Cloud-based email phishing attack using machine and deep learning algorithm.

Area of Science:

• Cybersecurity
• Computer Science
• Machine Learning

Background:

• Email is a primary communication channel for sensitive data.
• Phishing attacks exploit user trust to steal confidential information.
• Existing methods for detecting phishing emails require enhancement.

Purpose of the Study:

• To develop and evaluate machine learning models for accurate email phishing detection.
• To compare the performance of Support Vector Machine (SVM), Naive Bayes (NB), and Long Short-Term Memory (LSTM) algorithms in classifying phishing emails.
• To improve the security of email communication by identifying and mitigating phishing threats.

Main Methods:

• A modified dataset of legitimate and phishing emails was created.
• Feature extraction was performed on the dataset, resulting in CSV and label files.
• The Support Vector Machine (SVM), Naive Bayes (NB), and Long Short-Term Memory (LSTM) algorithms were applied for email classification.

Main Results:

• The SVM classifier achieved the highest accuracy at 99.62%.
• The LSTM classifier demonstrated strong performance with 98% accuracy.
• The NB classifier achieved 97% accuracy in detecting email phishing attacks.
• All tested algorithms showed high efficacy in identifying phishing attempts.

Conclusions:

• Machine learning algorithms, particularly SVM, NB, and LSTM, are highly effective for detecting email phishing attacks.
• The proposed methods offer a robust solution for enhancing email security.
• Accurate classification of phishing emails is crucial for protecting users and sensitive data.