A Method for Selecting Fit Data in Software Effort Estimation
Koji Toda, Akito Monden, Kenichi Matsumoto · 2009
To construct a better multivariate regression model for software effort estimation, this paper proposes a method to automatically select (or not to select) projects as a fit data from a given project data set based on estimation target’s features. As a result of an experimental evaluation using the ISBSG data set, the proposed method showed better estimation performance than the conventional method (of constructing a regression model using all project data). The median of MRE (Magnitude of Relative Error) was improved from 0.452 to 0.367, and the median of MER (Magnitude of Error Relative) was improved from 0.357 to 0.336. This paper showed the necessity of fit data selection, and showed that the proposed method was one of the effective and systematic meant to do the selection. 1. はじめに