应用Bays概率法早期预测流脑的流行

为了早期预测流脑在北京市10个远郊居民区的流行,应用贝叶斯(Bayes)概率法建立了数学模型,并以逐步判别分析法作为对比算法。所用基础资料为1960~1980年这10个居民区的流脑旬报数据和人群额外流动量。以此算得1960~1980年的理论发病率及1981~1983年的预测发病率与实际报告值颇相符,且与用逐步判别分析法算得理论值一致。由于该算法简单易学,故适应于县级防疫站应用。

Using Bayes Method of Probability for the Early-Forecast on Prevalence of Epidemic Meningcoccal Meningitis

A mathematical model,based upon Bayes Method of probability (BMP), was established for the early forecast of epidemic cerebrospinal meningitis (ECM) in 10 communities of Beijing outskirts,with the comparision of stepwise Discriminatant Analysis Method (SDAM). Every 10 days''case report of ECM and the records of extra population flowing during 1960~1980 in these communities were analysed in establishing this model.The results of 1981~1983 calculated from this model were 70~100% precise to the observed morbidity levels and trends of (ECM) from 1960 to 1983, and agreed with that calculated from SDAM.Since BMP is far easier to learn than SDAM, it can be used by antiepidemic stations in county level.