Phishing Attack Prediction using Several Machine Learning Techniques

With significant improvements in social media, banking, web technologies, smart grids, cloud, and online mobile environments, etc., cyber security is a rapidly evolving area that require close consideration. Cyber-attacks involve an organization using cyberspace to destroy, disrupt, disable, or maliciously control an IT infrastructure or compromise data confidentiality, integrity or steal organized information. The state of cyberspace bodes well for the future fragility of the Internet and the growth of its customers. Many attacks such as DoS attack, phishing attack has been identified from the review paper. Various machine learning techniques have been effectively used to solve such serious problems in the field of cyber security. Machine learning techniques are used in many scientific fields due to their unique properties such as scalability, ability, adaptability to predict security problems and quickly adapt to new and unknown problems. Deep learning methods, ensemble learning, Behavioral analysis, Natural Language Processing (NLP), Hybrid approaches and metaheuristic algorithms are used to solve the challenges like adaptability to evolving phishing tactics, data scarcity and class imbalance. This article predicts phishing attack using machine learning such as Support Vector Classifier, Logistic Regression and Random Forest Algorithm to achieve higher accuracy. The presented hybrid method will help network users to verify genuine websites, reducing the threat of phishing websites and ensuring a safe Internet experience.