I've just started learning about fuzzy logic and I would greatly appreciate your help on this. I have conducted a survey to find the factors that influence cost overrun in residential buildings. The questions were qualitative on a scale of 1 to 6 (1=never...6=always). Then the RII for each factors were calculated and top factors were selected (5 factors).
I'm trying to do some risk assessment using fuzzy logic (I am hoping the end result would be a user input form), however, I am facing difficulties with how to find the membership function of the the 6-scale ranking. I've found some papers, but unfortunately, they don't explain how they did it.
If I wanted to use trapezoidal MF, how should that be written?
Should I have 5 MF for each factor? What would the parameters be?
How can I use the RII and the weights of each factor in the fuzzy approach?
Your help is greatly appreciated!
This is an incomplete answer proposal: first, could you please describe your problem a bit more, and what you have tried already?
To the question about the number of membership functions, without knowing more, I'd suggest (based on a feeling) that it heavily depends on factors. Do you have natural interpretations for the membership functions? (What have you tried this far?)
A paper A Fuzzy Decision Making Approach in Evaluating Ferry Service Quality seemed to contain much details.
Another paper also contained quite nice Measurement and fuzzy scales exposure: identify the problem, fuzzify inputs, construct rules, aggregation, defuzzify, sensitivity analysis.
Have you seen those & do they answer to some of your questions?