Classification of Phishing Mail Detection Using Hybrid Approach of Fusion Model

Email is one of the most common tactics used by cybercriminals today. The main reason is that emails are very easy to find, and sending bulk emails is quite cheap, and some of these users will definitely fall victim. This study was designed to gain insight into the development of T-planning components in current research for phishing prevention. Our study followed the guidelines of systematic review and meta-analysis. The research delineates a taxonomy of phishing detection methodologies employing a hybrid approach and evaluates their efficacy in tackling contemporary difficulties. Our study has created a customized corpus of phishing and legitimate emails and provided efficient methodologies for identifying phishing emails by integrating deep contextual embeddings with machine-learned metadata representations inside a cohesive neural framework. The hybrid model attains an accuracy of 99.18%, precision of 98.98%, F1 score of 99.19%, and recall of 99.18% on the in-house corpus dataset. This encouraging outcome surpasses some current detection technologies and confirms the efficacy of the Fusion model in identifying phishing emails.

Hali tarjima qilinmagan