Currently, if I exclude some rows in a data set and then run partition, the analysis automatically assigns the excluded rows as a validation set. Even if I have a validation proportion identified, the excluded rows seem to take priority.
I'm having trouble reproducing this. Can you provide some more detail on exactly what's happening for you? What are you looking at to determine the validation rows?
Here's what I tried. I opened the Boston Housing data and excluded the first 9 rows. I also colored them red.
Now I launch Partition and specify a .10 validation proportion.
In the resulting report I don't see any red rows in the plot, meaning the first 9 rows are excluded and the validation portion show 55 rows.
If I go to the All Rows node and click the red triangle and choose Select Rows, I can see in the data table that the training rows (the selected ones) and the 55 other validation rows (the unselected ones).
