We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
An Assistant Diagnosis Method Based on Association Paths Representation Learning for Diseases.
- Authors
Ying Zhong; Xiaojun Luo
- Abstract
The introduction of Electronic Health Records (EHRs) has created a new set of challenges and opportunities for clinical research, stimulating the transition of medical health-related studies to a data-driven approach and opening up new prospects for more personalised health care. In light of the intricate correlation between medical data, this paper presents a learning approach to represent medical information connection pathways. First, the method uses extracted medical pathways to enhance potential associations between diseases and learns highly representative patient embedding vectors. Second, by combining the medical information association path, the representative medical concept representation vector is learned. Meanwhile, the self-attention reasoning network learns the patient feature embedding vector using the patient’s previous admission records in order to properly anticipate the patient’s future health state. Finally, the experiment demonstrates that the proposed method can effectively use the medical path and the patient’s historical disease record for disease prediction, and achieve excellent prediction results.
- Subjects
ELECTRONIC health records; DIAGNOSIS methods
- Publication
IAENG International Journal of Computer Science, 2023, Vol 50, Issue 4, p1470
- ISSN
1819-656X
- Publication type
Article