Requires analysis based on software maintainability

Requirements analysis is the most important phase of the software life cycle process. Some studies have shown the most faults of software are from the requirements phase. Therefore, the quality of Software Requirements Specification has become the key to project success, which correctness, consistency, no ambiguity of software requirements specification is more important. This paper presents a classification method based on natural language processing techniques and grey similar correlation. The first step of this method is that keywords refined from various functional requirements through segmentation of natural language processing, thus made up of heavy weight vector based on the weight of functional requirements, such a functional requirement corresponds to a weight vector. The second step is that the related technology of grey system is used to compute grey correlation coefficient between two weight vectors, in order to construct a correlation matrix. Finally, the appropriate statistical tools are used to classify functional requirements statements. The clustering results based on this method can provide work guidance for requirement analysts, software developers, software testers, software maintenance.