Applying Machine Learning for Identifying Fraud Sites

Abstract

Offenders looking for sensitive information create illicit clones of legitimate websites and e-mail accounts. The email will contain actual company logos and phrases. When a User selects one of these hackers’ links, the hackers obtain availability of all the significant user data, including credit card detailsinformation, personal login passwords, and photos.Random Forest Decision Tree methods and are employed often in current systems, and their accuracy must be improved. The current models to be low latency. Existing systems lack a specialised user interface.Not all algorithms are compared in the present system. When consumers read the e-mails or links given, they are sent to a phoney website that looks to be from the legitimate firm. The models are used to detect phishing Websites based on URL importance factors and to discover and execute the best machine learning model. The Multinomial linear regression and other predictive modelling techniques are included in a comparison. Simple Bayes, and XG Boost. Logistic Regression beats the other two algorithms.