Text Data Processing on Non-Functional Requirement for the Similarity Between Requirement Elicitation with Deployment Diagram and Recommendation for SRS Improvement

The discrepancy between the developer’s understanding of the client’s wishes in the Software Requirement Specification document is one of the backgrounds in this research. This research uses SRS Baker’s Corner, which is an Android-based cake and pastry product ordering platform. The primary purpose of this research is to process text data on Non-Functional Requirements, where the extraction results can be used as a reference for comparing the suitability between Requirement Elicitation and Deployment Diagram with the text pre-processing method. This research has produced two similarity values, namely between Requirement Elicitation with Non-Functional Requirement and Non-Functional Requirement with Deployment Diagram based on the results of text data extraction in d1-d15. The naming of d1 to d15 is an identity for a text document. The value "d" stands for "document." In the search process for Requirement Elicitation with Non-Functional Requirement, the cosine similarity value closest to 1 is 0.751304, obtained from d2 and d6. Based on the cosine similarity value, the kappa index value is 0.54488 or has a moderate proportion of agreement, and Gwet’s AC1 value is 0.88041, which has an almost perfect proportion of agreement. In the search process for Non-Functional Requirement with Deployment Diagram, the cosine similarity value, which is relatively high, is 0.604605, which is obtained from d6 and d11. Based on the cosine similarity value, the kappa index value is 0.30845 or has a fair proportion of agreement, and Gwet’s AC1 value is 0.95837, which has an almost perfect proportion of agreement. In addition to producing similarity values, there are improvement recommendations for documents with relatively low kappa index values.