Some ideas on the mathematical model of COVID-19 epidemic
Next, we will use the data of suspected cases to make predictions. The reasons are as follows:
- Clinical diagnosis will be included in the number of confirmed cases, which means that the number of confirmed cases is disturbed, and the degree of interference is very large~
- Confirmed cases come from suspected cases. In the diagnosis and treatment plan for pneumonia caused by novel coronavirus infection (trial version 5), the “clinical diagnosis” is added to the case diagnosis classification of Hubei Province. The clinical diagnosis is from the suspected cases, and the confirmed cases are from the clinical diagnosis and the suspected cases.
- The diagnosis process of confirmed cases is limited by hardware conditions (such as the number of nucleic acid reagents and other medical equipment), and the number of missed cases should be large; The criteria for judging suspected cases are lower. As long as their epidemic history and whether they have fever, the number of missing cases should be less.
Therefore, we will use antigen detection to judge suspected cases for subsequent prediction.
Step 1: capture data
From Tencent’s real-time epidemic data, we captured the daily increase data. The captured results are as follows:
Step 2: estimate the actual number of confirmed cases
We need to do the following:
- Estimate the proportion of confirmed cases in suspected cases
- Determine the number of confirmed patients according to the daily increase data
- According to the SI model, determine the final confirmed number and the approximate time of the end of the epidemic
We will first look at how to estimate the proportion of confirmed cases in suspected cases through existing data.
Second, we determined the number of confirmed cases according to the daily increase of suspected cases.
- To determine the total number of suspected cases per day, we need to accumulate and function cumsum ()
- To determine the total number of confirmed cases per day, we only need to multiply 0.5960971132108407 by the total number of suspected cases per day
It can be seen that our predicted total number of confirmed cases per day is higher than the official one, but it has started to overlap in recent days. In fact, it can be understood that under the current state’s high attention and full publicity, individuals and governments at all levels no longer hide information, so it can be considered as true data.
If all suspected cases are turned into confirmed cases, the total number of actual infections is more than 70000, as shown by the green line~