"Matrix close to singular or badly scaled" after using squeeze function for mnrfit

Hi all!

I am trying to use mnrfit on two matrices, but get the following error message:

> In mnrfit>nominalFit (line 455)
  In mnrfit (line 202)
  In samplecode(line 28) 
Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND =  4.731375e-21.

My matrix A has 1 column and 200 rows, and my matrix B has 300 columns and 200 rows. I think the problem may be due to how I define B:

B = squeeze(data(:,50,:));

Indeed data is a 3D matrix and I want it to be a 2D matrix.

Once I have B, I simply do:

C = mnrfit(B,A);

Any help would be very much appreciated Many thanks

[Jan]

It does not matter how you create B and if the values are coming from a 3D array. The value of A or B is the only problem. Can you provide the data? Does the documentation explain, why the problem occurres?

[Lulu Dulac]

Thank you a lot for your answer. The full error message is

>> samplecode
> In mnrfit>nominalFit (line 455)
  In mnrfit (line 202)
  In samplecode (line 28) 
Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND =  6.968113e-21. 
> In mnrfit>nominalFit (line 455)
  In mnrfit (line 202)
  In samplecode(line 28) 
Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND =  4.731375e-21. 
Operation terminated by user during mnrfit>nominalFit (line 449)

In mnrfit (line 202)
        [b,hess,pi] = nominalFit(x,y,m,pi,n,k,p,pstar,parallel);

In samplecode(line 28)
C = mnrfit(B,A);

I am not sure how to provide the values as there are so many, but for A they range from -3 to 3 (e.g., -1,18217390027104 0,578412947454769 0,917299333541635 0,544267094172225 1,06462747652037 -1,69546501866701) and for B they range from 1 to 8 (only nondecimal numbers)

[Jan]

You can create a MAT file by save and attach it here in the forum.