A new combinatorial framework for software services development effort estimation

Abstract

Since software services have become one of the most important parts of companies and organizations in recent years, estimation of the effort required for their development is a critical necessity. Increased accuracy of an estimation model is important for both industry and academia. Different methods and models have been proposed for estimating the effort, but none are suitable for all services and different types of data-sets. In fact, the performance of each model can be right or wrong based on conditions. In this paper, a new method is presented based on a combination of different estimation models and mathematical assembling of their answers. The new method was evaluated on two quite different real data-sets. The results show that the proposed method has improved the estimation accuracy and performance. The proposed method flexibly estimates the effort without any additional overhead. It is completely independent of methods employed in the combination so that any new method can be smoothly added.