topic Re: Big Data course - predicted and residual in SAS Academy for Data Science

Big Data course - predicted and residual

odesh — Tue, 09 Jul 2019 22:22:57 GMT

Hello,

I do not follow why the predicted values for 2 observations must be the mean. ( see attached)

Please help me understand.

Thanks.

Odesh.

Re: Big Data course - predicted and residual

Reeza — Wed, 10 Jul 2019 01:35:11 GMT

Without any other context this cannot be answered.

Re: Big Data course - predicted and residual

odesh — Wed, 10 Jul 2019 13:30:22 GMT

Thanks Reeza. I think that you are right. The context is presented in the Big Data course , Statistics component, ANOVA section.

Odesh.

Re: Big Data course - predicted and residual

Cynthia_sas — Wed, 10 Jul 2019 14:01:38 GMT

Hi:

Here's the context -- the statement on this slide from the Statistics 1 class:

I've asked the course instructors to take a look at this posting and comment.

Thanks,

Cynthia

Re: Big Data course - predicted and residual

MarcHuber — Wed, 10 Jul 2019 14:21:43 GMT

This is the context for this question. The table in question is the output predicted values from an ANOVA problem. In ANOVA (ANalysis Of VAriance, of course), the so-called "predicted value for any observation in the sample is the calculated arithmetic mean of the target variable (in this case, the target is TCellCount) for the group that the observation belongs to (group being the Medication value). So, in this example, each and every Predicted TCellCount for Medication 2 must be the same - in this case 905, just as each and every Predicted TCellCount for Medication 1 must be the same - in this case 795, and each and every Predicted TCellCount for Medication 3 must be the same - in this case 615. This MUST be the case because this is one-way ANOVA (ANOVA with only one classification variable as a predictor).

Re: Big Data course - predicted and residual

MarcHuber — Wed, 10 Jul 2019 14:25:03 GMT

Just to be clear, in my previous reply, I should have mentioned that 905 MUST be the group mean for Medication 2, 795 MUST be the group mean for Medication 1, and 615 MUST be the group mean for medication 3. It must be because predicted values in one-way ANOVA are the group means. And this is a listing of an output table from a one-way ANOVA.

Re: Big Data course - predicted and residual

odesh — Wed, 10 Jul 2019 16:09:19 GMT

Thanks very much to everyone for helping me understand. The key piece that
I was missing was that " It must be because predicted values in one-way
ANOVA are the group means. "

Odesh.

Re: Big Data course - predicted and residual

Reeza — Wed, 10 Jul 2019 16:32:43 GMT

lol,

Re: Big Data course - predicted and residual

odesh — Wed, 10 Jul 2019 16:39:19 GMT

I know Reeza. I had forgotten that basic fact.

Re: Big Data course - predicted and residual

Reeza — Wed, 10 Jul 2019 16:45:26 GMT

Sorry, that was submitted early and was not directed towards you! I was actually thinking how as you progress in your career it's easy to forget the basics sometimes! I have a Masters degree in Statistics and didn't remember that 😞