COVID-19 growth prediction using multivariate long short term memory
Coronavirus disease (covid-19) spread forecasting is an important task to track the growth of pandemic. Existing predictions are merely based on qualitative analysis and mathematical modeling. The use as much as possible of available big data with machine learning is still limited in covid-19 growth prediction even though the availability of data is abundance. To make use of big data in prediction by using deep learning, we use Long Short Term Memory (LSTM) method to learn correlation of covid-19 growth over time. The structure of LSTM layer is searched heuristically until achieving the best validation score. Firstly, we trained training data containing confirmed cases from around the globe. We achieve favorable performance compared to RNN method with comparable low validation error. The evaluation is done based on graph visualization and RMSE. We found that it is difficult to achieve exactly the same quantity of confirmed cases over time, however, LSTM is able to provide similar pattern between actual and prediction. In future, our proposed prediction can be used for anticipating the forthcoming pandemics. The code is provided here: https://github.com/cbasemaster/lstmcorona
READ FULL TEXT