Details

Autor(en) / Beteiligte

• Anand, Amann
• Joshi, Aman
• Pandey, Nishant Sagar
• Singh, Raghvendra
• Tikkiwal, Vinay Anand
• Kar, Arpan Kumar

Titel

A Deep Learning Based Hybrid Approach for Short-Term Forecasting of Spread of COVID-19

Ist Teil von

• IoT, Big Data and AI for Improving Quality of Everyday Life: Present and Future Challenges, p.261-278

Ort / Verlag

Cham: Springer International Publishing

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• COVID-19 is currently a major threat to humanity. The rapid spread of this pandemic from China to the rest of the world is claiming many lives and disrupting the economy. This turn of events makes it imperative to conduct a detailed study of the trend it follows to develop adequate short-term prediction models to combat its impact. This study utilizes the European Centre for Disease Prevention and Control (ECDC) data for three days and one week ahead forecasts of new cases for India, USA, and Russia. A Long Short Term Memory (LSTM), Bidirectional Long Short Term Memory (BiLSTM), Gated Recurrent Unit (GRU), and a hybrid GRU-Cuckoo Search (GRU-CS) model where the network parameters of the GRU network are optimized using Cuckoo Search algorithm is developed for this purpose. The performance of all the four models is compared, and it is found that the proposed GRU-CS model produces significantly better forecasting results, which can aid in keeping track of the ongoing increase in COVID-19 cases and assisting public health systems in developing strategic plans to reduce mortality and morbidity rates.