Development of Intelligent Effort Estimation Model Based on Fuzzy Logic Using Bayesian Networks

Accuracy gain in the software estimation is constantly being sought by researchers. On the same time new techniques and methodologies are being employed for getting capability of intelligence and prediction in estimation models. Today the target of estimation research is not only the achievement of accuracy but also fusion of different technologies and introduction of new factors. In this paper we advise improvement in some existing work by introducing mechanism of gaining accuracy. The paper focuses on method for tuning the fuzziness function and fuzziness value. This document proposes a research for development of intelligent Bayesian Network which can be used independently to calculate the estimated effort for software development, uncertainty, fuzziness and effort estimation. The comparison of relative error and magnitude relative error bias helps the selection of parameters of fuzzy function; however the process can be repeated n-times to get suitable accuracy. We also present an example of fuzzy set development for ISBSG data set in order to elaborate working of proposed system.