Segmented parametric software estimation models: using the EM algorithm with the ISBSG 8 database
Miguel Garre, J. J. Cuadrado, Miguel‐Ángel Sicilia, M. Charro, Daniel Rodríguez · 2005
Abstract. Parametric software estimation models rely on the availability of historical project databases from which estimation models are derived. In the case of large project databases with data coming from heterogeneous sources, a single mathematical model cannot properly capture the diverse nature of the projects under consideration. In this paper, a clustering algorithm have been used as a tool to produce segmented models. A concrete case study using a modified EM algorithm is reported. In cases in which the clustering algorithm produces an statistical characterization of each of the resulting clusters, as happens with the EM algorithm, such representations can be used as similarity metrics to derive analogical estimates. This can be put in contrast with deriving partial parametric models for each cluster. The paper also provides a comparison of quality of adjustment of both approaches.