Fast Generation of Clinical Pathways including Time Intervals in Sequential Pattern Mining on Electronic Medical Record Systems
| Autor: | Haruo Yokota, Hieu Hanh Le, Yuichi Honda, Tomoyoshi Yamazaki, Muneo Kushima, Kenji Araki, Henrik Edman |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Computer science
business.industry Process (engineering) Electronic medical record 02 engineering and technology computer.software_genre Health informatics 03 medical and health sciences 0302 clinical medicine Clinical pathway 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing 030212 general & internal medicine Data mining Sequential Pattern Mining business computer |
| Zdroj: | 2017 International Conference on Computational Science and Computational Intelligence (CSCI). |
| DOI: | 10.1109/csci.2017.300 |
| Popis: | Machine-based generation of clinical pathways that utilizes sequential pattern mining to extract the pathways from historical electronic medical record (EMR) systems has gained much attention. We previously proposed a method to generate clinical pathways including time intervals that provides rich information to medical workers. However, this method is difficult to use in real applications because of slow clinical pathway generation as a large number of duplicate patterns are included. In this paper, to speed up the clinical pathway generation, we deploy an occurrence check that adds only closed sequential patterns to the results during mining while considering time intervals between events. Experiments on real data sets showed that our proposal can be more than 13 times faster than our earlier method and can significantly improve the decision-making process for medical actions at large hospitals. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|