SignalExplainer

SignalExplainer is a specialized tool designed for deep learning models that work with signal-based data. It provides visual insights to explain the decision-making process of the model when classifying signal data.

How to Use

Usage: python signalExplainer.py Example: python signalExplainer.py model/multi_convlstm.h5 data/dataset.npz

NPZ File Structure:

• X: A numpy array with shape (n_samples, signal, n_channels).
• y: A numpy array with shape (n_samples, 1).
• channel_names: A numpy array with shape (numberOfChannels, 1).
• class_names: A numpy array with shape (numberOfClasses, 1).

TF Model Structure:

• model: A Keras model.

Note: Ensure that X is ordered consistently with the channels in the model.

Citation

If you find this tool useful, please consider citing our work:

@inproceedings{jalayer2023model,
  title={A Model Identification Forensics Approach for Signal-Based Condition Monitoring},
  author={Jalayer, Masoud and Shojaeinasab, Ardeshir and Najjaran, Homayoun},
  booktitle={International Conference on Flexible Automation and Intelligent Manufacturing},
  pages={12--19},
  year={2023},
  organization={Springer}
}