Hey all,
I have the following table: data(pnum text, distance float8, route text).
I would like to remove the outliers in distance, i.e. lets say i get
the avg dist of pnum for each route and the std deviation of the
distance what is the best way to identify the outliers?
Rhys.
--
Sent via pgs
Im asking how to get the ones that dont fall near the avg so for
example lets say i have the following distances:
10,11,12,11,10,9,9,10,11,12,10,11,99
then 99 would be an outlier. the avg would be like 16 or 17 i reckon
with the 99. so i want a way to find aan outlier, remove it and then
recal
Rhys A.D. Stewart escribió:
> i did some seraching about outliers and most of hits are about R or
> spss or some other statistical package.so looking for a way to do
> it wholly in pgsql.
Well, then, maybe PL/R?
--
Alvaro Herrerahttp://www.CommandPrompt.com/
Rhys A.D. Stewart wrote:
Im asking how to get the ones that dont fall near the avg so for
example lets say i have the following distances:
10,11,12,11,10,9,9,10,11,12,10,11,99
then 99 would be an outlier. the avg would be like 16 or 17 i reckon
with the 99. so i want a way to find aan outlie
Rhys A.D. Stewart wrote:
Hey all,
I have the following table: data(pnum text, distance float8, route text).
I would like to remove the outliers in distance, i.e. lets say i get
the avg dist of pnum for each route and the std deviation of the
distance what is the best way to identify the outliers
> Rhys A.D. Stewart wrote:
> >I would like to remove the outliers in distance
As others have said; an "outlier" is normally a human call and not
something that's generally valid to do automatically. The operator
would probably want to go in and look to see why it's that far out and
either fix the
>
> I'd agree, stddev is probably best and the following should do
> something
> reasonable for what the OP was asking:
>
> SELECT d.*
> FROM data d, (
> SELECT avg(distance), stddev(distance) FROM data) x
> WHERE abs(d.distance - x.avg) < x.stddev * 2;
>
[Spotts, Christopher]
Statis