shap_values.base_values become weird when the size of maskers is larger than 100.

[yipei-wang]

I tried a simple model model with input data x and background data X. The exact SHAP is implemented as

explainer = shap.explainers.Exact(model, X)
shap_values = explainer(x)

From the paper, I expect that shap_values.base_values[0,0] should be equivalent to model(X).mean(). But the I found that this is true only if X.shape[0]<=100. When the number of samples in X is greater than 100, the reference value becomes weird and I don't know how it is calculated.

I am confounded by this phenomenon, am I doing it wrong? I really appreciate it if anyone can help me with this issue!

[AnotherSamWilson]

This is probably caused by a hard coded limitation to 100 samples in the background set:
#1991

Caused by this guy:
https://github.com/slundberg/shap/blob/b6e90c859fdfc6bc145242d9a8082d4ad844e995/shap/maskers/_tabular.py#L243

[github-actions]

This issue has been inactive for two years, so it's been automatically marked as 'stale'.

We value your input! If this issue is still relevant, please leave a comment below. This will remove the 'stale' label and keep it open.

If there's no activity in the next 90 days the issue will be closed.

[github-actions]

This issue has been automatically closed due to lack of recent activity.

Your input is important to us! Please feel free to open a new issue if the problem persists or becomes relevant again.