@rawagah wrote:
Dears,
I have a dataset with no missing values or negative values but when I run HPGLM node in EM the number of observations used is 50% of the number of observations read.
How can I fix that?
As a minimum you should post the code you used. Best is to post the code along with the log, copy both and paste into a code box opened on the forum with the {I} icon.
You may have specified an option that restricted values or there might be something such as values are not missing but an operation may have a restricted domain such a LOG doesn't like negative or 0 values and excludes records based on that domain.