Detection of Phishing Sites Using Machine Learning

Detection of Phishing Sites Using Machine Learning

Pooja Maroti Jadhav1, Tejashree Devendra More2, Vaishnavi Hemant Bagul3, Prof. S.V. Shardul4

*1,2,3 Department of Computer Engineering, Shatabdi Institute of Engineering and Research, Nashik, Maharashtra, India.

*4 Professor, Department of Computer Engineering, Shatabdi Institute of Engineering and Research, Nashik, Maharashtra, India.

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - The content of phishing websites and online-based material provides a variety of indicators. One of the criminal attacks that are successful in the online world is phishing sites that direct users to a phishing website that looks and acts like a legitimate website in order to steal the victim's personal and sensitive information. The Extreme Learning Machine-based version that was suggested was excellent at spotting phishing websites. Interne page types vary greatly in terms of their characteristics. So, in order to protect against any phishing attack, we must use a set of web page features. To defend against these threats, a machine learning strategy is used.

The phishing dataset that is planned to be imported, authentic URLs from the database, and also the data that is collected are all pre-processed. Four groups of URL characteristics are used to detect phishing websites: domain, address, anomalous based, HTML, and JavaScript features. Data that has been analyzed is used to extract URL characteristics and produce URL attribute values. ML approaches are used to analyze URLs and determine the threshold value and range value for URL properties. This project's goal is to develop an ELM categorization for a number of database characteristics and a few phishing sites.

Keywords: Support Vector Machine (SVM), Extreme Learning Machine (ELM), URL Phishing Websites, Browser add-ons, Random Forest.