I tried to make a Logistic Regression in order to explain forest cover changes (Permanence, gain and loss) with four independent variables (Slope, Distance to water bodies, to roads and to settlements). I set the variables in the Logistic Reg window, but in the step 3 of 6 3%, the analisys seems to stop and a message appear: A singular matrix has ocurred. No meaningfull regression coefficients found. Adjust the combination of the independent variables.
What it mean this?
Hi Santiago,
This message is letting you know that your independent variables are correlated, which can result in a matrix that is singular. I recommend that you remove any variable that seems like it would be perfectly correlated with any of the other variables and try your logistic regression again. This may take some time and tinkering to determine which independent variables work best together. Please let us know if you continue to run into issues with this.
Best Regards,
Andrew Hostetler
Clark Labs Technical Support Team
Hi Andrew,
I have the same problem, i am using the data obtained in the tutorial on logistic regression, i eliminated any variable but the message is same.
What can I do?
Regards
William, you may want to go and redo your steps in the tutorial. There are many steps to create the independent variables and if these are incorrectly made you may get the result you did. When the steps are executed correctly, Logistic Regression runs correctly with or without the prediction series.
James Toledano
Clark Labs
Sorry ı have the same problem. ı did you say that but ı have same problem. How can ı fix this problem
Regards
how do we calculate the standard error and the significance levels of the regression coefficients with IDRISI?
