I am actually working with the iris dataset from sklearn and try to understand the ANFIS-Package for python. But that does not really matter! I have a more general question.
During thinking about adaptive neuro-fuzzy inference system (ANFIS), a general question came into my mind. I don't really understand: in general, why is ANFIS necessary?
So, for example, if I want to predict classes for this iris dataset, I also can use a supervised learning method or a neural network and I get the result.
In ANFIS, I do nothing other than splitting the input attributes into linguistic terms and give membership functions to it. At the end of the day, I will receive "predictions" for the input values, which are classes.
But - with the ANFIS-Package in Python - I cannot see, if my membership function has changed during the learning time or what rules the network constructed. So, I cannot really see why this is useful. Maybe it is just because I am usually using the iris dataset for supervised learning.
