Author(s)

C.W. Dawson

Abstract

One of the major problems with software development projects is that it is extremely difficult to accurately estimate their likely cost, duration and effort requirements. This invariably leads to problems in project management and control. Part of the problem is that during the early stages of these projects, very little is known about the problem domain, and initial estimates tend to be best guesses by a project manager. Artificial neural networks appear well suited to problems of this nature as they can be trained to understand the explicit and inexplicit factors that drive a software project’s cost. For this reason, artificial neural networks were investigated as a potential tool to improve software project effort estimation using project data supplied by BT. In addition, in order to deal with uncertainties that ex

Keywords