Automatic Risk Identification in Software Projects an Approach based on Inductive Learning英文文献资料.docVIP

Intelligent Information Management, 2012, 4, 291-295 /10.4236/iim.2012.425041 Published Online October 2012 (http://www.SciRP.org/journal/iim) Automatic Risk Identification in Software Projects: An Approach Based on Inductive Learning Julia Botan Machado, Silvio do Lago Pereira Department of Information Technology, S?o Paulo State Technological College—FATEC/SP, S?o Paulo, Brazil Email: julia.botan@, slago@pq.br Received August 30, 2012; revised October 4, 2012; accepted October 16, 2012 ABSTRACT Effective risk management is very important to increase the probability of success in software projects. Indeed, like other types of projects, software projects are also susceptible to various problems that can lead to the cancelation of their development or to the development of systems that do not meet the client’s requirements. One of the main activi- ties of risk management is the risk identification, because the list of risks generated in this activity is used all along the risk control process. Thus, this work proposes the creation of an expert system which is capable of identifying risks in software projects by using the lessons inductively learned from similar software projects already developed. By using this proposed expert system, project managers and software developers must be able to avoid errors of the past. Keywords: Risk Management; Risk Identification; Software Engineering; Expert System 1. Introduction work; Section 3 presents the concepts and techniques of artificial intelligence used to implement an expert system to solve that problem; Section 4 briefly discusses em- pirical results obtained with the system; and, finally, Sec- tion 5 presents the final conclusion. There are many different definitions of risk in literature. In this work, risks are defined as future events with some probability of occurrence and a potential for loss. Every project is subject to risks and the role of a project risk m