Nowcasting Influenza Using Google Flu Trend and Deep Learning Model

DOI

10.2991/aebmr.k.201128.079How to use a DOI?

Keywords

Influenza-like illness, Google Flu Trend, Deep learning Model, Now casting

Abstract

Understanding how Influenza-like Illness (ILI) spreads has a substantial societal impact on improving overall human well-being and effective government response. The accurate estimation of ILI occurrences in a timely manner helps to alert and make meaningful intervention decisions that could save lives. In this article, I leverage Google Flu Trends (GFT) data derived from Google general search queries in the United States during 2003-2015 to now cast the occurrence of ILI collected from CDC using various state-of-the-art deep learning models with different neural network architectures, such as a convolutional neural network (CNN), recurrent neural network (RNN), and sequence-to-sequence (Seq2Seq) models. The correlations between GFT and ILI data are first analyzed through the stationary test using classical time-series forecast models such as ARIMA and the recently developed Prophet model. Then, the deep learning models are used to predict the ILI occurrences at both the national and state levels. The results show that deep learning models are capable of now casting the ILI occurrence with good accuracy. Also, the model performance varies at the state level, the signaling occurrence of ILI may be impacted by idiosyncratic factors pertaining to the state and beyond what GFT may capture. Overall, this paper adds to the existing literature on using real-time alternative data sources such as general search queries to estimating ILI occurrence, which sheds light on achieving effective surveillance to achieve a better social good by preventing another pandemic.

Copyright

© 2020, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Ping Jiang
PY  - 2020
DA  - 2020/11/30
TI  - Nowcasting Influenza Using Google Flu Trend and Deep Learning Model
BT  - Proceedings of the 2020 2nd International Conference on Economic Management and Cultural Industry (ICEMCI 2020)
PB  - Atlantis Press
SP  - 407
EP  - 416
SN  - 2352-5428
UR  - https://doi.org/10.2991/aebmr.k.201128.079
DO  - 10.2991/aebmr.k.201128.079
ID  - Jiang2020
ER  -