基于谷歌趋势的乙型肝炎预测模型

目的 探讨结合谷歌趋势进行预测的状态空间模型应用于乙型肝炎的可行性。方法 采用SAS 9.13软件进行数据分析和建模。以2005年1月2日—2010年12月26日的重庆市乙型肝炎实际发病数和谷歌趋势数据建立状态空间模型，并用2011年1月2日—6月19日的实际发病数据进行模型验证。结果 重庆市乙型肝炎发病情况适合用结合谷歌趋势指数进行预测的状态空间模型，预测值与实际值的平均相对误差为7.44%，预测效果良好。结论 结合网络搜索引擎的数据对乙型肝炎的预测是可行的。

Hepatitis B prediction model based on Google trends

Objective To investigate the feasibility to predict hepatitis B with the state space model in combination of Google trends. Methods SAS 9.13 software was used for data analysis and modeling. The state space model was established with the data of actual incidence of hepatitis B in Chongqing and those from Google trends between January 2, 2005 and December 26, 2010, and the data of actual incidence of hepatitis B between January 2, 2011 and June 19, 2011 was used for model validation. Results The state space model in combination with Google trends was useful to predict the prevalence of hepatitis B in Chongqing. The mean relative error for predictive value and actual value was 7.44%, which indicated a favorable prediction result. Conclusion It is feasible to predict hepatitis B in combination with data from network based searching.