COCOMOST is a computer program for use in estimating software development costs. The goal in the development of COCOMOST was to increase estimation accuracy in three ways: (1) develop a set of sensitivity software tools that return not only estimates of costs but also the estimation error; (2) using the sensitivity software tools, precisely define the quantities of data needed to adequately tune cost estimation models; and (3) build a repository of software-cost-estimation information that NASA managers can retrieve to improve the estimates of costs of developing software for their project (see figure).

Example Model Output

COCOMOST implements a methodology, called "2cee," in which a unique combination of well-known pre-existing data-mining and software-development-effort- estimation techniques are used to increase the accuracy of estimates. COCOMOST utilizes multiple models to analyze historical data pertaining to software-development projects and performs an exhaustive data-mining search over the space of model parameters to improve the performances of effort-estimation models. Thus, it is possible to both calibrate and generate estimates at the same time. COCOMOST is written in the C language for execution in the UNIX operating system.

This program was written by Tim Menzies and Dan Baker of West Virginia University and Jairus Hihn and Karen Lum of Caltech for NASA's Jet Propulsion Laboratory.

This software is available for commercial licensing. Please contact Karina Edmonds of the California Institute of Technology at (626) 395-2322. Refer to NPO-44858.

This Brief includes a Technical Support Package (TSP).
Estimating Software-Development Costs With Greater Accuracy

(reference NPO-44858) is currently available for download from the TSP library.

Don't have an account? Sign up here.