Readiness Evaluations Using Multivariate Data Reduction
Abstract
Several statistical methods--principal component analysis, orthogonal factor analysis, classification, and clustering techniques--are tailored and combined into a system designed to digest high-dimensional vectors of data on operational readiness of Navy ships. Such data consist of large numbers of scores for individual ships assigned by experts. The purpose of the data reduction system is to provide a robust method of representing the data by a small number of scores that are meaningfully related to the original scores and that allow classification and clustering of the ships into homogeneous groups on relevant readiness scales. Simulated data drawn from mixtures of specified multivariate normal populations have been used to test the ability of the system to recover individual populations and to detect trends over time.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1978
- Accession Number
- ADA063758
Entities
People
- S. Zacks
- W. H. Marlow
- Zeev Barzily
Organizations
- George Washington University