How to solve evalfis error?

Prashant Kumar answered . 2025-03-26 03:52:24

The error message indicates that the evalfis function is expecting the first argument to be a trained Fuzzy Inference System (FIS) structure, but it received something else. In your code, the issue lies with how anfis1 and anfis2 are defined or trained.

Let’s resolve this step-by-step:

1. Verify Training of anfis1 and anfis2

The anfis function should output a trained FIS structure if provided with valid data and options. In your case, check the anfis training process:

anfis1 = anfis(data1, opt);
anfis2 = anfis(data2, opt);

Make sure:

• data1 and data2 are valid training datasets with at least 2 inputs and 1 output (3 columns).
• opt.InitialFIS is properly defined. Setting it to 7 or 6 might be incorrect. Typically, you create an initial FIS structure using genfis or provide an existing FIS.

2. Fixing opt.InitialFIS

Instead of assigning numeric values (7 and 6), initialize the FIS structure using genfis. Here's an example:

% Generate an initial FIS for anfis1
fis1 = genfis(data1(:,1:2), data1(:,3)); % Inputs: X, Y; Output: THETA1
opt.InitialFIS = fis1;

anfis1 = anfis(data1, opt);

% Generate an initial FIS for anfis2
fis2 = genfis(data2(:,1:2), data2(:,3)); % Inputs: X, Y; Output: THETA2
opt.InitialFIS = fis2;

anfis2 = anfis(data2, opt);

This ensures that the anfis function has a valid FIS to start with.

3. Using evalfis

Once you have trained FIS structures (anfis1 and anfis2), you can use them in evalfis:

XY = [X(:) Y(:)];
THETA1P = evalfis(anfis1, XY); % Predict THETA1
THETA2P = evalfis(anfis2, XY); % Predict THETA2

4. General Troubleshooting

If you still face issues:

• Use disp or class to check the type of anfis1 and anfis2 after training:

disp(class(anfis1));

• If the class is not mamfis or sugfis (FIS structures), then training failed.

• Inspect the training datasets (data1 and data2) for correctness:

  • data1 and data2 must have at least three columns.
  • The first two columns are inputs (X, Y), and the third column is the corresponding output (THETA1 or THETA2).

By ensuring anfis1 and anfis2 are valid trained FIS structures, you should be able to use evalfis without errors. Let me know if you encounter any specific issues during these steps!

Not satisfied with the answer ?? ASK NOW