1. Predictive article recommendation using natural language processing and machine learning to support evidence updates in domain-specific knowledge graphs
- Author
-
Gretchen Purcell Jackson, Bhuvan Sharma, Vanessa V. Michelini, Van C. Willis, Dilhan Weeraratne, Shang Xue, Kirk A. Beaty, Jane L. Snowdon, Brett R. South, and Claudia S. Huettner
- Subjects
0301 basic medicine ,AcademicSubjects/SCI01060 ,Computer science ,precision medicine ,Health Informatics ,Specific knowledge ,Machine learning ,computer.software_genre ,Domain (software engineering) ,03 medical and health sciences ,0302 clinical medicine ,Named-entity recognition ,Intelligence amplification ,Similarity (psychology) ,natural language processing ,business.industry ,Cognition ,artificial intelligence ,Term (time) ,Subject-matter expert ,machine learning ,030104 developmental biology ,030220 oncology & carcinogenesis ,Artificial intelligence ,AcademicSubjects/SCI01530 ,Brief Communications ,AcademicSubjects/MED00010 ,business ,computer ,Natural language processing - Abstract
Objectives Describe an augmented intelligence approach to facilitate the update of evidence for associations in knowledge graphs. Methods New publications are filtered through multiple machine learning study classifiers, and filtered publications are combined with articles already included as evidence in the knowledge graph. The corpus is then subjected to named entity recognition, semantic dictionary mapping, term vector space modeling, pairwise similarity, and focal entity match to identify highly related publications. Subject matter experts review recommended articles to assess inclusion in the knowledge graph; discrepancies are resolved by consensus. Results Study classifiers achieved F-scores from 0.88 to 0.94, and similarity thresholds for each study type were determined by experimentation. Our approach reduces human literature review load by 99%, and over the past 12 months, 41% of recommendations were accepted to update the knowledge graph. Conclusion Integrated search and recommendation exploiting current evidence in a knowledge graph is useful for reducing human cognition load.
- Published
- 2020