v.0.4.7.

SETS

• Bug in PyTorch model prediction #68
• Added train #67
• Fixes in the generation processes to enable the elimination of class shaplets

Dependencies

• Changes Dependencies for Python 3.12. Compatibility

v.0.4.6

Minor

• Deleted Print in TSR see #69

v.0.4.5

TSR:

• Fix in Plot Function for univariate data #56

Native Guide

• Fix in counterfactual swap generator for PyTorch #63

0.4.4.

TSR:

• Fix in Plot Function for univariate data

v.0.4.3

LEFTIST:

• Fix in Heatmap generation from component importance #59

NativeGuide:

• Fixed Name Inconsistencies in Default Param #58

v.0.4.2

Minor :

• Readme.md
• setup.py - h5py compatible for python > 3.10

SETS :

• First Release of TF-Version of "Shapelet-Based Counterfactuals for Multivariate Time Series" [1]

[1] Bahri, Omar, Soukaina Filali Boubrahimi, and Shah Muhammad Hamdi. "Shapelet-based counterfactual explanations for multivariate time series." arXiv preprint arXiv:2208.10462 (2022).

Contributors:

@belaboe97

Acknowledgement
We would like to thank Omar Bahri (@omarbahri) for letting us base this implementation on his original code https://github.com/omarbahri/SETS.

v.0.4.1

Update __version__.py

v.0.4.0

Multivariate Data, Resolved Shaping Issue #44 for:

• TSEvo
• COMTE
• TSR

v.0.3.4

Minor Fixes:

• Improved Error Handling for TSR #39
• .plot takes figsize-argument into account #40
• Improved Error Handeling for Native Guide #38

v.0.3.3

COMTE:

• Changed naming from AtesCF to COMTECF
• Eliminated MLrose dependency

TSR:

• opening up the masking parameters with **kwargs

NativeGuide

• minor in subarray NUN_CF