Supporting clinical trial data curation and integration with table mining
Download as PPTX, PDF2 likes282 views
This document summarizes a system for extracting and curating data from tables in clinical trial publications. It describes the challenges in extracting information from tables due to their complex structures and dense, ambiguous content. The system decomposes tables into different types such as list, matrix, super-row, and multi-tables. It then annotates the tables with metadata and semantics by linking data cells to existing knowledge bases. Evaluation showed 85% accuracy in decomposition. Further work will add more semantic annotations and expand linked knowledge bases.
Supporting clinical trial data curation and integration with table mining
- 1. Supporting clinical trial data curation and integration with table mining Nikola Milosevic1, Cassie Gregson3, Robert Hernandez3, Goran Nenadic1,2 1School of Computer Science, University of Manchester 2 The Farr Institute @HeRC 3AstraZeneca
- 2. Clinical trial publications • Around 800 000 clinical trials in PubMed • Difficult to digest/search • Text mining approaches • But tables and figures are often not processed
- 3. Tables in publications • Present factual information • Usually: • Experimental settings (i.e. demographics) • Findings and results (e.g. DDI, side effects, adverse events…) • Background information (previous research, datasets, etc.) • Examples • Important information about trials
- 4. Extraction and curation of table data
- 5. Challenges • Complex structure • Table dimensionality (1, 2, multi-dimensional) • Visual relationships • Dense content • Ambiguous short text • Lack of context • Acronyms and abbreviations • Incomplete information
- 8. Table types (1) • 4 types: list, matrix, super-row and multi-tables • List table:
- 9. Table types (2) • Matrix table
- 10. Table types (3) • Super-row table
- 11. Table types (4) • Multi-table
- 15. Results
- 16. Next steps • Add semantic annotations • Link patterns in data cells with its meaning • Build/Expand knowledge bases • Relate to existing knowledge on the semantic web
- 17. Annotation schema • Meta-data • Paper (name, abstract, authors, publisher) • Authors (names, emails, affiliations) • Table (caption, footers) • Cells (content, role) • Inter-cell relationships • Semantics (links to ontologies, dictionaries, knowledge bases)
- 18. Summary • Tables contain valuable information such as settings or results • System for extraction and curation of table data • Decomposition and annotation of the tables • Accuracy of 85% • Semantic analysis and information extraction