Novel Techniques for Process Topology Reconstruction and Fault Diagnosis (PTR-FD)

This repository contains the codebases for the three PTR-FD algorithms investigated in my MASc thesis. The three algorithms include:

1. Granger Causality [1]
2. Granger Net [2]
3. Extended Convergent Cross-Mapping [3, 4]

Setup

To use this repo, first clone it.

Next, install external module dependencies:

pip install -r requirements.txt

Quickstart

A guide on how to use these functions provided in packages can be found using the Jupyter Notebook Quickstart.ipynb.

The functions are also callable through the command line (examples provided below). These functions require a csv containing only the time series variables of interest for the analysis, and the functions visualize the resulting output as a causal heatmap.

Granger Causality

python3 -m packages.granger_causality.granger_causality <path_to_csv> <max_lag> --autocausation=True --pval=0.05

Granger Net

python3 -m packages.granger_net.core.analysis <path_to_csv> <max_lag> --autocausation=True --epochs=3000 --initial_batch_size=32 --threshold=0.1

Extended Convergent Cross-Mapping

python3 -m packages.eccm.models.eccm.eccm <path_to_csv> --cross_map_lags=5 --use_all_points=True --criterion=Peak --p_val=0.05 --verbose=True

References

1. Clive WJ Granger. Investigating causal relations by econometric models and cross-spectral methods. Econometrica: Journal of the Econometric Society, pages 424–438, 1969.
2. Alex Tank, Ian Cover, Nicholas J Foti, Ali Shojaie, and Emily B Fox. An interpretable and sparse neural network model for nonlinear granger causality discovery. arXiv preprint arXiv:1711.08160, 2017.
3. George Sugihara, Robert May, Hao Ye, Chih-hao Hsieh, Ethan Deyle, Michael Fogarty, and Stephan Munch. Detecting causality in complex ecosystems. Science, 338(6106):496–500, 2012.
4. Hao Ye, Ethan R Deyle, Luis J Gilarranz, and George Sugihara. Distinguishing time-delayed causal interactions using convergent cross mapping. Scientific reports, 5:14750, 2015.