Detecting Phishing Websites Using an Efficient Feature-based Machine Learning Framework

Abstract

Phishing is a form of digital crime where spam messages and spam sites attract users to exploit sensitive information on fishermen. Sensitive information obtained is used to take notes or to access money. To combat the crime of identity theft, Microsoft's cloud-based program attempts to use logical testing to determine how you can build trust with the characters. The purpose of this paper is to create a molded channel using a variety of machine learning methods. Separation is a method of machine learning that can be used effectively to identify fish, assemble and test models, use different mixing settings, and look at different mechanical learning processes, and measure the accuracy of the modified model and show multiple measurement measurements. The current study compares predictive accuracy, f1 scores, guessing and remembering multiple machine learning methods including Naïve Bayes (NB) and Random forest (RF) to detect criminal messages to steal sensitive information and improve the process by selecting highlighting strategies and improving crime classification accuracy. to steal sensitive information.