Research on Early Warning of MDR Risk Based on Machine Learning and Data Mining Technology
DU Hongyang1, WU Shifu2, HUANG Lin2, DU Weihong1, Wangshuyu3
1.Adverse Drug Reaction Monitoring Center of Heze, Heze Shandong 274000, China; 2.Shandong Center for ADR Monitoring, Jinan
Shandong 250012, China; 3.Heze Municipal Hospital, Heze Shandong 274000, China
Abstract:In view of the shortcomings of early warning function of medical device adverse events (MDR) in our country, combined
with the exploration results of early warning project of MDR data of Shandong Province in the past three years, the machine learning
and data mining techniques based on data cube into the field of warning of MDR were introduced firstly in this paper. Then the
warning system framework of MDR based word2vec and Cluster analysis was proposed, and the analysis was carried out combined
with the actual data of adverse event report of medical devices. According to actual needs, early warning can be realized from
production enterprises, registration certificate number, product batch number, production date and other dimensions, which provides a
new method for MDR risk early warning and decision support for medical device adverse event monitoring personnel.
杜红阳1,吴世福2,黄琳2,杜卫红1,王淑玉3. 基于机器学习和数据挖掘技术进行MDR风险预警的研究初探[J]. 中国医疗设备, 2018, 33(7): 138-141.
DU Hongyang1, WU Shifu2, HUANG Lin2, DU Weihong1, Wangshuyu3. Research on Early Warning of MDR Risk Based on Machine Learning and Data Mining Technology. China Medical Devices, 2018, 33(7): 138-141.