摘要
目的建立一种数字化评估出入境人员亚健康风险的方法。方法收集2008—2014年出入境人员的鹰演电子扫描系统的检查数据与临床体检数据,经数据预处理后,采用极限学习机(ELM)和支持向量机(SVM)同时进行数据挖掘,建立数学模型,并对两种方法的评估效果进行比较。结果 ELM对出入境人员呼吸系统、泌尿生殖系统、免疫系统、内分泌系统、神经系统、消化系统、循环系统和运动系统的训练准确率分别为99.6%、99.4%、99.9%、99.8%、99.7%、97.8%、98.5%和99.0%,SVM的训练准确率均为100.0%。ELM对8个系统的预测准确率分别为87.2%、91.7%、92.8%、93.6%、81.7%、84.7%、87.0%和84.4%,而SVM的预测准确率依次为69.4%、95.7%、81.5%、87.8%、71.1%、82.3%、85.9%和66.0%。结论采用ELM能够准确地对出入境人员亚健康情况进行数字化评估,对实现出入境人员的亚健康早期干预和健康管理有重要意义。
Objective To establish a method for the digital evaluation of human sub-health risks of entry-exit personnel. Methods Collect DDFAO electronic scanning system inspection data and physical examination data of entry-exit personnel from 2008 to 2014. After data preprocessing,establish the mathematical model using extreme learning machine (ELM) and support vector machine(SVM) at the same time,and then compared the evaluation effect of the two methods. Results The training accuracy of the health assessment model was 99.6%,99.4%,99.9%,99.8%, 99.7% ,97.8% ,98.5% and 99.0% ,by ELM on the respiratory system,urogenital system,immune system,endocrine system,nervous system,digestive system, circulatory system and skeletal system,respectively. The training accuracy was 100.0% by SVM on the eight systems. The forecast accuracy was 87.2%,91.7%,92.8%,93.6%,81.7%,84.7% ,87.0% and 84.4%,by ELM on the eight system respectively,while the SVM forecast accuracy were 69.4%,95.7%,81.5%,87.8%, 71.1%,82.3%,85.9% and 66.0%,respectively. Conclusion ELM was an important way to realize the digital evaluation of the sub-health risks of the entry-exit personnel. It's beneficial to realize the early health intervention and management of entry-exit personnel.
作者
王嫱
罗力涵
龙川凤
WANG Qiang LUO Li-han LONG Chuan-feng(Shenyang International Travel Healthcare Center, Shen yang, Liaoning 110016, China)
出处
《中国国境卫生检疫杂志》
CAS
2016年第6期404-406,共3页
Chinese Journal of Frontier Health and Quarantine
基金
质检公益性行业科研专项项目(201310083)
关键词
亚健康
出入境
极限学习机
支持向量机
Sub-health
Entry-exit
Extreme learning machine
Support vector machine