Dr. George Yuan, distinguished professor of Center for Financial Engineering in Soochow University, correctly predicts COVID-19 peaking in China by mid-February using a mathematical model called iSEIR.1
by Shaun Tan Yi Jie
As the COVID-19 situation in China tails off, much attention has fallen onto a particular professor from Soochow University who predicted on 7 February 2020 – earlier than anyone – that the epidemic would peak around mid to late February in China.2-5
“If the isolation control program currently implemented continues with full force and the disclosed statistics are true, then considering the current situation and combining it with the simulation analysis of our internal iSEIR model,1 we believe 1 February 2020 is the start of the turning point of the COVID-19 epidemic situation in China, and a peak should be reached around middle to end of February,” wrote Dr. George Yuan Xianzhi, distinguished professor of Centre for Financial Engineering in Soochow University, in a WeChat post that has since gone viral.2-4
Professor Yuan said that the iSEIR model is an extension of the SEIR model used to describe the spread behaviour of infectious diseases, where “S” refers to those who are susceptible to disease but have not been infected yet; “E” refers to the exposed group who are infected but are not infectious yet; “I” refers to those infected who also become infectious; and “R” refers to those who have recovered from the infection, who may or may no longer be infectious.1
According to Prof. Yuan, the SEIR model does not consider the fact that every individual has a subject-specific probability to become a spreader. “With the iSEIR model, where ‘i’ stands for ‘individual’, we are able to study the distribution of individual behaviors by studying each node in the corresponding multiplex network,” he explained.
In this scenario, the model focused on two critical parameters Prof. Yuan dubbed as “Delta” and “Gamma”, terminologies borrowed from financial risk management. “Delta” refers to the change in percentage of infected people in China each day, while “Gamma” refers to the speed of change in percentage of infected people in China each day, a concept also known as second derivative in calculus.
Prof. Yuan’s group observed that “Delta” and “Gamma” were below 20 percent and 2 percent, respectively, for five consecutive days since 1 February 2020, which led them to conclude that 1 February 2020 heralds the start towards the situation peaking.
Their iSEIR model also calculated that at the national level (excluding Hong Kong, Macao and Taiwan):
- The number of infectious and asymptomatic (E) will reach a peak in about two days, and then after about seven days, the rate of increase will decrease to less than 10 percent.
- The daily increase in the number of people infected and symptomatic (I) is already in a steady downward trend, and the rate will decrease to within 10 percent after about 14 days.
- The proportion of people recovering from infection (R) will return to more than 80 percent after about 17 days.
As a result, they earmarked the second half of February as the period when the peak would happen (see exclusive figures below provided by Yuan to APBN).
Turning Period, Not Turning Point
In a phone interview with APBN, Prof. Yuan stressed that their model does not predict a specific date for the peak, but rather an approximate interval encompassing a larger time period. “We think the so-called ‘turning point’ is not an accurate term. It is not a single time point; it should be a period of time interval. We suggest the term ‘turning period’,” said Prof. Yuan.2
This “turning period” of the latter half of February is backed by the World Health Organization (WHO), whose own data appears to show that cases levelled off sometime during the week of 14 February 2020. WHO director-general Tedros Adhanom Ghebreyesus also stated in a press conference on 24 February 2020 that the epidemic in China “has been declining steadily” since 2 February 2020.6-7
In a follow-up analysis to determine the accuracy of their model, Prof. Yuan’s team collated official data from 11-29 February 2020. They highlighted these three findings:3-4
- Since 17 February 2020, the “Gamma” of the number of close contacts has declined rapidly to 4 percent (Table 1).
- Since 17 February 2020, the death rate has been kept within 3 percent (Table 2).
- Since 17 February 2020, the “Gamma” of the number of infected people has also declined to within 2–4 percent (Table 3).
In addition to the fact that 17–18 February 2020 had the highest number of confirmed cases in China (Table 4), “we have reason to believe that the inflection point has emerged on 17 and 18 February 2020, thus validating the prediction we made on 7 February 2020,” concluded Prof. Yuan.2-4
The accuracy of the model’s prediction is made all the more impressive considering that Prof. Yuan and his colleagues “basically used the daily data from 23 January to 6 February 2020, only three weeks data available,” he disclosed to APBN. Prof. Yuan further revealed that this iSEIR model they employed for the project was developed almost two years ago.
