How can I write different ouput values for each for-loop?

I am trying to write the outputs in a text file using with the for loop and a function. After working the code in thee workspace I see the outputs for MAPE values at the each loop but in the text file it gets the one value it does not change. I use neural network function for MAPE values.How can I write the different MAPE values for each loop in the text file?

fileID=fopen('Donguu.txt','w+');
      for lrate=0.1:0.2:0.2
      for trainingrate=0.2:0.4:0.8
      for n1=3:2:7
      for n2=6:2:10
      NeuralNetwork(Input, Target, trainingrate, n1, n2, lrate);
fprintf(fileID, '%d, %d, %d, %d, %d\n', lrate, trainingrate, n1, n2, MAPE);
      end 
      end
      end
      end

lrate=6.000000e-01, trainingrate=2.000000e-01, n1=3, n2=6

MAPE =

    0.6875

lrate=6.000000e-01, trainingrate=2.000000e-01, n1=3, n2=8

MAPE =

    0.6354

lrate=6.000000e-01, trainingrate=2.000000e-01, n1=3, n2=10

MAPE =

    0.7353

[Walter Roberson]

You do not appear to be writing to a text file: your fprintf() does not have a file identifier as the first parameter, and you are not passing a file identifier to NeuralNetwork() to write to.

Are you using diary() to write to a text file?

[Perihan Bilgeç]

I edited the code, I am using the fprintf(fileID, '%d, %d, %d, %d, %d\n', lrate, trainingrate, n1, n2, MAPE) code for writing the text file and first row I am creating the text file. Without writing the text file on the command view I see different values for MAPE, but in the text file there must be a wrong thing.

[Geoff Hayes]

Perihan - do you close the file with fclose?

[Walter Roberson]

MAPE is not calculated by any of the code you posted. Is NeuralNetwork doing an assignin('caller', 'MAPE', ...) or is it simply displaying the value ?

[Perihan Bilgeç]

I closed the file but the result doesn't change. I added the code below.

fclose(fileID)

[Perihan Bilgeç]

Yes I calculated the MAPE value in the function that named NeuralNetwork. The function is below with input and output arguments.

function [net ye Yv MAPE R2] = NeuralNetwork(Input, Target, trainingrate, n1, n2, lrate)

Answer 1

[Walter Roberson]

You need to change

NeuralNetwork(Input, Target, trainingrate, n1, n2, lrate);

to

[~, `, ~, MAPE, ~] = NeuralNetwork(Input, Target, trainingrate, n1, n2, lrate);

[Perihan Bilgeç]

I tried and it is correct now. Thank you very much for the help. What is the wrong thing about the usage of NeuralNetwork(Input, Target, trainingrate, n1, n2, lrate)?

[Walter Roberson]

NeuralNetwork(Input, Target, trainingrate, n1, n2, lrate)

calculates an internal variable named MAPE but does not assign to MAPE in the workspace of the calling function.

Any assignment you make inside a function definition does not affect the calling function, except for one of the following circumstance:

• you declared the variable as global in all applicable places; or
• you are using nested functions with shared variables; or
• you use assignin('caller') -- and if you do then increasingly MATLAB might refuse to recognize the assignment unless you already had assigned to a variable of the same name
• you are using a handle object and setting a property of the handle object. handle objects act similar to pointers
• you are calling the function from the command line or from a script that is being invoked from the command line or being invoked as-if it was from the command line (callbacks specified as strings are invoked as-if from the command line), and you use assignin('base') -- and if you do then increasingly MATLAB might refuse to recognize the assignment unless you already had assigned to a variable of the same name

All this applies even for variables listed as output variables. For example,

function primus
  A = 5;
  secondus();
  disp(A)
function A = secondus()
  A = 22;

Then it will be 5 that is displayed, not 22, because the A inside secondus is not the same variable as the A inside primus. If primus() contained the line

A = secondus();

then A would change to 22, but not because the variable names happened to be the same: assignment is strictly positional, so the first output of secondus is assigned to the first location listed on the left side of the assignment.

Answer 2

[Perihan Bilgeç]

Thank you very much for detailed explanation.