In today’s world, we are witnessing that the ever-growing advancement of technologies like e-commerce, ebanking, e-registration, etc had vast impact in a lot of factor on ours life. Attacks caused by Phishing have promptly manifested as a prime issue of cybersecurity. Fake web pages or phishing websites developed by attackers to fool and rob vital information of users such as username and password. However there are a number of methods to predict phishing, phisher's tactics were developed to avoid being detected. The most suitable way of predicting phishing is machine learning as maximum phishing attacks have common attributes which can be easily identified by machine learning algorithm. However, the precise capturing of fake webpages is a difficult topic as being directly proportional to dynamic aspects. Our study unfolds the Decision Tree (DT) classifier consisting significant attributes selection, to identifying fake websites with the aim of enriching the classification of webpages as fake or legal webpages. To perform the experiments we have used a publically accessible phishing website dataset from the UCI machine learning repository, which contains 4899 phishing webpages and 6158 legal webpages. In our study, our team firstly gather attributes form the dataset and then we train our DT model and at last we test it and we have achieved 98%(approx) accuracy by our feature selection technique, which surpassed the DT classification when compared to other feature selection techniques.



Software And Hardware