|
Title:
|
ARE SOME INTERVALS BETTER THAN OTHERS? |
|
Author(s):
|
John Byrne |
|
ISBN:
|
972-8924-16-X |
|
Editors:
|
Pedro Isaías, Maggie McPherson and Frank Bannister |
|
Year:
|
2006 |
|
Edition:
|
2 |
|
Keywords:
|
Statistical techniques, discrete data, confidence interval, data mining. |
|
Type:
|
Short Paper |
|
First Page:
|
304 |
|
Last Page:
|
308 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
In this paper we consider the use of confidence intervals in decision making when the data is count data using the Poisson
distribution as our example. We make use of statistical techniques to produce a random confidence interval rather than
one based on considering percentiles as endpoints, which is a common technique in data mining. The paper explains why
the standard Poisson confidence interval may not be appropriate and also why the first normal approximation is totally
inappropriate. We review the literature on exact confidence intervals and suggest two appropriate approximate and
averaging intervals for consideration. Finally using a marketing example, we develop appropriate intervals and make
suggestions for circumstances when there are larger data sets. |
|
|
|
|
|
|
