|
Activity Number:
|
415
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
| Abstract - #304943 |
|
Title:
|
Determining the Future Rate of Poisson Random Variables After Removing Variables with Too Few or Too Many Occurrences
|
|
Author(s):
|
Matthew Lindsey and Kellie Keeling and Robert Pavur*+
|
|
Companies:
|
The University of Texas at Tyler and University of Denver and University of North Texas
|
|
Address:
|
College of Business, Denton, TX, 76203-5249,
|
|
Keywords:
|
Poisson process ; Future Rate of Occurrence
|
|
Abstract:
|
What is the rate of occurrences of the sum of remaining Poisson random variables after variables with observed values above a set threshold or below a certain threshold are removed? Applied to software reliability, the underlying problem can be stated: "What is the reliability of a software package after bugs having a certain observed error rate are removed?" This study investigates the distribution of a standardized estimator of the future rate of occurrences after observing occurrences over a fixed time. Unbiased estimates are proposed for the future rate of certain subsets of Poisson random variables and the squared difference of this estimate and the true future rate based on Ross (1985). The advantage of the proposed estimation procedure to the traditional approach in the case that observed values above a set threshold or below a certain threshold are removed is investigated.
|
