Chaotic Map-Based Weight Optimized Spiking Neural Network (Snn) For Binary Classification Of Breast Cancer

Abstract

Spiking Neural Networks (SNNs) currently attract broad attention in biomedical classification because of their biological background combined with their energy-efficient operation. Traditional SNNs achieve slow convergence rates together with weight tuning issues which negatively affects their classification accuracy. A Chaotic Map-Based Weight Optimization approach serves as a solution to optimize SNN learning dynamics for breast cancer binary classification. The implementation of Logistic Chaotic Map (LCM) optimization together with Spike Timing-Dependent Plasticity (STDP) serves to optimize weights and speed up convergence and avoid getting trapped in local optima. The chaotic weight update procedure undergoes constant improvement for synaptic refinement which results in generalized outcomes effectively. The chapter introduces mathematical aspects and learning procedures alongside algorithmic execution procedures and performance assessments of the developed model. The proposed model achieves better accuracy and precision as well as recall and F1-score metrics when evaluated on breast cancer datasets compared to traditional SNNs. Chaos-driven weight optimization demonstrates its effectiveness for biomedical classification tasks through the findings which strengthen the potential of SNN for medical diagnostics.