Boosting Guided Probabilistic Ensemble-based Approach For Phishing Website Detection
Avisheak Das, Fahim Irfan Alam, Sadia Sharmin, Rokan Uddin · 2022 International Conference on Innovations in Science, Engineering and Technology (ICISET) · 2022
Phishing is a cyber crime that targets naive online users to disclose their sensitive and personal information such as username, pin, password, social insurance numbers, credit card numbers, etc. To prevent such alarming event from taking place, an early detection of Phishing websites is significantly important and may keep users’ sensitive information safe. Thanks to the Machine Learning (ML) based techniques, there have already been automated methods available for detecting phishing websites. Although traditional ML approaches achieve success to some extent, there are still possibilities to improve the performance by introducing probabilistic approach in the solution. In this paper, we propose a probabilistic ensemble-based integrated solution that combines individual prediction probabilities from different classifiers and enhances the overall performance by introducing probabilistic estimation guided boosting classifier in the end. Additionally, we performed state-of-the-art data preprocessing methods to eliminate unwanted outliers and also, select statistically significant features from out data. Being evaluated with a benchmark dataset, we performed intensive experiments to verify the effectiveness of our proposed solution from which we obtained potential preliminary results.