Phishing Website Detection using Machine Learning

Phishing is a common and cunning method used by attackers to rob users of their personal details. They assume the identity of trusted sources, getting users to divulge bank information, usernames, and passwords. It is critical for cyber experts to develop authentic methods of identifying and blocking such advanced threats. This paper discusses using machine learning to identify phishing URLs. We try to produce strong models that can discriminate between real and spurious URLs on the basis of varied features of both. Decision tree, random forest, (SVM) the Support Vector Machine, XGBoost, Back-propagation CNN Convolutional neural network, also known as CNN, are the algorithms we employ for this. Its accuracy, false-negative and false-positive rates, and other indicators of the algorithms are compared. The aim is to find the most effective algorithm that can identify phishing URLs. According to the result, machine learning is employed to identify phishing URLs and obtain useful information that can strengthen the defense of cybersecurity. Support Vector Machine, or SVM, achieved a mean accuracy of 97.0% in the identification of a phishing URL for this project, and the model based on the neural network (Backpropagation) yielded well-balanced outputs for varying parameters, indicating that both models are helpful to make good phishing threat identification inputs.