Case Study
Regression for Cost Analysts
(The content of this paper is the opinion of Evin Stump, Sr. Systems Engineer at Galorath Incorporated)The question that regression methods can potentially answer for the cost analyst can be stated this way: Given a set of historical cost data, what relationships can be discovered that are potentially useful for estimating a future cost?1 The potential product of regression analysis in cost estimating is that the future may not be like the past. The second biggest obstacle is that the information about the past may be so sparse or so “noisy” that reliable clues about the regression analysis is a cost estimating relationship (CER). Typically a CER relates cost or a cost connected variable (such as labor hours) to parameters of the project’s product in the form of an algebraic equation. As an example, consider a project whose mission is to write computer software.
