Calibrating Function Complexity in Enhancement Project for Improving Function Points Analysis Estimation

Abstract

Producing a good software product on time and within budget, the initial software estimation takes a significant role. Reality has shown that most software fails because the initial software estimation is not correct. Many researchers have proposed methods for software estimation. It has been developed since the 70s of the last century, but it is still of great interest until now. We also know that creating a new software product is difficult; it is even more difficult to innovate. In the framework of this paper, we propose an improved method based on the FPA method of IFPUG. We named this proposed model is Calibrating Function Complexity in Enhancement Project (CFCEP). This method is based on the Linear Regression technique to give coefficients of function complexity. The experimental results based on the ISBSG dataset show that the estimation based on this new coefficient gives much better results than using the coefficients of the standard FPA method.