Automatic Cause–Effect Graph Tool with Informal Korean Requirement Specifications

Abstract

In requirement engineering, it is a very important issue to generate test cases with natural language automatically. However, no test case tools deal with informal Korean requirement specifications. In the Korean military software system and airspace industrial area, it is strongly suggested to automatically make just 30% of all possible test cases with requirements. Unlike the previous approaches, we adapted Gary E. Mogyorodi’s cause-effect graphing approach and the model-driven architecture (MDA) approach for automatic test case generation with natural language. In order to generate test cases with informal Korean requirement specifications, we propose an automatic cause–effect tool as an intermediate model for (1) simplifying complicated requirements; (2) modeling the C3Tree (that is, condition and result); (3) identifying incomplete requirements; (4) constructing causes, effects, and relationships; and (5) integrating with two units (that is, similar causes or effects) to remove redundant requirements. We evaluated the accuracy of two generated cause–effect graphs in two ways. With our approach, we can also remove requirement redundancy.

Keywords

• Korean natural language analysis
• automatic test case generation
• model-driven architecture (MDA)
• cause–effect graph
• C3Tree model
• requirement redundancy