A Framework of Business Rule Extraction from Historic Testing Cases

Business rules could be used to guide the composition of testing cases. However it is difficult for companies to represent or gain their business rules merely depending upon domain experts. Meanwhile, historic testing cases accumulated during previous testing projects provide valuable materials for digging business rules. In this paper, a framework is proposed to extract business rules automatically from historic testing cases to help testing engineers to write new testcases. Firstly, keywords are extracted from historic testing cases based on frequency and freedom of acquired attribute-value pairs. Then, testing cases will be clustered based on semantic similarity of the extracted keywords. Business rules will be discovered and represented by attribute-value pairs. Finally, a case study is demonstrated to verify the validity of the framework.