The Application Of Baidu Index In The Analysis Of Avian Influenza A H7N9 Virus And Dengue Fever Epidemics In China

With the development of the Internet information technology,the Internet-based method for infectious disease analysis has become an important part of the national public health research in recent years.Although a lot of research progresses have been made,most of the research focuses on conventional infectious diseases like flu rather than emerging infectious diseases.In addition,most of the works come from abroad diseases monitoring.With the rapid increase of Internet usage in China,there are huge amounts of data,such as Baidu search engine data and Sina weibo data.Using these data to monitor domestic infectious deseases outbreak is helpful for finding epidemic situation timely and taking effective prevention and control mearsures.In 2013 and 2014,Avian influenza H7N9 virus and Dengue epidemic were outbreaked in China seriously,which affected the public health safety.Therefore,this paper is aimed to study the comprehensive application of domestic Internet data for infectious disease analysis.Here,we apply Baidu Index to analyze the two emerging infectious disease epidemics.Firstly,as the H7N9 epidemic is still spreading,we make statistical and spatial-temporal analysis for the epidemic data of the H7N9 virus,and establish a sharing data analysis platform for the public and the related researchers;Secondly,For the study of the correlation between public attention and the actual outbreaks,we evaluate the consistency of user attention pattern about epidemic outbreaks within the epidemic areas quantitatively,and analyze the temporal and spatial correlation between Baidu Index and the actual human infected cases of H7N9 and Dengue epidemics;Finally,in order to predict the dynamics of H7N9 and Dengue,we select keywords related with the clinical characteristics and epidemic,and establish a Multiple Linear Regression Model(MLRM)based on Baidu Index with 12 key words.To improve the model prediction results,we join the offset historical cases data as independent variables in the MLRM.The experimental results show that both in H7N9 and Dengue epidemics,the reflecting pattern of the public in the epidemic regions for the epidemic is high consistency,Baidu Index is significantly positive correlated with the epidemic data both in temporal and spatial dimensions.So Baidu Index can well reflect the epidemics' dynamic and severity.For the research of epidemic forecasts,the results of the optimal prediction model predicted show that the MLRM based on Baidu Index can accurately predict H7N9 and Dengue epidemic dynamics,and it's better for Dengue with higher real-time performance.Overall,this work can provide some reference value and guiding significance to China's emerging infectious diseases analysis based on Internet data.