Designing a Test Case Catalog Based on Software Requirements Documents Using Text Mining

Software development projects involve a number of companies and individuals with diverse knowledge and expertise. This Failure of software development projects is generally caused by weaknesses in the testing process. Fixing and detecting bugs is a costly and time-consuming activity. Bugs can be addressed through the testing process, which utilizes test cases. However, creating test cases manually is expensive and time-intensive. This study presents a Test Case Catalog design that reuses existing test cases, focusing on XYZ campus system applications. Test cases are organized based on existing software requirements documents as a basis for testing. By utilizing a repository containing similar and relevant test cases, the Test Case Catalog offers a systematic tool for application testing of the XYZ project. The catalog is designed using Text Mining, including Bag of Words, to extract and analyze similar and relevant functions or features from the software requirements documents. Based on the threshold results with a percentile value of 70 not only relevant feature selection to generate test cases but also improves the testing process, as evidenced by the better success rate compared to the pretest phase, manual testing without using the Test Case Catalog. Thus, this catalog emerged as an effective and efficient tool in software testing.