Adaptive Phishing Threat Identification and Classification Using Hybrid ML Techniques in Cyberspace

Phishing attacks are becoming a more prevalent danger in the realm of cybersecurity, targeting individuals and organizations to obtain sensitive personal and financial information through fraudulent means. With the continuous advancement of phishing techniques, traditional detection methods similarbans and rule basedstructuresare finding it difficult to effectively counter the ever-changing tactics used by attackers. This paper reviews a broad range of ML algorithms applied to phishing detection, highlighting their effectiveness, limitations, and potential challenges in practical applications. It also identifies promising areas for future research, with an emphasis on improving classification accuracy, reducing false positives, and developing adaptive detection models that can enhance cybersecurity defences in real-time. The proposed ensemble machine learning process such as Random Forest and CNN obtains 97 % accuracy and also reduced error rate by 5 %. The testing results highlights proposed technique provides better efficiency in terms higher accuracy, and minimal computational overhead which formulated it appropriate for real-time phishing identification.