A Hybrid Intelligence System for Risk Estimation in Software Project

Abstract

The aim of this thesis is to investigate fuzzy parameters that may cause risk in software project and further leads to different issues in the product. In previous research the major factors for arising risk in software projects was taken into consideration. But risk which may create flaws does not remains constant to every software projects. This thesis first examines the fuzzy risk parameters in mainly five categories of risk common to all software projects by using fuzzy logic. In second stage the dataset of five categories of risk are provided as input parameters to neural network for the prediction of overall risk probability in software projects with the software project risk of NASA’93 dataset. Finally the result so obtained is the probability of overall risk in software projects lying between low and high or 0 and 1. Keywords: Software project, risk contingency, Project planning, Fuzzy logic, Neural network, Risk probability