I would like to ask one point , how do i decide these parameters , more appropriately how are you being so specific about the values of these hyperparameter . Just curious , how do i get more insights into rasa core ? i kind of have grasped rasa nlu but how do i grab onto the former one ? any suggestions ?
How did you grasp it for NLU?
evaluation would be the first step. when it comes to hyperparameters, it really depends with your problem definition, i have never built a large scale chatbot with an open domain like structure. I have used rasa to test/build task specific chatbots where assesing the max history is much easier because you know your target and what matters in a conversation. there are no general solutions
Also treat it as a Deep learning problem and specially an RNN and the steps you take to optimise your neural network ( gradient descent, activation)
You can learn about them in the Rasa core code as well and docs
Can you please elaborate your question a bit?