Classification of web-based phishing attacks using deep learning method

Web-based phishing attacks are a method used by attackers to attain their important information and to trick users in an online media. This type of attack is accomplished by emulating websites to mislead users. Means such as advertising or e-mail are used as mediators. It is a big problem to distinguish whether a website is phishing. Therefore, phishing web sites, users or customers to prevent theft of personal information and to use the classification methods to detect this situation plays an important role in solving the problem. In this study, it is aimed to prevent web-based phishing attacks that increase in years and to prevent users from being affected by these types of attacks. For this purpose, deep learning method was used for classification process. This method is an approach that allows the classification and analysis of unknown data by training the known data with a multi-layer artificial neural network. For the classification process, a data set with 5000 legitimate and 5000 phishing websites was used. Using this data set, classification was made with a trained model. As a result, high accuracy values were obtained in this classification process and it was seen that deep learning method showed successful results.