Requirement Engineering Techniques Selection and Modeling An Expert System Based Approach

Choosing suitable Requirement Engineering (RE) techniques for a particular project is a challenging and time-consuming task, requiring substantial expertise and efforts. In this paper, an expert system based approach is proposed to help solving this problem. This expert system is capable of modeling and selecting suitable RE techniques for a software project. An on-line questionnaire is created in the first place to collect the expertise available in the community. A new algorithm is proposed to convert the raw data to a training data set suitable for building a complete Bayesian Belief Network (BBN). The resulting BBN is used to build sub-BBNs for RE techniques in six RE phases. The sub-BBNs integrated with a user interface form the expert system. Empirical study shows that the expert system outperforms other predictors in selecting suitable RE techniques. A case study is conducted to show the application of the expert system in the real world.