Too Much Variables in Fuzzy Logic and ANFIS

Hi i have 13 variables and some of them has 5 membership. I can't define all rules (1 billion rules). How can i define that rules?

And my data has 16000 rows how can i train fis without memory problem?

P.S.I can only use some of GUI's so please keep it simple.

Thx

Answer 1

[Sam Chak]

Hi @YTU

With 13 inputs and 5 membership functions each, the total number of rules exceeds more than billion rules when using the grid partitioning technique to combine all possible rules. When setting up the ANFIS, consider using the Subtractive Clustering, a built-in method in ANFIS to identify the clusters of data, so that the rules can be expressed as follows:

Rule i: IF in1 & in2 & in3 & ... & in11 & in12 & in13, THEN output_i

where i = 1, 2, 3, ... 13.

This approach substantially reduces the total number of rules from 1.2 billion to only 13 rules.