|
Activity Number:
|
422
- Statistical Learning for Functional Data
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2018 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Government Statistics Section
|
|
Abstract #329495
|
Presentation
|
|
Title:
|
Regression Based Circular Error Probable: An Application to Ballistic Systems
|
|
Author(s):
|
Zachary Zimmer* and Casey Turner
|
|
Companies:
|
and ATEC
|
|
Keywords:
|
Circular Error Probable;
Bivariate Normal Regression
|
|
Abstract:
|
Circular error probable (CEP) is a key metric for the analysis of performance for ballistic systems. CEP is defined as a circular region containing a specified percentile of the geolocation data collected (miss distances for projectile systems). Oftentimes the median is used but other percentiles such as the 90th are considered based on additional requirements or objectives. Most applications assume the data follows an iid bivariate normal distribution, although some papers have looked into deviations from this assumption such as assuming the data follows a bivariate normal mixture models with known and unknown mixing percentages. In a recent paper the CEP was computed from a bivariate normal regression model including computing confidence limits based on two bootstrap approaches. This approach will provide an additional method for computing CEP estimates when the data deviates from the standard iid data assumption. "This work is independent of the author's employment at Merck & Co. Inc."
|
Authors who are presenting talks have a * after their name.
