International Science and Technology Journal

This research aims to study the effectiveness and performance of the proposed classifier to detect phishing emails, because there is an urgent need to develop information security systems that are accurately and proactively able to recognize phishing messages due to their increasing number and diversity of fraudulent capabilities. Since this type of phishing message manipulates human emotions leading to fears and creates a situation of urgency by claiming that the recipient must take immediate and swift action, which may lead to financial losses or significant data leakage losses. In order to overcome the human weakness in detecting and recognizing phishing emails, this study was conducted because we need to continuously enhance and improve the accuracy and effectiveness of automated and automatic phishing detection algorithms and methods. The proposed classification model utilizes a hybrid algorithm that combines deep learning (DL) multi-layer perceptron (MLP) neural network algorithms and natural language processing (NLP) methods on the body of the received email. This paper highlights the importance of examining the textual features of the body of a mail message for phishing detection, using multi-layer perceptron (MLP) neural networks to analyze the accuracy of detecting phishing through the message text, and because text features represent a relatively new direction of study in the field of email phishing detection. The proposed model was tested on a balanced and labeled dataset of 8579 different messages, and the results showed an improvement in classification accuracy and performance compared to other deep learning methods. The proposed classifier model was evaluated using the following metrics: (Recall, Accuracy, Precision, and F-measure), and the results were obtained - 98.3%, 98.2%, 98.5%, and 98.55%, respectively. The model also showed good performance and took a short time to detect; to produce an overall accuracy rate of over 98.1% and a low false positive rate (FPR) of 0.015................. Keywords: ................Phishing Email detection, MultiLayer Perceptron (MLP) Neural Network, Deep learning (DL), Natural language processing (NLP), Algorithms, Information security.