Predicting COVID-19 from a Quantitative Finance
Senior Advisor of China ABS Business, MSCI
Founder & CEO, CreditWise Technology Co., Ltd.
Chen Jian has recently used the quantitative model of credit risk analysis for
the prediction of COVID-19 outbreaks. This approach is proven to be robust,
flexible, and accurate. He has published two medical papers with his medical
colleagues on Lancet Preprints. His paper, “Predictive Modeling for Epidemic
Outbreaks: A New Approach and COVID-19 Case Study” is forthcoming on Asia-Pacific
Journal of Operational Research. He has been invited to discuss the Asian
experiences with Dr. Zhang Wenhong, Dr. David Ho, along with other leading
medical experts at the Brookings Institution online forum, to share insights
about the predictive model.
SIR-type models rely heavily on calibrated parameters, which poses serious
modelling challenges, regarding model stability, sensitivity, and accuracy.
This new approach applies state transition matrix techniques in coronavirus
infection and treatment predictions. The key innovation is trifold: 1. It is a
flexible model which can predict the intermediate states; 2. It doesn't need
too much parameter estimation and is mainly driven by empirical probabilities;
3. It treats all the government preventive measures as embedded in the observed
probabilities. Thus, this model greatly improves the predictive power.
April 2020, Dr. Chen has given more than 30 presentations about the forecast
model to various universities, financial institutions, and regulatory agencies,
with an estimated audience more than 10 million via media broadcasting.