyzed with our proposed algorithm to provide better prediction accuracy. Also, the proposed work has concluded that, the associate algorithm has achieved better accuracy, when compared to other existing methods.

Keywords

Deep learning malicious website

Full Article PDF Download Article PDF