Explainers¶
Explainer is very similar to the Model and in fact it inherits the same superclass, but it has different inference interface – accepting not only data, but also another Model, it should generate an explanation of some format. The idea here is similar to the model’s – to be able to incorporate in the benchmark as many types as possible the interface made very broad. One can have feature importance, example selection/synthesis or trees under one umbrella-class which allows common interface for handling different entities similarly.