An Inverse Regression Method for Determining an Ensemble of State Error Vectors from a Covariance Matrix.
Abstract
A method is described to determine an ensemble of initial condition state vector errors for use in a Monte-Carlo type error model. The state vector errors are found from a given (N x N) error covarance matrix and are properly correlated with one another. The method used is an inverse regression scheme which randomly introduces a gaussion distribution of errors with determined variances about linear regression curves calculated from the given covariance Matrix. A FORTRAN coded implementation of the method is included. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 05, 1977
- Accession Number
- ADA054214
Entities
People
- Ronald S. Brunsvold
Organizations
- Naval Ordnance Laboratory