Continuous Artificial Prediction Markets as a Syndromic Surveillance Technique
This work addresses syndromic surveillance for public health by offering an incremental improvement over existing models like Google Flu Trends.
The paper tackles the problem of early outbreak detection in syndromic surveillance by applying continuous Artificial Prediction Markets (c-APM) to improve upon Google Flu Trends models, showing that c-APM typically achieves lower Mean Absolute Error (MAE) across years, with particularly large improvements from 2011 to 2013.
The main goal of syndromic surveillance systems is early detection of an outbreak in a society using available data sources. In this paper, we discuss what are the challenges of syndromic surveillance systems and how continuous Artificial Prediction Market [Jahedpari et al., 2017] can effectively be applied to the problem of syndromic surveillance. We use two well-known models of (i) Google Flu Trends, and (ii) the latest improvement of Google Flu Trends model, named as GP [Lampos et al., 2015], as our case study and we show how c-APM can improve upon their performance. Our results demonstrate that c-APM typically has a lower MAE to that of Google Flu Trends in each year. Though this difference is relatively small in some years like 2004 and 2007, it is relatively large in most years and very large between 2011 and 2013.