A Multivariate Spatiotemporal Model of COVID-19 Epidemic Using Ensemble of ConvLSTM Networks
- PDF / 2,433,301 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 54 Downloads / 167 Views
ORIGINAL CONTRIBUTION
A Multivariate Spatiotemporal Model of COVID-19 Epidemic Using Ensemble of ConvLSTM Networks Swarna Kamal Paul1
•
Saikat Jana1 • Parama Bhaumik2
Received: 10 July 2020 / Accepted: 30 October 2020 Ó The Institution of Engineers (India) 2020
Abstract The high R-naught factor of SARS-CoV-2 has created a race against time for mankind, and it necessitates rapid containment actions to control the spread. In such scenario short-term accurate spatiotemporal predictions can help understanding the dynamics of the spread in a geographic region and identify hotspots. However, due to the novelty of the disease there is very little diseasespecific data generated yet. This poses a difficult problem for machine learning methods to learn a model of the epidemic spread from data. A proposed ensemble of convolutional LSTM-based spatiotemporal model can forecast the spread of the epidemic with high resolution and accuracy in a large geographic region. The feature construction method creates geospatial frames of features with or without temporal component based on latitudes and longitudes thus avoiding the need of location specific adjacency matrix. The model has been trained with available data for USA and Italy. It achieved 5.57% and 0.3% mean absolute percent error for total number of predicted infection cases in a 5-day prediction period for USA and Italy, respectively. Keywords Covid-19 Spatiotemporal model Convolutional LSTM Ensemble learning Forecasting
& Swarna Kamal Paul [email protected] Saikat Jana [email protected] Parama Bhaumik [email protected] 1
Tata Consultancy Services Kolkata, Kolkata, India
2
IT, Jadavpur University, Kolkata, India
Introduction Wuhan city in China initially observed an outbreak of Covid-19 disease caused by SARS-CoV-2. Eventually it became a pandemic and more than 200 countries are fighting hard to contain the infection. One of the best ways to contain the infection is rapid identification of positive cases and isolation. Forecasting regional spread can help identify future hotspots and distribution of infection which would eventually help to take containment measures. A spatiotemporal epidemic spread model can accommodate both spatial and temporal correlations in data. However, most of the models either require diseasespecific domain knowledge [1] or are too spatially coarse [2]. Deep learning models can learn the dynamics of epidemic spread with high spatial resolution and high degree of accuracy with minimal initial bias due to its capability of high nonlinear representation. Deep neural network-based spatiotemporal models [3] have already been applied to predict epidemic spread. However, this model is experimented on a small localized region and influence of external factors are ignored. Deep learning models also tend to overfit due to its high representational capability. Due to availability of very limited dataset, the problem of overfitting looms large in this case. Thus, modelling of Covid-19 spread in a wide region with high spat
Data Loading...