Chinese Journal of Computers   Full Text
Title	Predicting Software Stage Effort with Sequence Changing Ratio
Authors	WANG Yong SONG Qin-Bao SHEN Jun-Yi
Address	(Department of Computer Science and Technology, Xi’an Jiaotong University， Xi’an 710049)
Year	2009
Issue	No.7(1346—1355)
Abstract &   Background	Abstract Software stage effort has the features of data starvation and uncertainty. It is difficult to use the current methods (e.g. regression) to make predictions. This paper proposes a novel prediction method, which gets the effort sequence changing feature—“changing ratio” from the completed stage effort sequences, and gets the “changing ratio threshold” from historical projects by machine learning methods, then uses grey models to make predictions. The experimental results on 10 real world software engineering datasets show that, compared with linear regression method, the prediction accuracy of the proposed method has been improved by 20%~80%. This is very encouraging and indicates that the method has considerable potential. Keywords project cost prediction; software cost; stage cost; grey model Background The software stage effort estimation plays an important role in the area of the software project management. There are about 75% of the projects all over the world overrun the schedules because of the inaccurate software cost prediction. In practice the software projects are often out of control, can’t be finished on budget and with poor quality for the absence of effective project cost prediction, and it is difficult to enter the international software market. Although there are several software cost prediction methods, however, no one method is consistently effective. Without considered the special characters of software development process, these methods can not deal with the uncertain data, and they require a large data samples with certain distribution. The Grey System Theory (GST) based on the uncertainty of small samples, doesn’t require a typical distribution, and has been successfully and widely applied in many areas. Prof. Song, one of the authors of this paper, and Prof. Shepperd et al. have used this theory successfully addressed the whole project cost estimation problem. Encouraged by their job, in this research the authors focus on the software stage effort evolution and estimation questions during the development process by the GST. The experimental results show that, compared with the current method, the prediction accuracy of the proposed method has been improved by 20%~80%. The research is supported by the National Natural Science Foundation of China (60673124， 61673087， 90718024) and the National High Technology Research and Development Program (863 Program) of China (2006AA01Z183).