A Gompertz Model for Software Cost Estimation 


Vol. 15,  No. 2, pp. 207-212, Apr.  2008
10.3745/KIPSTD.2008.15.2.207


PDF
  Abstract

This paper evaluates software cost estimation models, and presents the most suitable model. First, we transformed a relevant model into variables to make in linear. Second, we evaluated model's performance considering how much suitable the cost data of the actual development software was. In the stage of model performance evaluation criteria, we used MMRE which is the relative error concept rather than the absolute error. Existing software cost estimation model follows Weibull, Gamma, and Rayleigh function. In this paper, Gompertz function model is suggested which is a kind of growth curve. Additionally, we verify the compatability of other different growth curves. As a result of evaluation of model's performance, Gompertz function was considered to be the most suitable for the cost estimation model.

  Statistics


  Cite this article

[IEEE Style]

S. U. Lee, "A Gompertz Model for Software Cost Estimation," The KIPS Transactions:PartD, vol. 15, no. 2, pp. 207-212, 2008. DOI: 10.3745/KIPSTD.2008.15.2.207.

[ACM Style]

Sang Un Lee. 2008. A Gompertz Model for Software Cost Estimation. The KIPS Transactions:PartD, 15, 2, (2008), 207-212. DOI: 10.3745/KIPSTD.2008.15.2.207.