InterpretML | 11A: Local explanations

InterpreML provides local explanation.
This is a model for predicting recurrence of breast cancer based on patient and tumor attributes. For this patient, the model has predicted no recurrence. Interpret ML makes clear why the model made this prediction (Guideline 11), providing a local explanation that lists attributes in order of their importance for the prediction (Pattern 11A). In this case, a smaller tumor size of 10-14 (range maximum is 59) and a low degree of 1.0 malignancy (range maximum is 3) are the top two most important attributes for the prediction. Image captured July 2020.