Research Article
Predictive Analysis for COVID-19 Cases in India Based on LSTM Algorithms
@INPROCEEDINGS{10.4108/eai.28-10-2022.2328465, author={Zixuan Zhou}, title={Predictive Analysis for COVID-19 Cases in India Based on LSTM Algorithms }, proceedings={Proceedings of the International Conference on Financial Innovation, FinTech and Information Technology, FFIT 2022, October 28-30, 2022, Shenzhen, China}, publisher={EAI}, proceedings_a={FFIT}, year={2023}, month={4}, keywords={machine learning; linear regression; random forest model; long short term memory; coronavirus-related predictions}, doi={10.4108/eai.28-10-2022.2328465} }- Zixuan Zhou
Year: 2023
Predictive Analysis for COVID-19 Cases in India Based on LSTM Algorithms
FFIT
EAI
DOI: 10.4108/eai.28-10-2022.2328465
Abstract
COVID-19 continues to hurt the global economy and living standards. As the number of confirmed cases rises, it is crucial to predict COVID-19 transmission and treatment effectively. This paper aims to compare the results under different machine learning methods and improve the efficiency of predictive analysis. Using valid data provided by Indian government, machine learning regression and the random forest model are implemented to predict the correlation between confirmed cases, deaths, and cured cases. The training and testing dataset scale is 80% to 20% for both models. Comparing the numerical and graphical results of the linear regression and random forest models, while both models generate highly accurate predictions compared with the actual data, the later one contains smaller errors. Another aspect of this paper uses the Long Short-Term Memory (LSTM) model, a popular neural network in artificial intelligence, to predict the future trend of COVID-19 cases, and 70% of the dataset is used for training. As a result, LSTM could make relatively accurate predictions on the overall trend of the confirmed cases, while there appear to be some statistically significant errors compared with actual data.
