Improving the COSMIC approximate sizing using the fuzzy logic EPCU model

Abstract

In software engineering, the standards for functional size measurement require, for accurate measurement results, that the functionality to be measured be fully known. Therefore, in the early phases of software development when there is a lack of details, approximate sizing approaches must be used instead of the standards themselves: such approximate sizing techniques are typically based on the analysis of historical data of the functional size of a number of completed projects within an organization. This paper revisits a fuzzy logic size approximation technique - the EPCU model, and presents an improved version, which lifts a number of constraints on its design, considering the Vogelezang dataset used in the literature to define the Equal Size Bands approximation approach.