A methodology for improving the accuracy of ongoing software development effort prediction

Abstract

Poor project scheduling can lead to software project failure. Schedule overrun can be caused by the additional effort in entertaining software requirement change, which is an unavoidable activity in a real software development environment. When a change is applied to a partially developed artefact, it will incur more effort and cost than if the change is done on a non -developing artefacts. This paper presents the methodology proposed to improve effort prediction accuracy when a software project manager is confronted with software change requests during the course of a software project development. Careful decisions have to be made based on the int egration of a few effort prediction models. The methodology discussed is organized in such a way that it can be used to predict the effort due to software change request during an ongoing software project and that it can be reproduced by anyone who wish to perform the same prediction.