A quantitative approach for similarity comparison of the terminologies in standard documents

Huixing Meng, Kanittha Setthapitayakul, Yan Fu Li*

*此作品的通讯作者

科研成果: 期刊稿件会议文章同行评审

摘要

The application of standards is effective for ensuring the expected function of a system in a specific scenario. A key step of the standard enaction or implementation is to precisely comprehend the terminologies in the standards via literal comparison. Current methods for such comparison usually rely heavily on labor, which is time consuming and error-prone. With the rapid increase in the number of the standard documents, it is essential to develop an approach to automatize the comparison process. In this work, we proposed a methodology for the computerized comparison of the terms and definitions in standards. Based on the standard structures, the methodology is developed in three steps: the PDF (Portable Document Format) text conversion, the terms and definitions extraction, and comparison. (1) According to the PDF types, either scanned or digitally created, we provided corresponding methods for converting the PDF files. (2) Regarding the terms and definitions extraction, we identified content structures and logical elements, and extracted definition sections and terms. (3) For the comparison step, we evaluated the similarities between the terminologies from the semantic and syntactic aspects. Reliability, availability, maintainability, and safety (RAMS) are crucial attributes to evaluate the performance of a system. In the experimental studies, we compared the RAMS standards issued by IEC, IEEE, ISO, and the Society for Automotive Engineering (SAE). The results show that the proposed methodology is capable of evaluating the similarities of the terms and definitions in standards.

源语言英语
页(从-至)3254
页数1
期刊Proceedings of the International Conference on Industrial Engineering and Operations Management
2019
MAR
出版状态已出版 - 2019
已对外发布
活动9th International Conference on Industrial Engineering and Operations Management, IEOM 2019 - Bangkok, 泰国
期限: 5 3月 20197 3月 2019

指纹

探究 'A quantitative approach for similarity comparison of the terminologies in standard documents' 的科研主题。它们共同构成独一无二的指纹。

引用此