A novel optimized neural network model for cyber attack detection using enhanced whale optimization algorithm

Area of Science:

• Cybersecurity and Network Security
• Artificial Intelligence in Cybersecurity
• Computational Intelligence

Background:

• Cybersecurity is paramount in our interconnected digital world.
• Credential stuffing attacks, leveraging stolen credentials, pose a significant threat to online account security.
• Widespread reuse of passwords across multiple platforms exacerbates this vulnerability.

Purpose of the Study:

• To address the challenges of detecting and predicting credential stuffing attacks.
• To introduce a novel model for enhanced cybersecurity defense.
• To improve the security of online accounts against unauthorized access.

Main Methods:

• Development of a novel Enhanced Whale Optimization Algorithm (EWOA) for training.
• Integration of EWOA with an Artificial Neural Network (ANN) to create the EWOA-ANN model.
• Utilizing the EWOA-ANN model for credential stuffing attack detection and prediction.

Main Results:

• The proposed EWOA-ANN model demonstrates effectiveness in identifying credential stuffing attacks.
• The model shows promise in predicting the occurrence and nature of these cyber threats.
• Empirical comparisons are planned to validate the model's performance against existing security analyses.

Conclusions:

• The EWOA-ANN model offers a robust solution for combating credential stuffing cyberattacks.
• This approach enhances the detection and prediction capabilities crucial for online security.
• The study contributes a novel optimization technique for improving neural network performance in cybersecurity applications.