regress function error for matrix multiplication

Hi,

I tried to regress stock returns on market returns and I get this error:

Applying the function '@(x,y){regress(y,x)}' to the 10562nd group of data generated the following error:

Incorrect dimensions for matrix multiplication. Check that the number of columns in the first matrix matches the number of rows

in the second matrix. To perform elementwise multiplication, use '.*'.

This is my code.

%% creating variables
A = [ones(size(subset1.MRET)), subset1.MRET];
b = subset1.RET;

%% run regression 
[group5, years4, ID3] = findgroups(subset1.years, subset1.ID);
m1 = splitapply(@(x,y) {regress(y,x)}, A, b, group4);

A is a 64121190x2 Matrix

The first column are ones to include the constant and the second column are the Marketreturns.

b is a 64121190x1 Vector

With the stock returns in it.

Does somone know how to deal with this error.

It worked for me, when i used a subset of only 2 stocks.

The dataset is unfortunaltly to big to ubload.

Thank you in advance

[Shashank Gupta]

For some reason, I am not able to regenerate the error which you are getting, but by the look of it, it seems like some kind of mismatch in input dimension provided to the function. I can further investigate if you can provide me "subset1.Years" and "subset1.ID" may be.

[Luca]

Thank you for your answer. I found the problem there were to many NaN in some years to run the regression.