Optimal Features Based Phishing URL Detection Using Decision Tree - Learning Approach for Web Applications

  • Mr. M. Sathish Kumar, Dr. B.Indrani

Abstract

In the field of phishing site identification, numerous methodologies are considered to recognize the phishing URL, for example, the blacklist technique and some machine learning models. Yet, it's having some impediment in the training and testing process. To defeat issues, we will build up the inventive model to distinguish the fake or malicious or phishing URL site. Generally, the databases are acquired is data slicing, and afterward the for the most part entire database partitioned into two sections (i) Training and (ii) Testing. Our proposed, fake detection process extricate the essential features from chosen training database, the feature like using IP address, request URL, page setup, login details, lexical features (like URL length, prefix/suffix so on.), and some different features. From that features, optimal features is selected based on Discrete Bat Algorithm (DBA). These optimal features can be easily detecting the fake URL sites by performing the fitness as similarity measure. When the optimal features are getting from the trained URL site, the classification techniques Decision Tree (DT) is utilized to detect the considered site is phishing, non-phishing or suspicious. Performance measures of the proposed work is analyzed such as confusion matrix based precision, recall and F-measure

Published
2020-06-01
How to Cite
Mr. M. Sathish Kumar, Dr. B.Indrani. (2020). Optimal Features Based Phishing URL Detection Using Decision Tree - Learning Approach for Web Applications. International Journal of Advanced Science and Technology, 29(7s), 4527 - 4544. Retrieved from https://sersc.org/journals/index.php/IJAST/article/view/25691