[Eeglablist] Question regarding regression ( GLM and Logistic)

[Iman Mohammad-Rezazadeh]

Hi  ,
Anyone knows in generating a GLM/ logistic regression model using both categorical and non-categorical data:

  1.   how should I treat with coefficient with insignificant p-values?  Should I include them in the model? And how should I interpret the model with insignificant coefficient ( even after stepwise process) ?
  2.  If I want to see their effect size then should I standardize (zscore) the predictors first?
  3.  If I have unbalanced data for training a logistic regression model ( suppose in my database, I have 1000 pairs for output category '1' and only 50 data for output category '2')  then how should I select the training data size?

Best,
Iman


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20170616/aa1ca897/attachment.html>