PROJECT TITLE :

OFS-NN An Effective Phishing Websites Detection Model Based on Optimal Feature Selection and Neural Network

ABSTRACT:

Phishing attacks have become a major menace to people's daily lives and social networks. Attackers might persuade users to access phishing URLs by masking unlawful URLs as genuine ones in order to obtain private information and other benefits. To combat the threats posed by phishing assaults, effective ways of detecting phishing websites are urgently needed. The neural network is extensively used to detect phishing assaults because of its active learning capability from large data sets. However, many unnecessary and minor effect features will trap the neural network model in the dilemma of over-fitting during the training data sets stage. This issue frequently results in a trained model that is unable to detect phishing websites. This research introduces OFS-NN, an effective phishing website detection model based on the optimal feature selection approach and neural network, to address this issue. A new index, feature validity value (FVV), is first introduced in the proposed OFS-NN to evaluate the impact of sensitive characteristics on phishing website identification. After that, an algorithm is created based on the new FVV index to identify the best features from phishing websites. This approach is capable of alleviating the underlying neural network's over-fitting problem to a great extent. The underlying neural network is trained using the selected optimal features, and then an ideal classifier is built to detect phishing websites. The OFS-NN model is accurate and reliable in detecting various types of phishing websites, according to the findings of the experiments.