|
カワイ ユキコ
KAWAI YUKIKO
河合 由起子 所属 京都産業大学 情報理工学部 情報理工学科 職種 教授 |
|
| 言語種別 | 英語 |
| 発行・発表の年月 | 2016/10/17 |
| 形態種別 | 研究論文（国際会議プロシーディングス） |
| 査読 | 査読あり |
| 標題 | After the boom no one tweets: Microblog-based influenza detection incorporating indirect information |
| 執筆形態 | その他 |
| 掲載誌名 | ACM International Conference Proceeding Series |
| 出版社・発行元 | Association for Computing Machinery |
| 巻・号・頁 | pp.17-25 |
| 著者・共著者 | Shoko Wakamiya,Yukiko Kawai,Eiji Aramaki |
| 概要 | The recent rise in popularity and scale of social media services (SNSs) has created the growing necessity for SNS-based information extraction systems. A popular application of SNS data is health surveillance for predicting the outbreak of epidemics by detecting diseases from text messages posted on SNS platforms. Such applications share the same logic: they incorporate SNS users as social sensors. Such social sensor approaches also share a common problem: SNS-based surveillance can be reliable if sufficiently numerous users are active, but small inactive populations produce inconsistent results. To overcome this problem, this paper presents a novel approach using indirect information covering both urban areas and rural areas within the posts. This system uses not only direct information, but also indirect information that mentions other places. Indirect information is less reliable (too noisy or too old) than direct information. Therefore, we do not use the described data directly. Instead, we consider that indirect information inhibits direct information. For example, when indirect information appears often, we consider that everyone already has a known disease, leading to a small amount of direct information. In the experiment using three years' long collection of tweets (7 million influenza-related tweets in Japanese) described herein, this proposed approach improved the detection performance not only in rural cities, but also in urban cities, thereby demonstrating the feasibility of this approach. |
| DOI | 10.1145/3007818.3007822 |
| DBLP ID | conf/edb/WakamiyaKA16 |
| PermalinkURL | http://dblp.uni-trier.de/db/conf/edb/edb2016.html#conf/edb/WakamiyaKA16 |