Statistics and R software
Blog by JMSS
 

COVID-19: The Case of Spain

ESTA ENTRADA NO TIENE ACTUALIZACIONES EN ESTA WEB / THIS ENTRY DOES NOT HAVE UPDATES ON THIS WEB

HA SIDO MOVIDA A / HAS BEEN MOVED TO: https://github.com/josemss/Covid19

————————

(1) SIR MODEL

I will use the same model used in this post: Epidemiology: How contagious is Novel Coronavirus (2019-nCoV)?. You can find all the details there and in the comments. It’s a SIR model (https://es.wikipedia.org/wiki/Modelo_SIR).

The results should not be taken seriously since the absence of real data makes the model inconsistent. This entry was created for informational purposes on SIR models. I will try to update it daily. More reliable information can be found at https://www.imm.upv.es/covid-19/

Daily data: Wikipedia

(click plots to zoom)  Corona_datos_01 Corona_datos_02

Since we do not know the true data of infected people, an important variable to take into account is the daily increase in infected people. In the following graph we observe this increase seems to have a downward trend in recent days (despite variability increases)

(click plots to zoom)

Corona_datos_03

Estimates with the SIR model:

(click plots to zoom)

SIR-model_01SIR-model_02

According to this model, the rate of infection is 2.07, the height of the pandemic will be reached by 12/04/2020.

About 7778775 people would be infected by then, which translates to about 1439073 severe cases, about 388939 cases in need of intensive care and up to 54451 deaths.

Infected forecast for tomorrow: 35272

21mar -> predicted=24896 ; observed=25374 ; error=478 (1.8%)

22mar -> predicted=30117 ; observed=28572 ; error=-1545 (5%)

23mar -> predicted=35272; observed=33089; error=-2183 (6.6%)

(2) REGRESSION MODEL

I use a cubic polynomial regression model. See Wikipedia.

The fit of the model is very good since the coefficient R2 is very close to 1.

Infected forecast for tomorrow: 37873 with confidence (36831.19 , 38914.46)

(click plots to zoom)

Poly_model_01

#StayAtHome #QuedateEnCasa

jose

About jose

Profesor titular del Dpto. de Estadística e I.O. de la Usal. Entusiasta del software estadístico R.
Aún no hay comentarios.

Deja un comentario


*

Política de privacidad