Adaptive fuzzy logic-based framework for software development effort prediction

An adaptive fuzzy logic framework for software effort prediction that tolerates imprecision, explains prediction rationale through rules, incorporates experts knowledge, offers transparency in the prediction system, and could adapt to new environments as new data becomes available is presented.

Abstract

Algorithmic effort prediction models are limited by their inability to cope with uncertainties and imprecision present in software projects early in the development life cycle. In this paper, we present an adaptive fuzzy logic framework for software effort prediction. The training and adaptation algorithms implemented in the framework tolerates imprecision, explains prediction rationale through rules, incorporates experts knowledge, offers transparency in the prediction system, and could adapt to new environments as new data becomes available. Our validation experiment was carried out on artificial datasets as well as the COCOMO public database. We also present an experimental validation of the training procedure employed in the framework.