Summary: Since December 2019, the novel coronavirus disease (COVID-19) has spread to most of the countries of the world. This chapter applies machine learning (ML) algorithms in order to predict the spread of COVID-19. For this, the datasets of India and the USA are taken into consideration as these countries have the highest number of confirmed cases as of 30 October 2020. Data till 23 September, 2020 are used as training data and data from 24 September 2020 to 10 October 2020 as the test or evaluation data. Different ML regressions including polynomial regression, support vector regression, Holt’s linear model, autoregression (AR), moving-average model, and autoregressive integrated moving average model are applied to the datasets. For this, first, the time-series training data is studied to investigate the appropriateness of a model. If the model is found to fit the data, it is then used for prediction purpose. Experiments are performed using the Scikit-learn library of Python programming language. Experimental results show that for both India and the USA, the AR model shows the best performance in accurately predicting the confirmed cases of COVID-19.
For the entire collection see [Zbl 1472.92003].