Identification of Fraudsters Involved in Phishing by Different Machine Learning Models

Abstract

With digitization of the current age, number of fraudsters in the digital realm has increased manifolds. Although the internet can be used for much good of the general population, the increase in number of unscrupulous people in online is a grave danger to the general public. Among many of the vices in the internet, one of the common one is phishing. To tackle phishing many approaches has been taken, of them ML based approach is one of the leading approaches. In our research work, we compared and contrasted many ML models to find out which one is most suitable for phishing detection. Our research is unique in regards that we have integrated data preprocessing and reduced the number of features for complexity reduction. Among these models XGBoost brought the highest accuracy after the hyperparameter tuning which was 97.0455%.