A Study of Analogy Based Sampling for interval based cost estimation for software project management

Abstract

Software cost estimation is one of the most challenging activities in software project management. Since the software cost estimation affects almost all activities of software project development such as: biding, planning, and budgeting, the accurate estimation is very crucial to the success of software project management. However, due to the inherent uncertainties in the estimation process and other factors, the accurate estimates are often obtained with great difficulties. Therefore, it is safer to generate interval based estimates with a certain probability over them. In the literature, many approaches have been proposed for interval estimation. In this study, we propose a navel method namely Analogy Based Sampling (ABS) and compare ABS against the well established Bootstrapped Analogy Based Estimation (BABE) which is the only existing variant of analogy based method with the capability to generate interval predictions. The results and comparisons show that ABS could improve the performance of BABE with much higher efficiencies and more accurate interval predictions.