Thank you for making your code public. I run the demo notebook "flow_example.ipynb" you proposed. I now have the output with the Shap values associated with the features that most influence the n anomalies with greater reconstruction error. This output cannot be directly used for the visualization proposed in the original Shap paper [https://github.com/slundberg/shap]. How can I adapt the result I got from your implementation?
Kind regards,
Alessandro
Thank you for making your code public. I run the demo notebook "flow_example.ipynb" you proposed. I now have the output with the Shap values associated with the features that most influence the n anomalies with greater reconstruction error. This output cannot be directly used for the visualization proposed in the original Shap paper [https://github.com/slundberg/shap]. How can I adapt the result I got from your implementation?
Kind regards,
Alessandro