Define interfaces to define input and output of a MachineLearningModel
Two interfaces should bridge the gap between the Knowledge-Base and the Machine-Learning-World. This enables the decoupling of machine learning model and the rest of the knowledge base, as well as an easy way of extending eraser with new machine learning models.
The input interface defines the connection of Item states and values for the input neuron(s) or decision tree rules.
The output interface defines how to interpret the values of the output neuron(s) or decision tree leafs and get assignable values for certain items.
Both interfaces need more details to be specified as of now.