2011—2017年基于百度搜索指数的全国手足口病预测研究

Prediction of hand-foot-mouth disease in China based on Baidu search index in 2011-2017

目的:利用百度关键词搜索指数和全国手足口病发病数建立手足口病预测模型。方法:从疾病名称、症状、治疗和预防等方面选择关键词,获取关键词的百度指数,筛选出相关系数大于0.7的关键词来构建综合百度搜索指数,结合全国手足口病发病数进行自回归移动平均(ARIMA)模型的拟合和预测。结果:相关性分析表明手足口病发病数与综合百度指数高度相关(r=0.94,P<0.05),且两者变化趋势非常一致。结合百度搜索指数建立的ARIMA模型的平均绝对百分比误差为24.86%,均方根误差百分比为29.86%;单独基于发病历史数据建立的ARIMA模型的平均绝对百分比误差为27.58%,均方根误差百分比为35.50%,前者的预测精度更好。结论:百度搜索指数可以提升ARIMA模型预测性能,结合百度搜索指数和手足口病发病数建立的ARIMA模型能较好地预测手足口病发病数。

Objective:To establish a prediction model of hand-foot-mouth disease(HFMD)by using Baidu keyword search index and the number of HFMD cases in China.Methods:Key words were selected from disease name,symptom,treatment and prevention of HFMD.Baidu Index of keywords were obtained and the key words with correlation coefficient greater than 0.7 were selected to build the comprehensive Baidu search index.The autoregressive integrated moving average(ARIMA)model combined the incidence of HFMD in China was used to fit and predict.Results:The correlation analysis showed that the incidence of HFMD was highly correlated with the comprehensive Baidu Index(r=0.94,P<0.05),and the trend of the two changes was very consistent.The average absolute percentage error of ARIMA model based on Baidu search index is 24.86%,and the root mean square error is 29.86%;The average absolute percentage error of ARIMA model based on the incidence history data alone is 27.58%,and the root mean square error is 35.50%.The prediction accuracy of the former is better.Conclusion:Baidu search index can improve the prediction performance of ARIMA model,and the ARIMA model combined with Baidu search index and HFMD incidence can better predict HFMD incidence.