TY - JOUR
T1 - The predictive skill of convolutional neural networks models for disease forecasting
AU - Lee, Kookjin
AU - Ray, Jaideep
AU - Safta, Cosmin
N1 - Publisher Copyright:
Copyright: © 2021 Lee et al.
PY - 2021/7
Y1 - 2021/7
N2 - In this paper we investigate the utility of one-dimensional convolutional neural network (CNN) models in epidemiological forecasting. Deep learning models, in particular variants of recurrent neural networks (RNNs) have been studied for ILI (Influenza-Like Illness) forecasting, and have achieved a higher forecasting skill compared to conventional models such as ARIMA. In this study, we adapt two neural networks that employ one-dimensional temporal convolutional layers as a primary building block—temporal convolutional networks and simple neural attentive meta-learners—for epidemiological forecasting. We then test them with influenza data from the US collected over 2010-2019. We find that epidemiological forecasting with CNNs is feasible, and their forecasting skill is comparable to, and at times, superior to, plain RNNs. Thus CNNs and RNNs bring the power of nonlinear transformations to purely data-driven epidemiological models, a capability that heretofore has been limited to more elaborate mechanistic/compartmental disease models.
AB - In this paper we investigate the utility of one-dimensional convolutional neural network (CNN) models in epidemiological forecasting. Deep learning models, in particular variants of recurrent neural networks (RNNs) have been studied for ILI (Influenza-Like Illness) forecasting, and have achieved a higher forecasting skill compared to conventional models such as ARIMA. In this study, we adapt two neural networks that employ one-dimensional temporal convolutional layers as a primary building block—temporal convolutional networks and simple neural attentive meta-learners—for epidemiological forecasting. We then test them with influenza data from the US collected over 2010-2019. We find that epidemiological forecasting with CNNs is feasible, and their forecasting skill is comparable to, and at times, superior to, plain RNNs. Thus CNNs and RNNs bring the power of nonlinear transformations to purely data-driven epidemiological models, a capability that heretofore has been limited to more elaborate mechanistic/compartmental disease models.
UR - http://www.scopus.com/inward/record.url?scp=85109595718&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85109595718&partnerID=8YFLogxK
U2 - 10.1371/journal.pone.0254319
DO - 10.1371/journal.pone.0254319
M3 - Article
C2 - 34242349
AN - SCOPUS:85109595718
SN - 1932-6203
VL - 16
JO - PloS one
JF - PloS one
IS - 7 July
M1 - e0254319
ER -