Answers are grounded in this paper and cite their sources.
As the complexity of smart networks continues to rise, ensuring their security while maintaining data privacy has become a critical challenge. The integration of blockchain technology with machine learning offers a promising solution by enabling decentralized data processing and enhancing privacy. However, there are still significant obstacles in securing these networks, particularly when it comes to intrusion detection systems (IDS) that must accurately identify threats while preserving sensitive data. This abstract presents a novel approach that combines blockchain-based federated learning (BCFL) with advanced machine learning techniques to address these challenges. The proposed methodology incorporates several key components: data collection, preprocessing, feature extraction, feature selection, and classification. For feature extraction, an Enhanced Convolutional Neural Network with Squeeze-and-Excitation blocks (ECNNSE) is employed to capture and enhance important patterns in network traffic data. To further improve model performance, hyper-parameter optimization is achieved using the Butterfly Optimization Algorithm (BOA), ensuring the ECNNSE model is finely tuned for optimal results. Feature selection is performed using the Correlation-Based Feature Selection (CFS) method, which identifies and selects the most relevant features, reducing computational complexity and improving model efficiency. The integration of BCFL ensures secure, privacy-preserving decentralized training across multiple devices, allowing collaborative learning without exposing sensitive data. The Attacks Prediction dataset achieved an accuracy of 97% , a precision of 96% , a recall of 95% , and an F1-score of 95% .
No takes yet. Share an insight, caveat, or question.
Binbusayyis et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: