Frequency-Selective Memory Networks for Adaptive Signal Processing - Source Code and Datasets

This repository contains the complete implementation and experimental materials for the paper "FSM-Net: Frequency-Selective Memory Networks for Adaptive Signal Processing in Deep Learning". **Contents:**- Complete Python source code for FSM-Net architecture- Preprocessed CWRU bearing fault dataset and MIT-BIH Arrhythmia database- Training and evaluation scripts with hyperparameter configurations- All experimental result datasets and performance metrics- Publication-ready figures and visualization materials- Comprehensive documentation and reproduction instructions **Key Features:**- Learnable frequency decomposition with adaptive spectral boundaries- Dual-pathway LSTM processing for target enhancement and interference suppression- Context-aware adaptive fusion mechanisms with multi-head attention- Cross-domain validation on mechanical and biomedical signals **Requirements:**- Python 3.8+- PyTorch, NumPy, Matplotlib, and other dependencies