Skip to main content
. 2018 Feb 1;15:2. doi: 10.1186/s12976-017-0074-5

Table 1.

Summary of the used data sets in the reviewed studies

Method category Method name Study reference SNS Language Timeframe Location
Graph data mining Graph data mining [23] Twitter English Oct 2008 - March 2009 US
Text mining Historical patterns [45]
Co-occurrences [44] Twitter English March 2009 - Jul 2009
Topic models ATAM [46] Twitter English May 2009 - Oct 2010 US
ATAM+ [47] Twitter English May 2009 - Oct 2010 US
HFSTM [48] Twitter Dec 2012 - Jan 2014 South America
Machine learning Neural network [61] Twitter Jun 2009 - Dec 2009 UK
SVM [57] Twitter English Sep 2012 - May 2013 US
[56] Twitter Nov 2008 - Jun 2010 Japan
[59] Twitter Portuguese March 2010 - Feb 2012 Portugal
[58] Chinese Sina Chinese Sep 2013 - Dec 2013 China
[60] Twitter English May 2009 - Oct 2010 US
[55] Twitter English Nov 2011 - Feb 2015 US
Prediction market using SVR [64] Twitter April 2009 - Jun 2009
Naive Bayes [63] Twitter English Oct 2015 - Nov 2015 Ottawa
Math/Statistical models SNEFT [67] Twitter Oct 2009 - Oct 2010
ACF [65] Twitter English Aug 2008 - Sep 2008
Numerical-based analysis (SEHA using BOW) [68]
Mechanistic disease models Metpopulation model [70] Twitter 2014-2015, 2015-2016 US, Spain, Italy
Compartmental model [35] Twitter Dec 2012 - Jan 2014 South America
Agent-based model [73]
Keys/Documents filtration Keys/Documents filtration [74] Twitter English Sep 2009 - May 2010 US