Title:	DETECTING AND PREDICTING INFLUENZA EPIDEMICS IN JAPAN USING INTERNET-BASED DATA
Author(s):	Masashi Inoue, Shinsaku Hasegawa, Masayuki Kakehashi
ISBN:	978-989-8704-11-5
Editors:	Piet Kommers, Pedro Isaías, Claire Gauzente, Miguel Baptista Nunes, Guo Chao Peng and Mário Macedo
Year:	2014
Edition:	Single
Keywords:	Influenza, predicting future epidemics, web-based data, Twitter.
Type:	Poster/Demonstration
First Page:	414
Last Page:	416
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	We investigated whether websites operated by Twitter and local newspapers are promising data sources for monitoring the incidence of influenza. It was found that the Twitter corpus is a promising data source for monitoring national morbidity of influenza, and content from the websites of many local newspapers seems to reflect current local influenza morbidity. We then performed a time series analysis to construct a suitable model for predicting future morbidity. As the result of time series analyses, the nearest neighbor method, which used together Internet-based data with Japanese infectious disease surveillance data, was found to yield the best fit predictions. The results of the obtained prediction function have been made available on our website, which provides up-to-date information about infectious diseases in Japan.
	Go Back