期刊名称：International Journal of Engineering and Computer Science
印刷版ISSN：2319-7242
出版年度：2015
卷号：4
期号：1
页码：10136-10239
出版社：IJECS
摘要：Many risks are involved in the development of software project and Risk assessment methods are most important component inthe process of risk management. They are a critical component of software project management and software testing. Practitioners,particularly researchers, are mostly interested in the evaluation of these methods for their applicability, strengths, and weaknesses forparticular scenarios. So far, no model has proved to be successful at effectively and consistently predicting software development cost. FuzzyEx-COM (Fuzzy Expert COCOMO) that combines the advantages of a fuzzy technique with Expert COCOMO methodology for riskassessment in a software project which leverages existing knowledge and expertise from previous effort estimation activities to assess therisk in a new software project. A novel Neuro-fuzzy Expert Constructive Cost Model is proposed to improve the accuracy of risk assessmenttechnique. With the introduction of the Neuro-Fuzzy Risk Methodology which combines the non-linear learning features of neuralnetworks with fuzzy logic that has capability to deal with sensitive and linguistic data and generate risk rules using Artificial NeuralNetwork(ANN) techniques to improve the accuracy of risk assessment technique. This paper shows the workflow required for implementingthe Neuro-Fuzzy Risk Methodology on the original Fuzzy Ex-COCOMO methodology
