nmrval returning larger matrix than expected.

I've done a simple logstic regression for one predictor varilable (X) (a continuous set oberseved measruements) and one repsonse variable (Y) (fail or sucess taken for each oberserved measurement).

So without loads of detail, you can see this is a simple problem and the regression should take the form of

Why is it when i run mnrval(B,X) i get n-by-p+1 matrix back, instead of a n-by-p matrix?

Shouldn't nmrval(B,X) just return a 1 column vector for my problem as there is only only one predictor varible?

Thank you for you're help,

im very aware its likely a stupid question.

Answer 1

[Steven Lord]

Does your X matrix contain a column of 1s? If so see the Note in the first item of the Description section on the documentation page for the mnrval function.

[Daniel Theobald]

It is not a column of 1s, i did clock that in the documentation page.

It is actually a column of values between 0 and 1 (probabilities).

The weird thing is that if i add (:,1) and (:,2) i get exactly a soltuion of 1s.

This means that one column is prob and the other is prob-1, i just dont know which is which.

Strange right?

Thanks!