Dear aart,

I have worked with ANFIS by more than 6 inputs and have encountered NO limitation. Do you use MATLAB toolbox for ANFIS implementation?

Dear aart,

I have worked with ANFIS by more than 6 inputs and have encountered NO limitation. Do you use MATLAB toolbox for ANFIS implementation?

Hi,

Thanks for replying. 

I am using the command line. What do you use?

please following this page:

http://www.cs.nthu.edu.tw/~jang/anfisfaq.htm

There are some very good discussions on ANFIS and its limitation in the following link:

https://www.researchgate.net/post/What_are_the_differences_between_genfis1_and_genfis2_when_used_in_Matlabs_ANFIS?_tpcectx=profile_questions

I have used ANFIS, I am not sure that 6 entries is a maximum, I worked with more than 6 entry parameters and it worked.

Aarti,

I use with MATLAB toolbox the ANFIS. I never worked with Mamdani but usually use Sugeno in my engineering applications. I do not see this limitation. What do you use?

You have two options, toolbox or command line. You can start with tool box to check the performance of your algorithm with visualization, then you can apply it with command line. Anyhow, there is no limitation as a tool. But you need to check you field and the problem it self, if  it applicable with this tool. Good luck

Thank you everyone,

I am using command line. But as per your suggestions, I tried to explore the ANFIS toolbox. I found that it can take only one output but many inputs. Am I correct?

The problem with which I am dealing is having around 37 inputs and 8 outputs. Can I use the toolbox now?

Due to the nature of the fuzzy rules under consideration, an ANFIS can only have one output, which limits its applicability to problems with one solution per time. Also, there are only two choices for the output membership function: constant and linear. This limitation of output membership function choices is because anfis only operates on Sugeno-type systems. For your case it is better to use ANN, instead.  

I don't remember the details right now, since I worked on Fuzzy Logic long time ago. However, if there is really a limit in the number of outputs to be just one, you can construct 8 parallel ANFIS systems each having 1 (different) output but 37 (same) inputs.

As far as the ANN is concerned, it can help you a great deal only if you have got sufficient dataset for all the combinations of the values of inputs and the corresponding values of outputs.

Hi,

We have solved the same problem in this paper:

A novel approach for ANFIS modelling based on Grey system theory for thermal error compensation

https://www.researchgate.net/publication/265508973_A_novel_approach_for_ANFIS_modelling_based_on_Grey_system_theory_for_thermal_error_compensation?ev=prf_pub

Conference Paper A novel approach for ANFIS modelling based on Grey system th...

Hi,

in the ANFIS toolbox, first select 'edit' at the top of page

then 'FIS properties'

then 'edit'

then 'add variable'

select the type of variable 'input or output'

your aim is accomplished

If you use Matlab toolbox there no limitation on the number of Inputs.

Anfis has limitations due to reasons attributed to the number of parameters. For this reason, if the number of parameters is high, calculation problems are encountered. You can reduce the size of the dataset with the recommended size reduction methods.

Deep Learning Autoencoders

Principle Component Analysis

so on

For multiple output case, you can try CANFIS- Coactive Neuro Fuzzy Inference systems.

Increasin the number of columns in the data provided in the function in matlab

This means the increasing of input neurons in the system

If you use the user interface of Matlab is easier