Combinetf for Requirements Data Similarity Detection on AREM

• Main Authors: Delima, Rosa, Wardoyo, Retantyo, Mustofa, Khabib
• Format: Article PeerReviewed
• Language: English
• Published: ICIC International 2022
• Subjects: Information and Computing Sciences
• Online Access: https://repository.ugm.ac.id/283437/1/COMBINETF-FOR-REQUIREMENTS-DATA-SIMILARITY-DETECTION-ON-AREMICIC-Express-Letters.pdf; https://repository.ugm.ac.id/283437/; https://www.scopus.com/record/display.uri?eid=2-s2.0-85138349906&doi=10.24507%2ficicel.16.09.913&origin=inward&txGid=ba3b1f6fafeab5d743a3b9e3c66d45b3; https://doi.org/10.24507/icicel.16.09.913
• Institution: Universitas Gadjah Mada

The Automatic Requirements Engineering Model (AREM) is a model that can automate the requirements engineering process. This model accepts input in the form of requirements data from several stakeholders. The similarity of the description of the requirements of one stakeholder with other stakeholders is very likely to occur. Therefore, the collected requirements data are to be processed and tested for similarity so that there is no duplication of requirements in system modeling. In this study, the CombineTF method was developed to check the similarity of the data requirements. CombineTF is a hybrid method that combines a term-based approach with Term Frequency (TF) and character-based similarity. In this research, CombineTF is integrated with the Jaro-Winkler algorithm and Levenshtein distance as a character-based similarity. The experimental results show that CombineTF has a good performance for measuring the similarity of requirements documents with a threshold of more than 0.5.