Background. After the COSMIC Functional Size Measurement method was proposed, the issue of convertibility between traditional function points and COSMIC function points arose. Several studies have been performed, to evaluate whether a correlation between the two measures. Specifically, it has been suggested that the relation between the measures of FP and CFP may not be linear, i.e., it is a curve that increases its slope for bigger projects. Objective. This paper aims at verifying this hypothesis using the available datasets that collect both FP and CFP size measures. Method. Rigorous statistical analysis techniques are used. In particular regression analysis is used, and the conditions for its applicability are systematically applied. Results. It appears that the analysis of the correlation between FP and CFP measures based on non linear regression models yields different results for the various available datasets: non-linear models can be the best representation of the correlation between FP and CFP, but can also be just equivalent to linear models, or even hardly applicable. Conclusions. There is no evidence that the non-linear models represent well all the available datasets correlating FP and CFP measures. Both linear and non linear regressions should be applied to a given dataset in order to identify the best correlation of FP and CFP.

A Study of Non-linearity in the Statistical Convertibility of Function Points into COSMIC Function Points

LAVAZZA, LUIGI ANTONIO;MORASCA, SANDRO
2010-01-01

Abstract

Background. After the COSMIC Functional Size Measurement method was proposed, the issue of convertibility between traditional function points and COSMIC function points arose. Several studies have been performed, to evaluate whether a correlation between the two measures. Specifically, it has been suggested that the relation between the measures of FP and CFP may not be linear, i.e., it is a curve that increases its slope for bigger projects. Objective. This paper aims at verifying this hypothesis using the available datasets that collect both FP and CFP size measures. Method. Rigorous statistical analysis techniques are used. In particular regression analysis is used, and the conditions for its applicability are systematically applied. Results. It appears that the analysis of the correlation between FP and CFP measures based on non linear regression models yields different results for the various available datasets: non-linear models can be the best representation of the correlation between FP and CFP, but can also be just equivalent to linear models, or even hardly applicable. Conclusions. There is no evidence that the non-linear models represent well all the available datasets correlating FP and CFP measures. Both linear and non linear regressions should be applied to a given dataset in order to identify the best correlation of FP and CFP.
2010
Alain Abran, Cigdem Gencel, Ales Zivkovic
Workshop on Advances in Functional Size Measurement and Effort Estimation (FSM '10)
Workshop on Advances in Functional Size Measurement and Effort Estimation
Maribor (Slovenia)
21 giugno 2010
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11383/1717139
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? ND
social impact