Hybrid butterfly optimization and back propagation neural network for enhanced smart city data classification

Area of Science:

• Computer Science
• Artificial Intelligence
• Data Science

Background:

• Smart cities generate vast, imbalanced datasets from urban systems like traffic and utilities.
• Accurate classification of this heterogeneous data is crucial for resource efficiency and citizen well-being.
• Existing methods struggle with the scale and complexity of smart city data.

Purpose of the Study:

• To develop a novel hybrid classification framework (HBPNNBO) for enhanced smart city data analysis.
• To improve the accuracy and efficiency of classifying imbalanced urban datasets.
• To ensure data integrity and privacy through blockchain and advanced encryption.

Main Methods:

• A hybrid classification framework integrating Butterfly Optimization Algorithm (BOA) with Back Propagation Neural Network (BPNN).
• Data preprocessing using HADASYNBSID for dataset balancing and Hybrid Chicken Swarm Genetic Algorithm (HCSGA) for feature selection.
• Integration of blockchain technology with a hybrid AES-CSO encryption method for secure data handling.

Main Results:

• The HBPNNBO model achieved a classification accuracy improvement of 94.76%.
• The framework demonstrated a minimal processing time of 23.62 ms.
• Evaluated on intrusion and traffic datasets, HBPNNBO outperformed conventional classifiers.

Conclusions:

• The HBPNNBO framework offers a highly accurate and efficient solution for smart city data classification.
• Blockchain-based security ensures end-to-end data integrity and privacy in urban analytics.
• The proposed model is suitable for real-time, secure smart city data processing and analysis.