As a guest user you are not logged in or recognized by your IP address. You have
access to the Front Matter, Abstracts, Author Index, Subject Index and the full
text of Open Access publications.
Analogy-based estimation is a widely adopted method in software cost estimation that identifies analogous projects to the one under estimation and uses their data to derive an estimate, i.e. it is a Case Based Reasoning approach. The similarity measures between pairs of projects are critical for identifying the most appropriate historical data from which the estimation will be generated. Usually the similarity measures are selected empirically, using jackknife-like procedures. Typically, the measures that identify the most similar projects in most of the cases are considered the most appropriate ones and are applied in every new estimation procedure. However there are situations that the default similarity measures may not be the most appropriate ones. In this study we determine the situations in which the default parameters are not the best and we propose the similarity measures for these cases. In particular we provide rules that point out which projects are not accurately estimated with the default parameters.