Time Series: Explain Predictions

[chukarsten]

Addresses: #1504

Oh my goodness. You've been waiting for it, so here it is.

[codecov]

Codecov Report

Merging #1818 (cb0e9cc) into main (2747fbc) will increase coverage by 0.1%.
The diff coverage is 100.0%.

@@           Coverage Diff           @@
##            main   #1818     +/-   ##
=======================================
+ Coverage   99.9%   99.9%   +0.1%     
=======================================
  Files        255     255             
  Lines      20571   20608     +37     
=======================================
+ Hits       20549   20586     +37     
  Misses        22      22             

Impacted Files	Coverage Δ
...derstanding/prediction_explanations/_algorithms.py	97.2% <100.0%> (ø)
...prediction_explanations/_report_creator_factory.py	100.0% <100.0%> (ø)
...tanding/prediction_explanations/_user_interface.py	100.0% <100.0%> (ø)
...nderstanding/prediction_explanations/explainers.py	100.0% <100.0%> (ø)
...s/prediction_explanations_tests/test_explainers.py	100.0% <100.0%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2747fbc...cb0e9cc. Read the comment docs.

[freddyaboulton]

Doing this to avoid having to compute the features twice.

[freddyaboulton]

@chukarsten Looks good to me! It was a pleasure getting this done together. I just have some minor clean-up style comments.

[freddyaboulton]

This is because the error metrics are now stored in a numpy array.

[freddyaboulton]

We're not checking the output against a specified value. The other tests check that the logic for selecting the rows that are explained and displaying the right features/shap values is correct so I think running real data through and verifying it completes without showing NaNs is sufficient.