Comparison of the Application of Neural Networks with K-Fold Cross Validation and Sliding Window Validation for Forecasting Covid-19 Recovered Cases

Abstract

The Covid-19 virus first appeared in China resulting in millions of confirmed cases, deaths and recovered cases to date. The spread and increase in the death rate due to Covid-19 is very worrying. Health workers and researchers continue to struggle to improve recovery from Covid-19 cases. There is a need for future forecasting to predict recovery from cases that occur, so that the public or government can understand the spread, take precautions and prepare for action as early as possible. Several previous studies have carried out forecasting the future impact of Covid-19 using Machine Learning methods. Neural Network and Sliding Window are appropriate methods for forecasting time series data. In this research, it has been proven that the application of a Neural Network with a Sliding Window can improve performance which is much better than without using a Sliding Window in forecasting Covid-19 recovery cases in China.