基于百度指数的人感染H7N9禽流感疫情预测

Prediction of human infection with avian influenza H7N9 based on Baidu index

目的已有研究在基于百度指数预测传染病疫情时大多未考虑各省人口规模对疫情严重程度评估的影响以及公众搜索行为随疫情发展的变化。以人感染H7N9禽流感为例,克服上述不足,基于百度指数预测该疫情的发展。方法引入各省每年每1000万常住人口中感染病例数来评估各省疫情的严重程度,选择疫情较为严重的省份。基于关键词“H7N9”的百度指数建立支持向量机回归来预测该省疫情的发展。结果若不考虑人口规模的影响,仅利用总病例数评估各省疫情的严重程度,福建省疫情会被低估。进一步选择基于百度指数预测福建省疫情,发现:随着疫情的发展,福建省公众的搜索行为发生变化。因此,考虑到公众的搜索行为,基于福建省关键词“H7N9”的百度指数分波段建立支持向量机回归来预测疫情的发展,该回归能够准确预测疫情的变化趋势以及峰值暴发的时间。结论根据公众的搜索行为,分波段建立疫情预测模型,可以实现疫情变化趋势和暴发时间的准确预测。

Objective Most of the current studies did not take into account the impact of population scale on assessment of the severity of the epidemic in each province and the change in public search behavior with the development of infectious diseases in predicting infectious diseases based on the Baidu index.Taking human infection with avian influenza H7N9as an example,this paper overcame the above shortcomings and predicted the development of the epidemic based on Baidu index.Methods Introducing the number of infected cases per10million permanent residents to assess the severity of the outbreaks in each province.Then,Establishing a support vector machine regression model based on the Baidu index of"H7N9"to predict the development of the epidemic.Results The outbreaks of avian influenza H7N9in Fujian Province could be underestimated if we just utilize the total number of cases to assess the severity of the outbreaks in each province with neglecting the influence of the population scale.Moreover,we chose to predict the epidemic in Fujian Province based on Baidu index.The analysis showed that the public search behavior changed as the disease spread.Therefore,according to the public search behavior,this paper established a support vector machine regression model based on the Baidu index of"H7N9"for different waves to predict cases infection with avian influenza H7N9in Fujian province.The model proposed in this paper could accurately forecast the epidemic trend of H7N9and the peak time of its outbreaks.Conclusion According to the public search behavior,the prediction model for different waves could be established,which can accurately predict the trend and the outbreak time of epidemic.