National University of Defense Technology Utilizes WCT Algorithm to Predict New Confirmed COVID-19 Cases
Professor Lv Xin's team at the National University of Defense Technology proposed a prediction algorithm (WCT) for new confirmed COVID-19 cases based on Weibo big data. This algorithm inputs historical case numbers and processed Sina Weibo data to accurately predict the daily number of new confirmed cases, significantly improving upon the classic Google Trends prediction technology (Google Flu Trends). It provides an automated solution for mining epidemic development trends from large-scale social media data, offers a highly adaptive method for feature engineering with third-party data in epidemic prediction, and constructs a general framework based on big data for predicting the development of infectious diseases when many epidemiological features are unknown in the early stages. By analyzing population mobility data from mobile positioning during the COVID-19 outbreak, the team found that with strict prevention and control measures, COVID-19 transmission could be predicted in the long term through the distribution of the initial high-risk population. This research integrates multi-source data while ensuring personal privacy, providing a decision-making basis and model for emergency management of major public health events, and publicly available data for research teams to reference. Future applications of the algorithm could also extend to monitoring other diseases or sudden public events.