Regression Diagnostics in Practice -- Experiences from Modelling Jet Engine Costs,
Abstract
This paper describes how regression diagnostics were used to help develop revised cost-estimating relationships for jet engines. The goal was to derive meaningful, yet easy to use models based on an updated collection of few observations and many variables. First, specific criteria were established for selecting explanatory variables. A variety of numerical and graphical techniques were then used to critique candidate models by examining residuals and evaluating the influence of individual engines. The final models are not only intuitively satisfying, but generally provide better predictions and are easier to use than earlier models. Additionally, the user is provided with a greater understanding of the design and sensitivity of the models, and therefore a better understanding of the actual estimates.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1983
- Accession Number
- ADA154199
Entities
People
- J. B. Garfinkle
- J. L. Birkler
Organizations
- RAND Corporation