DLKN-MLC: A Disease Prediction Model via Multi-Label Learning

Bocheng Li, Yunqiu Zhang () and Xusheng Wu
Additional contact information
Bocheng Li: Department of Medical Informatics, School of Public Health, Jilin University, Changchun 130021, China
Yunqiu Zhang: Department of Medical Informatics, School of Public Health, Jilin University, Changchun 130021, China
Xusheng Wu: Shenzhen Health Development Research and Data Management Center, Shenzhen 518028, China

IJERPH, 2022, vol. 19, issue 15, 1-15

Abstract: With the increasingly available electronic health records (EHR), disease prediction has recently gained immense research attention, where an accurate classifier needs to be trained to map the input prediction signals (e.g., symptoms, auxiliary examination results, etc.) to the estimated diseases for each patient. However, most of the current disease prediction models focus on the prediction of a single disease; in the medical field, a patient often suffers from multiple diseases (especially multiple chronic diseases) at the same time. Therefore, multi-disease prediction is of greater significance for patients’ early intervention and treatment, but there is no doubt that multi-disease prediction has higher requirements for data extraction ability and greater complexity of classification. In this paper, we propose a novel disease prediction model DLKN-MLC. The model extracts the information in EHR through deep learning combined with a disease knowledge network, quantifies the correlation between diseases through NodeRank, and completes multi-disease prediction. in addition, we distinguished the importance of common disease symptoms, occasional disease symptoms and auxiliary examination results in the process of disease diagnosis. In empirical and comparative experiments on real EHR datasets, the Hamming loss, one-error rate, ranking loss, average precision, and micro-F1 values of the DLKN-MLC model were 0.2624, 0.2136, 0.2190, 88.21%, and 87.86%, respectively, which were better compared with those from previous methods. Extensive experiments on a real-world EHR dataset have demonstrated the state-of-the-art performance of our proposed model.

Keywords: disease prediction; multi label learning; disease prevention; deep learning (search for similar items in EconPapers)
JEL-codes: I I1 I3 Q Q5 (search for similar items in EconPapers)
Date: 2022
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/1660-4601/19/15/9771/pdf (application/pdf)
https://www.mdpi.com/1660-4601/19/15/9771/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jijerp:v:19:y:2022:i:15:p:9771-:d:883182

Access Statistics for this article

IJERPH is currently edited by Ms. Jenna Liu

More articles in IJERPH from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().