This chapter introduces a hybrid deep learning model for COVID-19 spread forecasting. Specifically, the proposed approach combines the desirable characteristics of bidirectional long short-term memory (BiLSTM), convolutional neural networks (CNN), and an attention mechanism. Importantly, this combination, called BiLSTM-A-CNN, is intended to amalgamate the ability of LSTMs to model time dependencies, the capability of the attention mechanism to highlight relevant features, and the noted ability of CNNs to extract features from complex data. The use of the BiLSTM-A-CNN model is expected to improve the forecasting accuracy of future COVID-19 trends.
Bibliographical noteKAUST Repository Item: Exported on 2022-04-20
Acknowledged KAUST grant number(s): OSR-2019-CRG7–3800
Acknowledgements: This chapter is based on work supported by the King Abdullah University of Science and Technology (KAUST), Office of Sponsored Research (OSR) under Award No: OSR-2019-CRG7–3800.