I am trying to fit a GLM model to my data (7 million rows, 153 variables) using R. More precisely I am using the Revoscaler package, but I suppose my issue would apply to other software as well.

The calculation results in an error, saying that the GLM model is singular. I tried to find an explanation, without success.

Could someone explain me what it means that the model is singular, and what needs to be done to solve the problem?


Singularity refers to the covariance matrix for the covariates. A matrix is singular when at least one variable can be expressed as an exact linear combination of some of the others. The best thing to do is to find which variables are creating the singularity and remove at least one of the variables from the model.

  • $\begingroup$ Thanks. How could I test linear combinations? $\endgroup$ – Benoit_Plante Aug 14 '12 at 1:46
  • 3
    $\begingroup$ You should be able to figure it out from the covariance matrix. But another way would be to regress a set of variables on one of the covariates. When you get an R square =1 you will have the linear combination. $\endgroup$ – Michael R. Chernick Aug 14 '12 at 1:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.