VARAT: Variable Annotation Tool for Documents on Manufacturing Processes
Building physical models is essential for realizing digital twins in the manufacturing industry. This task, however, is labor-intensive and requires a deep understanding of target processes and extensive knowledge from various literature sources. Although this extensive workload can be mitigated by...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2025-12-01
|
| Series: | Journal of Chemical Engineering of Japan |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/00219592.2025.2454461 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Building physical models is essential for realizing digital twins in the manufacturing industry. This task, however, is labor-intensive and requires a deep understanding of target processes and extensive knowledge from various literature sources. Although this extensive workload can be mitigated by automated extraction of information from the literature, developing such methods necessitates domain-specific datasets lacking in chemical engineering. To address this problem, we developed an algorithm for extracting variable symbols from documents and a variable annotation tool, VARAT, based on this algorithm. Our proposed algorithm, tested on 47 papers on physical models of five manufacturing processes, achieved a recall of 97% and a precision of 96%. VARAT was subsequently employed to create a dataset containing 1,988 variable symbols from the 47 papers. This tool reduced the annotation time per paper by more than half. VARAT is expected to accelerate the development of datasets vital for chemical engineering information extraction and ultimately facilitate the development of physical models. |
|---|---|
| ISSN: | 0021-9592 1881-1299 |