Improving Analogy-Based Software Cost Estimation through Probabilistic-Based Similarity Measures.
Abstract
The performance of software cost estimation based on analogy reasoning depends upon the measures that specifying the similarity between software projects. This paper empirically investigates the use of probabilistic-based distance functions to improve the similarity measurement. The probabilistic-based distance functions are considerably more robust, because they collect the implicit correlation between the occurrences of project feature attributes. This information gain enables the constructed estimation model to be more concise and comprehensible. The study compares 6 probabilistic-based distance functions against the commonlyused Euclidian distance. We empirically evaluate the implemented cost estimation model using 5 real-world datasets collected from the PROMISE repository. The result shows a significant improvement in terms of error reduction, that implies an estimation based on probabilistic-based distance functions achieve higher accuracy on average, and the peak performance significantly outperforms the Euclidian distance based on Wilcoxon signed-rank test. Keywords-Software Cost Estimation, Analogy, Heterogeneous Distance Function, Machine Learning I. I NTRODUCTION Accurate software development cost estimation is vital for software development. Unreliable cost model can affect the software product quality or even worse as causing project cancellation. Cost estimation therefore is an important area in software engineering research. For decades, much effort has been given to improve the estimation accuracy. The cost estimation methods most possibly encountered in the literatures are analogy-based reasoning estimation (ABE) and regression-based model construction. They have shared a proportional interest, thus; up to now, numerous attempts were failed to define the best solution, as many research conclusions have produced conflicts [1], [2]. The essential hypothesis of software cost estimation based on ABE reasoning is that a new software project should acquire an approximate cost from similar past projects. ABE estimation accuracy heavily depends on 5 parameters, which are feature subset selection, number of project analogs to build the model, similarity m