Phishing URL Classification Using K-Nearest Neighbour and Logistic Regression Machine Learning Approaches

With rising Internet use across the developing countries, increasing network security is vital because it minimizes the dangers of social privacy spoofing, identity or information theft, and financial fraud. Phishing and spam emails are two of the most prevalent network security breaches because they are vulnerable to transmit a virus or a malicious website, potentially leading to widespread fraud. Phishing attacks are defined as the theft of consumers' personal information by a non-trusted source acting as a trustworthy source; however, this is not always the case. Phishing occurs when a link pretends as legitimate in order to deceive unaware consumers and drive them to do an activity that they would only take if they trusted the source. To distinguish phishing from legitimate URLs, link classification methods like Logistic Regression and K-Nearest Neighbors (K-NN) may be used. In this study, we proposed a machine learning-based phishing URL system by analysing the URLs using these two distinct categorization methods and comparing the findings with prior studies using different datasets. The results of the experiments demonstrate how well the recommended models function according to the proposed methods.