Hello Sir, Thank you for responding to my query.

My field of work is Machining where I have conducted some experiments to identify the accuracy of a machining.

In my experimental study I have used 4 machining inputs to do the tests and I have measured different accuracy parameters as outputs.

Actually the number of outputs that I have measured exceeds the number of inputs (i.e., 4 inputs and 6 outputs). Also, I can make only a few number of (less than 15) 'IF and THEN Fuzzy rules.'

In this case, is it recommendable to use Fuzzy analysis for modelling?

Thanks in advance Mr. Mubashar Hussain Shah

In fuzzy we generally get MPCI. This is one output after defuzzification.

The solution of your problem is rather simple. If each of m outputs depends on all n inputs create m separate fuzzy models for each of the m outputs. Models with one output are, in the general case, more precise than multi-output models.

Greetings- AP

How many membership functions do you have for each input?.

Remember, if you have m inputs and n membership functions (lingustic values) for each input, there are at most n^m possible rules (outputs). It is, all possible combinations of premise lingustic values for m inputs

It sounds like you are speaking about variable inputs versus variable outputs, in that case the answer is yes: Imagine you have 3 inputs being related to weather, like temperature, humidity and atmospheric pressure. And you want to model, with that 3 inputs, a lot of different possible variable outputs, from vegetable ripening speed, possibility of fog, athletes tiredness rate, oxidation rate of certain material, fungus growth, etc, etc, etc. Sounds right? Yes. We can have more output variables than inputs, if data is available. The membership functions may vary. If your problem sounds like this example you may be able to still model it.

Thank you, Ms. Virginia Corral for your suggestion.