A Neural Network Based Heuristic Phishing Pattern Prediction Method in Cloud

Abstract

Cloud Computing has gained huge attention over the past decades because of continuously
increasing demands. There are several advantages to organizations moving towards cloudbased
data storage solutions. These include simplified IT infrastructure and management,
remote access from effectively anywhere in the world with a proper internet connection as
well as its cost effectiveness. But with growing acceptance there lies the challenges of
privacy and security that requires further exploration. As millions of ordinary citizens are
using networks for banking, shopping, and filing their tax returns, network security is
looming on the horizon as a potentially massive problem. One of the most popular
information theft techniques is phishing. Nowadays information theft can cause a great deal
of trouble to the victim in a very short interval of time. Thus, a predictive model which can
predetermine threat is required of current scenario. We have also decided to create a highly
accurate predictive engine using Backpropagation Neural Networks. This, prediction engine
can detect and predict several types of threats. We then trained our prediction model to
specialize in phishing websites detection. Training was done using an available dataset. A
single page application was then created which provided the UI for prediction of phishing
websites. This application uses the prediction engine as the backend. General public can use
this application to determine whether a website is legitimate or not, only by providing certain
characteristics of that website as input features. Generally, prediction engines are used only
for research purposes, but since we have created an application which can be distributed, so
more and more internet users can become aware of phishing threats and get saved from
potential phishing attacks.