Enhanced Lightweight Model for Detection of Phishing URL Using Machine Learning

Abstract

In the past few years, many forged sites are created on the Internet to impersonate genuine websites, with the goal of acquiring sensitive and valuable information from the people and firms. Such type of attack which is carried online is known as phishing, and it continues to carry serious threats for consumers and business and the many shareholders’ hundreds of million dollars. Many cyber intrusions are successful through phishing. Phishing attacks are the attacks wherein Internet users are fooled by disguising a fake website as a legitimate website. As technology progresses, the phishing detection methods need to get advanced and there is a terrible requirement for improved mechanism to avoid, check, as well as detect these phishing attacks. Thus, useful remedial and preventive measures that can precisely find out phishing sites are a machine learning approach. Machine learning (ML) is a current tool for data analysis and lately has revealed capable leads in fighting phishing problems. Although many varieties of classification algorithms for detecting phishing are proposed, examined, and analyzed in many papers, it observed that advanced and depth of phishing threat are being continuously increased at steady rate. In this paper, we project a new method called enhanced lightweight model for phishing site detection. We further suggest the use of various URL features and boosting algorithm. We will analyze and examine the impact of boosting algorithm against the performance of other classification algorithms to find better result.