Entering the Second Half
While the value of deducing the turning point might not seem intuitive at first, Yuan took pains to underscore the importance of doing so. “Once the peak has been identified, a different approach has to be taken with regard to city management, economic development and epidemic control measures.” This is in line with President Xi Jinping’s remarks on 23 February 2020, where he stressed the need to resume production and normal life in an orderly way..
Prof. Yuan further added that since the COVID-19 virus could be spread asymptomatically and at the present date does not have an official treatment method nor a vaccine, making sure that the total numbers are under control via modelling would be reassuring to the public and stock markets. “We have definitely entered the second half in the fight against the virus,” said Prof. Yuan, using a sports analogy..4-5
Prof. Yuan’s viewpoint was shared by Bruce Aylward, an epidemiologist who led a 25-member team from WHO to conduct a 9-day field study trip to Beijing, Guangdong, Sichuan and Hubei. “They are using big data, artificial intelligence in places,” Aylward said, in a nod to Prof. Yuan’s work. “It’s a technology-powered and science-driven agile response at a phenomenal scale.”.6-7
Prof. Yuan’s team comprised Professor Di Lan from School of Artificial Intelligence and Computer Science in Jiangnan University, Engineer Gu Yudi from Information Construction and Management Center in Jiangnan University, Professor Qian Guoqi from Department of Mathematics and Statistics in University of Melbourne, in addition to Professor Qian Xiaosong and Professor Yue Xingye from Center for Financial Engineering in Soochow University, and a number of graduate students majoring in financial engineering and big data analysis with focus on Fintech and artificial intelligence. [APBN]
All graphs and figures presented here are credited to Professor George Yuan Xianzhi.
- iSEIR model: A dynamic epidemic model for rumor spread in multiplex network with numerical analysis (by George Yuan et al.), Working Paper (Internal Report), Soochow University (Suzhou, China), April 2018. http://arxiv.org/abs/2003.00144.
- “A brief explanation for the concept called ‘Turning Period’ for COVID-19 in China”, February 7, 2020. Retrieved from: https://mp.weixin.qq.com/s/7OfL3g6z_Bb-R2yl7-GIYg.
- “The fight against the new coronavirus entered the second half”, February 16, 2020. Retrieved from: https://mp.weixin.qq.com/s/azAkcR7nzST-dppDHFUZ0g.
- “The Report on China COVID-19”, February 16, 2020. Retrieved from: https://news.windin.com/ns/findsnap.php?sourcetype=1&id=487860593&code=24A05CBC501C&show=wft&device=android&terminaltype=wft.m&version=6.2.1&share=wechat.
- “Report by Members on China COVID-19”, China Society of Systems Engineering, February 18, 2020. Retrieved from: http://www.sesc.org.cn/htm/article/article1199.htm.
- “China has changed course of COVID-19 outbreak through pragmatic approach: WHO expert said”, February 24, 2020. Retrieved from: https://english.sina.cn/news/2020-02-26/detail-iimxyqvz5827520.d.shtml?vt=4.
- “China-WHO Joint report on COVID-19”, February 16-24, 2020, February 29, 2020. Retrieved from: http://www.nhc.gov.cn/jkj/s3578/202002/87fd92510d094e4b9bad597608f5cc2c..shtml.
About the Researcher
Professor George Yuan Xianzhi is a distinguished professor from Center for Financial Engineering (with focus on Fintech and Financial Engineering), Soochow University (Suzhou, China), Chair Professor in Sun Yat-sen University (Guangzhou, China), and holding a few visiting professorship position in a few other universities, in addition to being a Thousand Talents Plan expert of Shanghai and Sichuan, China. He is a senior advisor for Systems Engineering Society of China, and also serves as the deputy director on the Committee for Financial Mathematics, Engineering, and Actuarial Insurance of the China Society for Industrial and Applied Mathematics (CSIAM). He is also the Editor-in-Chief for the International Journal of Financial Engineering published by World Scientific.