Bayes判别分析在腹泻病诊断方面的应用研究

Application of Bayes discriminant analysis in diarrhea diagnosis

目的利用Bayes判别函数对腹泻病例的病原学诊断结果进行预测,为基层医疗机构诊断腹泻病提供参考。方法采用中国疾病控制信息系统收集腹泻病病原学监测数据,构建Bayes判别函数,利用收集到的与腹泻病诊断有关的一些因素对腹泻病例的病原学诊断结果进行预测。结果利用2010年1～12月甘肃省各哨点医院实验室检验的437例腹泻病例的相关数据,构建根据病人年龄、症状、体征预测腹泻病原体情况的Bayes判别函数,经检验,预测准确率为57.67%,简单kappa系数为0.20。结论 Bayes判别函数对腹泻原因预测的结果可以从统计学角度对腹泻的病因诊断提供参考信息。

Objective To forecast the result of diarrheal diseases etiology diagnosis by using Bayes discriminant function,so as to provide reference for medical institutions to diagnose diarrheal diseases.Methods Diarrhea etiology monitoring data were collected from Disease Surveillance Information Reporting System.Bayes discriminant functions were constructed,and the etiologic diagnosis results of diarrhea cases were forecasted through the factors which were associated with the diagnosis of diarrheal.Results Using related data of 437 diarrheal diseases cases collected by sentinel hospitals in Gansu Province during 2010,Bayes discriminant functions were constructed,it can forecast the result of diarrheal diseases etiology diagnosis by the patients age,symptoms,signs.After examination,prediction accuracy rate was 57.67%,simple kappa coefficent was 0.20.Conclusions The prediction of the cause of diarrheal diseases by Bayes discriminant functions can provide diarrheal diseases etiology diagnosis reference information.