http://www.sciencedirect.com/science/article/pii/S0165027005001172

you have increase the number of iterations and select correct parameter values. It will help to improve your efficiency.  

You can increase the number of membership function or rule, and selection input parameters. and you can combine the ANFIS with another soft computing it will be improved your efficiency

Dear Furat,

I used classification tree for your data and results are:

Error for training data - 378%

Error for test data - 13,85%

I did not optimize tree parameters (no crossvalidation, no pruning).

You can compare results with ANFIS.  

and check:

Inputs=xlsread('data.xlsx',1,'A1:c608') ;

Targets=xlsread('data.xlsx',1,'p1:p608') ;

Test=xlsread('data.xlsx',1,'A1:c260') ;

TestTarg=xlsread('data.xlsx',2,'p1:p260');

It should be:

Test=xlsread('data.xlsx',2,'A1:c260') ;

I guess.