[CentralOH] yesterdays statistics question

[Bryan Harris]

As I understand it, the T-Test tells you whether two populations with
normal distributions have a different mean.  If you plot the prediction
errors do they seem to have a normal distribution?  The T-Test checks the
hypothesis:  "These populations have different means."

You probably also ought to do an F-Test.  The F-test checks the hypothesis:
"These populations have different variances."

If you have reasonable sample sizes, and both populations are reasonably
normal (they have a bell curve) then these tests work OK.

Are you doing this for work or pleasure?

Bryan Harris, PE
Research Engineer
Electro-Mechanical Systems
University of Dayton Research Institute
bryan.harris at udri.udayton.edu
http://www.udri.udayton.edu/
(937) 229-5561

On Tue, May 17, 2016 at 9:46 AM, Eric Miller <miller.eric.t at gmail.com>
wrote:

> http://stats.stackexchange.com/questions/78079/confidence-interval-of-rmse
>
>
>
> On Tue, May 17, 2016 at 9:46 AM, Eric Miller <miller.eric.t at gmail.com>
> wrote:
>
>> I just happen to be sitting next to a real data scientist.  Shared your
>> problem w him, and he responded w this:
>>
>>
>
> _______________________________________________
> CentralOH mailing list
> CentralOH at python.org
> https://mail.python.org/mailman/listinfo/centraloh
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/centraloh/attachments/20160517/93c63491/attachment.html>