Software Development Effort Estimation using CBR: A Review

Abstract

The software community acknowledges that timely, accurate estimates of development effort are important to the success of major software projects. Estimates are used when tendering bids, for evaluating risk, resource scheduling and progress monitoring. Inaccurate estimates have implications for all these activities, and so, ultimately, the success or failure of a project. To date most work has focused upon building algorithmic models of effort, for example COCOMO (Constructive Cost Model). These can be calibrated to local environment. This paper presents an alternative approach to estimation based upon the use of analogies (or CBR). The underlying principle is to characterize projects in terms of features (e.g. number of interfaces, the development method etc.). Completed projects are stored and then the problem becomes one of finding the most similar projects to the one for which prediction is required. Similarity is defined as Euclidean distance in n-dimensional space where n is the number of project features. Each dimension is standardized so all dimensions have equal weight. The known effort values of the nearest neighbors to the new project are then used as the basis for the prediction. Our analysis is based on IBM dataset of 24 projects. Our results show that estimation by analogy is a viable technique that, at the very least, can be used by the project mangers to complement current estimation techniques.