Software Project Effort Estimation Based on Multiple Parametric Models Generated Through Data Clustering

Parametric software effort estimation models usually consists ofonly a single mathematical relationship. With the advent of softwarerepositories containing data from heterogeneous projects, thesetypes of models suffer from poor adjustment and predictive accuracy.One possible way to alleviate this problem is the use of a set ofmathematical equations obtained through dividing of the historicalproject datasets according to different parameters into subdatasetscalled partitions. In turn, partitions are divided into clustersthat serve as a tool for more accurate models. In this paper, wedescribe the process, tool and results of such approach through acase study using a publicly available repository, ISBSG. Resultssuggest the adequacy of the technique as an extension of existingsingle-expression models without making the estimation process muchmore complex that uses a single estimation model. A tool to supportthe process is also presented.