Hi,
I am exploring the ordinary regression formula in splus. When I give an
unclean dataset ( with NA) I can enter a command na.action=na.exclude and
it runs the regression and gives me some parameter fit. However when I
clean the data so that any row with an NA gets removed from the dataset and
then run the regression without the paramter na.action=na.exclude, the fit
produces different coefficients.
Can someone explain me why? On the face of it they appear to be equivalent
operations.
Best Regards
Arjun Bhandari
Quantitative Investment Analyst
Equities Europe Internal
Abu Dhabi Investment Authority
P O Box 3600
Sheikh Sultan Building
Corniche Road
Abu Dhabi.
Tel : +9712 626 2689 ( Dir )
Tel : +9712 626 6500 Ext: 2936
Cell : +97150 491 8457
Fax :+9712 626 4616
Office Hours:
Abu Dhabi Time - Time Diff 3/4 hrs
Monday - Thursday 9am to 2pm & 5pm to 8pm -
Sunday: 8 am to 2 pm
********************************************************************************
The contents of this mail are personal opinions of the Author.
ADIA disclaims all responsibility and accepts no liability, whatsoever.
********************************************************************************
|