For software project planning control and management, an accurate esti
mate of software development cost is important. Past research has focu
sed on using parametric models to predict development cost based on at
tributes such as lines of code or function points. This requires resea
rchers to identify the set of factors that influence cost estimation b
efore the system is constructed. We propose a non-parametric approach
that integrates a neural network method with cluster analysis to estim
ate development cost. The integration of the two techniques not only a
llows for a more accurate cost estimate but also leads to an increase
in the training efficacy of the network. (C) 1998 Elsevier Science B.V
. All rights reserved.