A COVID-19 Control Policy Evaluation and Prediction Method Based on Clustering

ABSTRACT

COVID-19 has brought huge losses to the economy all over the world. To solve this problem, we delivered an epidemic situation evaluation and prediction system based on dynamic data clustering, which was established to cluster the epidemic data in different regions and make evaluations and predictions through the Markov chain. Using the method of streaming data to cluster, we set the data at the same cluster as the sampling results from the same distribution to classify the epidemic situation. We used the Markov chain model to estimate the future development of the epidemic situation. According to the characteristics of stream data, the system can avoid the impact of epidemic data not meeting the assumption of independent homodistribution and only assess the epidemic situation based on local areas. © 2022 IEEE.