Experimental Evaluation of a Spreadsheets Transformation in the Context of Domain Model Engineering
Авторы: Yurin A.Y., Dorodnykh N.O.
Журнал: Proceedings - 2020 Ural Symposium on Biomedical Engineering, Radioelectronics and Information Technology (USBEREIT 2020)
Отчётный год: 2020
Аннотация: Spreadsheets are an important source of information and can be used to create domain models and knowledge bases. We implement an approach for transforming spreadsheets into domain models. This approach consists of 4 steps and based on the use of: an intermediate unified representation of spreadsheets in a canonicalized form, an original technique and software. A canonicalized form provides processing spreadsheets with arbitrary layouts and styles. The original technique provides a two-stage transformation: source spreadsheets to canonicalized tables and canonicalized tables to domain model fragments. Original software is presented by two tools: TabbyXL as a table transformation tool and Personal Knowledge Base Designer as a knowledge base prototyping tool. We present an experimental evaluation of the proposed approach for generating domain models. Two datasets are used for our experiment: Troy200 and ISI-161. We obtained high values of recall and precision (more than 90%) for both datasets. At the same time, the content of resulted domain model fragments was evaluated for the ISI161 dataset. This evaluation showed that 24% of concepts of the domain model for industrial safety inspection tasks were generated automatically by using our approach.
Индексируется WOS: 0
Индексируется Scopus: 1
Индексируется РИНЦ: 0
Публикация в печати: 0
Добавил в систему: