COVID-19 Outbreak Estimation Approach Using Hybrid Time Series Modelling

• Published in: Innovations in Intelligent Computing and Communication pp. 249 - 260
• Main Authors: Chakraborty, Soham, Mishra, Sushruta, Tripathy, Hrudaya Kumar
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing
• Series: Communications in Computer and Information Science
• Subjects: ARIMA; Corona-virus; Epidemic transmission; LSTM networks; Machine learning; Time series forecasting
• ISBN: 3031232321; 9783031232329
• ISSN: 1865-0929; 1865-0937
• DOI: 10.1007/978-3-031-23233-6_19

In the beginning of March 2020, coronavirus was claimed to be a worldwide pandemic by the World Health Organization (WHO). In Wuhan, a region in China, around December 2019, the Corona virus, also known as the novel COVID-19 was first to arise and spread throughout the world within weeks. Depending upon publicly available data-sets, for the COVID-19 outbreak, we have developed a forecasting model with the use of hybridization of sequential and time series modelling. In our work, we assessed the main elements to forecasting the potential of COVID-19 outbreak throughout the globe. Inside the work, we have analyzed several relevant algorithms like Long short-term memory (LSTM) model (which is a sequential deep learning model), used to predict the tendency of the pandemic, Auto-Regressive Integrated Moving Average (ARIMA) method, used for analyzing and forecasting time series data, Prophet model an algorithm to construct forecasting/predictive models for time series data. Based on our analysis outcome proposed hybrid LSTM and ARIMA model outperformed other models in forecasting the trend of the Corona Virus Outbreak.