The purpose is to try to classify the samples to each control group.
What I did is using MSExcel chart to display these sample and control
values in a line chart so I know where sample data point relative to
control data point, so I decide each sample belongs to which control.
(this is a classification with a single variable)

But it is kind of arbitrary in the above process. So I wonder if there
is a statistic to indicate the confidence /likelihood of my
classificatin for each sample. For example. which is better, consider
1.23 as control type: 0.8 or as control type 1.7?


"Anon." <[EMAIL PROTECTED]> wrote in message news:<[EMAIL PROTECTED]>...
> Henry wrote:
> > I have a set of sample results and controls for 3 levels. 
> > 
> > Control values like : 0.8; 1.7; 2.5
> > sample values: 0.45, .75, 0.9, 1.2, 1.23, 1.35, 1.45, 1.55, 1.54,
> > 1.78, 1.89, 1.98, 2.01, 2.33, 2.45, 2.65, 2.8, 3.0
> > 
> > In practics, I simply based on each sample data point position
> > relative to control in Excel chart to classify them to a control
> > group. 
> 
> I'm sorry, you'll have to re.write this - I can't parse the sentence, 
> and I suspect few others can too.  I can make a guess about what you're 
> asking (in which case, there is a solution to your problem), but I'd 
> rather wait to be confused later.
> 
> I wonder if there is any statistical parameter I can compute to
> > support my claim that 1.23 should be more like control of 0.8, but not
> > control of 1.7?  is  there some kind of likelihood or probabilty to be
> > computed?
> > 
> > Seems silly, but it does happen a lot in reality.
> 
> Sounds like life.
> 
> Bob
> 
> -- 
> Bob O'Hara
> 
> Rolf Nevanlinna Institute
> P.O. Box 4 (Yliopistonkatu 5)
> FIN-00014 University of Helsinki
> Finland
> Telephone: +358-9-191 23743
> Mobile: +358 50 599 0540
> Fax:  +358-9-191 22 779
> WWW:  http://www.RNI.Helsinki.FI/~boh/
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to