# How can they predict the shape of the bell?

#### luchins

##### Member
How can they predict the shape of the bell?
(distribution of the possible numbers of infections over the days)

During this corona virus I have seen many pics of normal distributions which should predict the number of the infections over the days.
As predictive model I only know the regression.

My question is:WHere do These distribution predicting the number of infections of corona virus over the day come from?

Are they calculated based on a previous regression made or do they use another kind of model to compute them?I have seen that they also have confidence intervals all over the line that draws the ''bell shape'' as if it was a regression line.. but it was a distributionCan someone explain to this statistic ignorant?

Last edited:

#### staassis

##### Active Member
This is not a question for outside experts. Ask the newspapers and blogs that published those histograms. Whatever they did, it was probably very simple. In general, calculating histograms on time series data has its flaws.

Last edited:

#### hlsmith

##### Less is more. Stay pure. Stay poor.

If you are referencing the epidemic curves, which they are trying to flatten - this is different from what you are thinking. The curve represents the increasing number of cases over time. So the numbers will increase until its transmission starts to decrease due to distancing, hygiene and a decrease in the number of susceptible individuals around you. So each person in the incline spreads the infection to greater than 1 person, thus the positive slope, but at some point people are going to spread it to only one person then less than one person and the number of cases will decrease representing the decline in the curve. The 'curve' is based on differential calculus, so fitting the spread and number of cases across time.

You should look up compartmental models (SIS, SIR, SEIR, etc.) to see how the calculation occurs and also the effective reproduction rate (R_E).