A system for supporting project effort planning to make more accurate cost estimates

Abstract

The purpose of this study was to evaluate if system support helps project leaders reduce non-billable hours in effort estimation during offer creation. In effort estimation, many practitioners still rely on expertise-based learning without system support. Various methods to support effort estimation were investigated in the last decades, but the majority of the recent investigations focus on automatic learning oriented approaches, such as articial neural networks. However, learning-oriented approaches are usually not transparent enough to gain the trust of a project leader. Therefore, within this study, learning-oriented methods were combined with expertise-based learning by involving the project leader in the process. While integrating project leaders in the case-based reasoning process, learned recommendations are proposed to transparently support effort estimation. The project features used for learning recommendations should be independent of an industry (e.g. software projects)....