Hybrid Bacterial Foraging and Particle Swarm Optimization for detecting Bundle Branch Block

Identifying abnormal cardiac rhythms remains a critical challenge in diagnosing heart conditions. While ECG feature extraction is essential for detecting heart diseases, traditional methods often struggle with accuracy and efficiency. Prior research has shown that heuristic optimization techniques, such as Bacterial Forging Optimization (BFO), can enhance feature selection processes. However, these methods may lack speed in reaching optimal solutions. No prior work had resolved the issue of combining local and global search strategies to improve ECG signal classification. This gap motivated the development of a hybrid approach that integrates BFO with Particle Swarm Optimization (PSO). That uncertainty drove the exploration of how PSO could complement BFO's chemotactic movement. It was already known that BFO relies on random search directions, which may delay convergence. This limitation suggested the need for a more efficient optimization framework.

The aim of this work is to improve the detection of left and right bundle branch block (LBBB and RBBB) in ECG signals. Bundle branch block occurs when electrical impulses in the heart face difficulty moving through the circulatory system. Detecting these anomalies is crucial for diagnosing heart ailments. The study proposes a hybrid optimization method that combines Bacterial Forging Optimization (BFO) and Particle Swarm Optimization (PSO). This BFPSO method seeks to enhance feature selection by integrating the strengths of both algorithms. The study addresses the problem of slow convergence in BFO by introducing PSO's global search capabilities. The motivation stems from the need to improve the accuracy and speed of ECG signal classification. This approach aims to provide a more efficient solution for identifying BBB patterns in ECG data.

The proposed method integrates two heuristic optimization algorithms: Bacterial Forging Optimization (BFO) and Particle Swarm Optimization (PSO). BFO simulates the chemotactic movement of bacteria to explore the solution space. PSO mimics the social behavior of swarms to guide the search process. The hybrid BFPSO method combines these two approaches to improve feature selection in ECG signals. BFO's chemotactic process is used for local search, while PSO handles global search. The BFPSO algorithm generates new individuals in each generation to refine the solution. The selected features from BFPSO are then fed into a Levenberg-Marquardt Neural Network classifier. This classifier is used to distinguish between normal and abnormal ECG patterns. The study evaluates the performance of BFPSO in detecting bundle branch block conditions.

The BFPSO method demonstrated improved performance in detecting bundle branch block patterns in ECG signals. The hybrid approach achieved higher accuracy compared to using BFO or PSO alone. The BFPSO method reduced the time required to reach an optimal solution. The Levenberg-Marquardt Neural Network classifier showed high sensitivity in distinguishing abnormal ECG patterns. The study reported a classification accuracy of 94.3% for detecting LBBB and RBBB. The BFPSO algorithm outperformed traditional methods in terms of convergence speed. The method successfully identified key features in ECG signals that are indicative of BBB. These results suggest that the BFPSO method is effective for ECG signal classification tasks.

The BFPSO method successfully combines Bacterial Forging Optimization and Particle Swarm Optimization to detect bundle branch block in ECG signals. The hybrid approach improves both the accuracy and speed of feature selection. The study shows that BFPSO outperforms individual optimization methods in ECG classification. The Levenberg-Marquardt Neural Network classifier effectively distinguishes between normal and abnormal ECG patterns. The results suggest that BFPSO is a viable method for detecting BBB conditions. The study does not propose new diagnostic criteria or treatment approaches. The findings align with the authors' goal of improving ECG signal classification. The BFPSO method may serve as a useful tool in automated heart disease detection systems.