Description

The article previously written on the failed simulation of the pandemic showed a poor adequacy related to the poor representativeness of the data. Today the tests being more numerous it seems interesting to repend the calculations of the recalibration of the SIR modeling by optimization with constraints. First, let's look at the raw data provided by the Politologist. Note that the first day corresponds to August 25, 2020 and that the 65th day is that of confinement (Friday, October 30). There is a seasonalization on the weekend, the peak, attributed to Sunday, showing a significant decrease in infections certainly comes from the decrease in weekend-related analysis. These points, indicated by a red cross, were offset by interpolation, indicated by a green circle. Red circles are anomalies that are passed 'by hand'. The first anomaly is due to a reported counting error (day 72). And the 84th day also appears as an anomaly but not explained. It is on this corrected basis that the calculations that will be presented in a future article will be made. The second figure shows on the last Sunday of September a decrease in infected individuals that is clearly visible when we overlay the real data with the simulated data in blue. The third curve at the bottom, brings a correction on the actual data of +4000 individuals per day from the date indicated to compensate for this decrease which does not seem to be a change in the dynamics of the epidemic. There is a better match between raw and simulated data. Comparison is not right, but it raises questions about the quality of the data published by the Politologist, and it would be interesting to find the cause of this anomaly which can have consequences on the quality of the predictions provided by the simulation whose re-alignment is based on these data.

Resources

Name	Format	Description	Link

Tags

Topics

Categories