Generation of Test Cases from Software Requirements Using Natural Language Processing

Abstract

Software testing plays an important role in early verification of software systems and it enforces quality in the system under development. One of the challenging tasks in the software testing is generation of software test cases. There are many existing approaches to generate test cases like using uses case, activity diagrams and sequence diagrams; they have their own limitations such as inability to capture test cases for non functional requirements and etc. Thus these techniques have restricted use in acceptance testing and are not effective for verification & acceptance of large software system. If software requirements are stated using semi-formal or formal methods then it is difficult for the testers and other third party domain experts to test the system. It also requires much expertise in interpreting requirements and only limited number of persons can understand them. This paper proposes an approach to generate test case from software requirements expressed in natural language using natural language processing technique.