Warning: Regression design matrix is rank deficient to within machine precision.

I'm getting this error when I'm trying to do a regression analysis with the function 'model_output = fitlim'. I read somewhere online that it may be due to too many variables (it's a lot in the thousands) but shouldn't Matlab be able to handle it? I'm very new to Matlab just this week so don't have a lot of experience with the software.

I tried to only using about 30 variables and then the function worked but I absolutely need to be able to do with all of my variables. The variables originate from Excel sheets which I only have done log function on beforehand.

[Jan]

Did you consider that the warning can be serious? If there is a problem with the rank of the matrix, than you might use the wrong approach to solve the problem. But perhaps a scalaing would solve the problem. Without knowing any details, this is hard to decide.

[Erik Öhrner]

A warning is something I would consider serious absolutely. As I said, I have little experience with Matlab. I'm a student and trying to use code which my supervisor recommended but maybe we will have to revise and use something different then.

[dpb]

As Jan alludes to, the problem here is likely not Matlab but the underlying data. How many variables versus how many observations do you have? Does that tell you something?

[Abhinav Gaur]

Is it necessary to normalize all the design variables between 0 and 1 before trying to use fitglm?