Modeling expert effort estimation of software projects

Effort estimation is important part of software project management. Based on applied strategy these models can be classified into groups of algorithmic and non-algorithmic models. In this study we present the model for expert effort estimation developed using data mining techniques - a multilayer perceptron (MLP) artificial neural network. The data set used in the study contains 785 records collected from five projects executed in company specialized for development of solutions in telecom domain. In total 20 estimators participated in the study. Study identifies objects relevant for production of expert effort estimate and presents methodology for its implementation in practice. Proposed model and study results show that DM techniques provide high accuracy effort estimates and therefore are suitable for implementation in real project environments. In future such a model can be used in practice to reduce estimation error and thus enhance expert effort estimation process.