Development of an Adjustment Factor for Improving Software Effort Estimation Utilizing the Use Case Point Metho

This Praxis describes a method for improving the accuracy of Use Case Point (UCP) effort estimation for software development projects. The thesis for the research is that an emergent factor exists in software development projects that is not considered in the UCP process. This emergent factor is the primary cause of error in the labor hour estimation. While the detailed nature of the emergent factor is unknown, a method for determining the quantitative value of the factor and using it to generate a more accurate labor estimate can be developed.;The method involves the application of an adjustment factor model (AFM) to UCP estimates. The AFM was developed by conducting multiple regression analyses on a sample of 44 software development projects provided by three different development organizations. The regression model was then optimized further through the use of a Generalized Reduced Gradient non-linear optimization algorithm, yielding an Optimized AFM.;Both the AFM and Optimized AFM were tested for validity against a set of 27 independent software projects. The test results show that the proposed application of an adjustment factor to use case based estimates reduced the mean magnitude of relative error by 50% and the weighted mean quartile by as much as 62% for the projects in the evaluation dataset. Moreover, 26 of the 27 projects improved to within 25% of their estimate with the AFM applied and all 27 when the Optimized AFM was applied. The tests also indicate that the application of the adjustment factors improved estimation error in 82% of the projects contained in the evaluation dataset.