Name
Abstract | ||
|---|---|---|
The popularity of social networks provides a new way for constant surveillance of unusual events related to a certain disease. Some researchers have begun to use twitter to estimate the situation of public health, as well as predict disease trends. However, previous studies usually focused on the infection data but not the data judged as non-infection, which was usually filtered directly in their studies. We believe that the non-infection data is also essential for monitoring disease activity, because of their inherently subtle connections. Firstly, we construct a time series outlier model that can detect flu outlier events of different region in China with high precision and good recall by mining all the flu related data. Secondly, those outlier events are used to find out hot topics by SN-TDT and use the twice iteration classification method which is designed to analyze users' status who published a flu-related weibo. These results could provide science reference for deploying sickness prevention resources, and make recommendation about which place pose a high risk of getting infected. |
Year | DOI | Venue |
|---|---|---|
2014 | 10.1007/978-3-319-11119-3_13 | WEB TECHNOLOGIES AND APPLICATIONS, APWEB 2014, PT II |
Keywords | Field | DocType |
weibo, outlier events, time series, twice iterate classification | Data science,Public health,Data mining,Disease,Social network,Computer science,Popularity,Outlier,Recall | Conference |
Volume | ISSN | Citations |
8710 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 10 | 5 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Quanquan Fu | 1 | 0 | 0.34 |
| Changjun Hu | 2 | 130 | 27.56 |
| Wenwen Xu | 3 | 8 | 3.17 |
| Xiao He | 4 | 2 | 0.70 |
| TieShan Zhang | 5 | 0 | 0.34 |