Advancing web security: Analysis of machine learning algorithms in detection of phishing websites

Reflective minds believe that the entire world has transitioned to the internet. From banking to shopping, almost everything can now be done online. Today, people have greater access to the internet than at any other time in history. The increased dependence on these digital platforms has also evolved cases of cybercrime. The word Cyber means which is related to the computer specifically Internet. With the fast growing world we became more dependent on technology and Internet which gives rise to Cyber threats. Cyber crime is illegal action performed using the Internet. Criminals have learned to use the digital environment for their own ills. Phishing is one among them; here, attackers will try to mimic the authentic sites and deceive others to make their sensitive information available to them. Phishing is an online assault that has racked up hundreds of millions of dollars in losses to the online community and its various stakeholders. Therefore, powerful countermeasures are also needed that can accurately classify phishing. This study examines the various machine learning techniques like random forest (RF), Support Vector Machine (SVM), KNN, Naive Bayes, and Extreme Gradient Boosting (XGBoost) are used to identify phishing websites. The research compared the classification models and found that Random Forest is the best model with 95.15% accuracy and The proposed Hybrid Model(DT+RF+XGB) outperformed the other with 95.63% accuracy for detecting phishing attacks based on website content.