Development and evaluation of a deep learning approach for modeling seasonality and trends in hand-foot-mouth disease incidence in mainland China

Yongbin Wang,Shengkui Zhang,Zhende Wang,Juxiang Yuan,Chunjie Xu,Li Yang,Ying Zhu

doi:10.1038/s41598-019-44469-9

Yongbin Wang, Shengkui Zhang + Show 5 more

Open Access

https://doi.org/10.1038/s41598-019-44469-9

Copy DOI

Abstract

The high incidence, seasonal pattern and frequent outbreaks of hand, foot, and mouth disease (HFMD) represent a threat for millions of children in mainland China. And advanced response is being used to address this. Here, we aimed to model time series with a long short-term memory (LSTM) based on the HFMD notified data from June 2008 to June 2018 and the ultimate performance was compared with the autoregressive integrated moving average (ARIMA) and nonlinear auto-regressive neural network (NAR). The results indicated that the identified best-fitting LSTM with the better superiority, be it in modeling dataset or two robustness tests dataset, than the best-conducting NAR and seasonal ARIMA (SARIMA) methods in forecasting performances, including the minimum indices of root mean square error, mean absolute error and mean absolute percentage error. The epidemic trends of HFMD remained stable during the study period, but the reported cases were even at significantly high levels with a notable high-risk seasonality in summer, and the incident cases projected by the LSTM would still be fairly high with a slightly upward trend in the future. In this regard, the LSTM approach should be highlighted in forecasting the epidemics of HFMD, and therefore assisting decision makers in making efficient decisions derived from the early detection of the disease incidents.

Highlights

Hand, foot and mouth disease (HFDM) is a common acute infectious disease in children, the majority (91%) of whom are under 5 years[1]
HFMD affects more than two million children annually in mainland China[11], and the number of cases and deaths caused by HFMD sporadics, epidemics and outbreaks invariably tops the list of monitored class C diseases every year since HFMD was designated as a notifiable disease in 20083,12
In view of the long short-term memory (LSTM) model’s flexible capacity to learn what to store and what to abandon during information-processing[25], to the best of our knowledge, this attempt is the first using an LSTM approach to model the long trajectory behaviors of HFMD incidence in mainland China

Summary

Introduction

Foot and mouth disease (HFDM) is a common acute infectious disease in children, the majority (91%) of whom are under 5 years[1]. The autoregressive integrated moving average (ARIMA) model is one of the best linear models in terms of performance for a specified time series[17]; the nonlinear autoregressive neural network (NAR) approach is among the nonlinear models with arbitrarily expected accuracy that can effectively extract the meaningful dynamic information of a data sequence[16] Both the ARIMA and NAR models are well suited to study the future trends of the morbidity or mortality time series of diseases with stationary short-term dependencies based on the aggregated long trajectories[16]. The simulating and predictive abilities of the LSTM model were compared with two especially useful estimation models, including the ARIMA and NAR methods, to seek the best-fitting time series modeling technique for HFMD, which will be of great help in initiating guidance planning and effective intervention measures for HFMD-prevention in mainland China

Objectives

Methods

Results

Conclusion