Phishing website prediction using base and ensemble classifier techniques with cross-validation.

Area of Science:

• Computer Science
• Cybersecurity
• Machine Learning

Background:

• The internet is increasingly vulnerable to various cyber threats, including phishing, which compromises personal data and causes significant user losses.
• Existing security measures like encryption and digital certificates are insufficient to combat sophisticated phishing attacks.
• Accurate detection and prediction of phishing websites are crucial for safeguarding internet users.

Purpose of the Study:

• To develop and evaluate machine learning classifiers for detecting and predicting phishing website URLs.
• To investigate the effectiveness of ensemble-based techniques in improving phishing detection accuracy.
• To provide a comprehensive performance analysis of different classification methods on distinct and merged datasets.

Main Methods:

• Utilized two distinct datasets containing features of website URLs for training and testing.
• Implemented primary machine learning classifiers and novel ensemble-based techniques.
• Conducted a three-phase study involving base classifiers, ensemble classifiers, and ensemble classifiers with and without cross-validation.

Main Results:

• The study analyzed the performance of various classification models in identifying phishing URLs.
• Ensemble classifiers demonstrated potential for enhanced accuracy in phishing website detection.
• Comparative analysis provided insights into the effectiveness of different methods and validation strategies.

Conclusions:

• Machine learning and ensemble techniques offer promising solutions for combating phishing.
• The findings contribute to the ongoing research in cybersecurity by providing a robust framework for phishing website detection.
• This study serves as a valuable resource for future research in developing more effective anti-phishing strategies.