Hybrid Machine Learning-Based Detection of Phishing Attacks through E-mails and SMSs in Electronic Vehicles

Electronic Vehicles (EVs) face rising phishing threats directed at e-mail and SMS communications because users and vehicle systems now spread across external networks. Such incidents result in user data loss, unauthorized access to individual information, and manipulate control of vehicle systems. In this context, this paper proposed a hybrid model for detecting EV phishing attacks that integrates Convolutional Neural Networks (CNN) together with long-short-term memory (LSTM). The proposed model included transformation, multiple preprocessing steps, tokenization along with stop-word removal. The developed system achieved 98.21% test accuracy in detecting SMS phishing while detecting email phishing with a 96.19% accuracy rate. The experimental results validate that the CNN-LSTM model is a powerful phishing protection system for EV communication networks that improves security in the advanced connected vehicle ecosystem.