The x_train looks like this (22 features):
total_amount reward difficulty duration discount bogo mobile social web income ... male other_gender age_under25 age_25_to_35 age_35_to_45 age_45_to_55 age_55_to_65 age_65_to_75 age_75_to_85 age_85_to_105
0 0.006311 0.2 0.50 1.000000 1.0 0.0 1.0 1.0 1.0 0.355556 ... 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0
1 0.015595 0.2 0.50 1.000000 1.0 0.0 1.0 1.0 1.0 0.977778 ... 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0
The label is 0 and 1, it's a binary classification problem, here's the code for building the model, and I was following this page to implement SHAP:
#use SHAG
deep_explainer = shap.DeepExplainer(nn_model_2, x_train[:100])
# explain the first 10 predictions
# explaining each prediction requires 2 * background dataset size runs
shap_values = deep_explainer.shap_values(x_train)
This gave me error:
KeyError: 0
During handling of the above exception, another exception occurred
I have no idea what this message is complaining, I tried to use SHAP with a XGBoost and Logistic Regression model and they both work fine, I'm new to keras and SHAP, can someone have a look for me and how I can solved it? Many thanks.
I think SHAP (whatever it is) is expecting a Numpy array and so indexing x_train like a Numpy array, it yields an error. Try:
shap_values = deep_explainer.shap_values(x_train.values)