目的 分析传染病发病数序列的时空特征,预测山东省传染病乙类、丙类发病情况,提高传染病监测预警能力。方法 收集传染病报告信息管理系统山东省法定传染病监测数据资料,应用EViews6.0软件对山东省2010年1月～2014年12月传染病逐月报告发病数构建ARIMA季节乘积模型,并预测2015年1～6月的发病数。结果 乙、丙类传染病均呈现一定的周期波动,其中每年的夏季为传染病的高发阶段,秋冬交替时段会有小幅回升;且在山东省17地市有一定差异;ARIMA模型预测与实际的变化趋势基本吻合。结论 应用ARIMA对山东省乙类传染病的预测效果优于丙类传染病预测效果,最小的相对误差仅为1.15%,提示该模型可以为传染病的预警提供支持。

Objective To improve the ability of disease surveillance, we analyzed the spatial and temporal characteristics of infectious diseases, and predict infectious hepatitis B and C incidence in Shandong province. Methods Collecting data from direct network reporting system of Shandong and using EViews6.0, the ARIMA model was constructed based on the time series between January 2011 to December 2014. The incidence number from January to June in 2015 was also predicted. Results Two types of infectious diseases showed a certain period fluctuation. Every year, summer was the high incidence period of infectious diseases, and there will be a slight rebound in alternating periods of autumn and winter. There were some differences in 17 cities in Shandong. ARIMA model prediction and the actual change trends coincide. Conclusion The forecasting effects for hepatitis B infectious diseases was better than hepatitis C, as the minimum relative error was only 1.15%,that the model could provide theoretical support for alarm of infectious diseases.