plot_shap

[jacobnikl]

I have a very short question:
Why does plot_shap call "shap_values = shap_values[:, :, 0]", doesnt this lead to the resulting plots (e.g., beeswarm plots) giving results for the "negative" class (assuming 0 is the negative class in the data)? For example in my dataset, 0 = no remission and 1 = remission from a mental illness. Using plot_shap would yield a beeswarm plot where positive values for a sample-feature combination would indicate higher probability for non remission correct? I see that this is not wrong but i wonder why this approach was chosen as it may confuse people (at least i was confused).