基于API的疾病预测系统数据导入研究  

Research on Data Import of Disease Prediction System Based on API

作　　者：邵荣强 陈燕[1] 龚庆悦[1] SHAO Rong-qiang;CHEN Yan;GONG Qing-yue(School of Artificial Intelligence and Information Technology,Nanjing University of Chinese Medicine,Nanjing 210023,China;Information Department of Yixing No.2 People's Hospital,Yixing 214221,China)

机构地区：[1]南京中医药大学人工智能与信息技术学院,江苏南京210023 [2]宜兴市第二人民医院信息科,江苏宜兴214221

出　　处：《软件导刊》2021年第3期101-105,共5页Software Guide

基　　金：国家社科基金重大项目(18ZDA322);江苏省中医药管理局资助项目(YB2017008)。

摘　　要：为了解决疾病预测系统在机器学习过程中数据导入效率不高、易出错、难以审计的缺点,提出一种通过API导入数据的方法。该方法基于SOAP和RESTful API技术,可实现跨编程语言的数据远程调用,最终实现数据集从数据库自动导入到机器学习服务器。实验结果表明,用10万条16字段的数据集测试,相对于基于文件导入方式,后者导入医疗数据的速度是前者的53倍以上,在1000次数据集测试导入过程中实现了医疗数据的零出错和操作的可审计性。基于API导入数据的方法可有效地提高医疗数据导入效率和稳定性,缩短机器学习模型训练周期,具有可审计性,进一步保障了医疗数据的安全性。To solve the problem of low efficiency of data import in disease prediction system during machine learning,this paper pro⁃poses a method of importing data through API.This method is based on SOAP and RESTful API technology,which can realize remote data calling across programming languages.The experimental results show that compared with the file based import method,the API based medical data import is faster and more stable in machine learning,and the operation is auditable.API data import method can ef⁃fectively improve the efficiency and stability of medical data import,shorten the model training cycle of machine learning,and further ensure the safety of data through auditability.

关 键 词：数据导入 API 机器学习 疾病预测 医疗数据 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]