important question especially in the current time

Dear Baydaa

Thank you for your message. I am hoping somebody who has some insight shares her/his knowledge.

According to morbidity, mortality and recovery rates, I think we should not feel panic of COVID-19, we should just follow hygienic measures, and work on developing vaccine.

The most cases about 85% are mild

Thanks Mohammed

i needed for any new information to this virus

The mortality rate will be very low and hence the situation will become normal in the globe soon.

Dear DR Veeman,

Thank you for your reaasurance.

Dear all,

The following document gives a glimpse of modeling.

http://handelgroup.uga.edu/files/presentations/2019SISMID_1_Modeling_Introduction.pdf

Dear Vadim,

Thank you for the insight.

I think one can use the Bass diffusion model to predict the total number of new daily cases. The graph I attached shows the prediction (as of today) I obtained using the Bass model. I guess this approach is similar to what you have on slide 15.

But if you want to predict the probability of new cases in a country where it has not yet occurred, I think you can use something like a random forest to estimate the probability of a new case tomorrow given a number of features for that country (population, GDP, number of flights per day, number of active cases in countries near that country (perhaps, with lags, etc.). But it will require 1) good database of existing cases, 2) considerable efforts invested in modelling. Also historical data will not be very representative as now many countries are better prepared to prevent the spread of the disease.

P.S. Of course, in order to obtain best forecasts you need to combine the estimates based on the historical data available with the information provided by experts and maybe then you can use the approach presented in this paper:

Technical Report A joint Bayesian forecasting model of judgment and observed data

One problem with the above forecast obtained using the Bass diffusion model is shows point estimates but the uncertainty associated with these estimates is very high (it is not shown on the graph, but the prediction intervals are very wide, so the forecast becomes not very reliable),

The approach presented in the paper allows using both estimates obtained from experts and the historical data in order to obtain better forecasts with lower levels of uncertainty.

See, e.g., examples on page 19:

Technical Report A joint Bayesian forecasting model of judgment and observed data

When there's no or little data, forecasts are mainly based on what experts say. But then these forecasts are updated when new data become available.

P.P.S. It is also very important to evaluate the estimates obtained from experts (usually such estimates should be calibrated and transformed into adequate probabilistic representation of future events). In this regard, maybe it would be interesting to look at the methods presented in Chapter 3 of my thesis:

Thesis Integration of judgmental and statistical approaches for dem...

For example, if you have estimates of the number of new cases obtained from experts, you can then try to transform those estimates into density forecasts.

Dear all,

An Italian researcher have fitted the Italian fatality to the logistic curve. see below

https://towardsdatascience.com/covid-19-infection-in-italy-mathematical-models-and-predictions-7784b4d7dd8d

Dear Andrey

Thank you for your contribution

Dear Andrey

I have downloaded your paper to read. Tanks.

COVID-19 is provided us a motivation to learn epidemiology and virology on our feet; while not scratching our face. We are learning the basics of good science: understanding how data is collected, and make sense of it; accepting that we can answer our inquiries by approximation and proxy.

Sirous Yasseri You're welcome.

I think instead fitting the logistic curve as was done in the link you sent a reasonable approach. As far as I understand, fitting the logistic curve is a special case of fitting the Bass diffusion model (see my graph above).

The new graph I attached shows the forecast I obtained using the diffusion model 2 days ago and the most recent two observations that are very close to what was predicted. So till this moment the diffusion model proves accurate (there are yet only two observations though). But generally it is known that the number of new cases of a disease can be described using a bell shaped curve. This was true for SARS as well.

Dear Andrey,

Thanks.

good information

Hi Baydaa,

I was hoping people with more insight share their knowledge.

Dear Balqees,

Thank you for your comment.

One thing i know,

should careful about this virus.

A colleague of mine believes that the number of people affected by Corona and the number of deaths are most probably incorrect for some countries. Simply governments hid it or don't want to know.

Here are his reasons

• There are not enough necessary kits to test to see who is positive for Corona
• The case of death deliberately of mistakenly is attributed to influenza
• They think the nation point out their incompetence.
• Mistakenly believe that they are exception and control Corona
• There are enough hospital beds to cater for ill people. If numbers are high then they have an excuse.
• ..................

This sound as a conspiracy theory. However, the progress of disease I followed indicates he is right.

I found this article to provide reasonable good analysis of the available data on this pandemic

https://medium.com/@tomaspueyo/coronavirus-act-today-or-people-will-die-f4d3d9cd99ca

This MIT and Harvard-Built App Could Slow the Spread of Coronavirus

https://futurism.com/coronavirus-app-mit-safe-paths

Good question

In this time , this is agood question

https://www.kaggle.com/sudalairajkumar/novel-corona-virus-2019-dataset/tasks?taskId=508

Here you may find some data. Not sure how to model it.

Dear Dr. Vrdoljak

Thank you for your contribution.

Dear DR Al-Dulamey

Thanks.

Dear Dr. Sen

Thank you for the info.

Agree with the comments of Rasha shakir

Agree with the Anton Vrdoljak answer

I have prepared a model about this topic but it is under processing sir.

Hi Emad,

I wait until you are ready to send me your model.

If you can read Farsi, I can send you a paper by an Iranian researcher who has used the system dynamics approach to model the effect of various measures to slow down the progress of epidemi.

Dear Rodolfo,

I don't understand your comment.

Would somebody translate it into English?

The book Blindness by José Saramago guides readers through an immersive experience of a terrifying predicament.

A driver waiting at the traffic lights goes blind. An opthamologist tries to diagnose his distinctive white blindness but is affected before he can read the textbooks. It becomes a contagion, spreading throughout the city. Trying to stem the epidemic, the authorities herd the afflicted into a mental asylum where the wards are terrorised by blind thugs. And when fire destroys the asylum, the inmates burst forth and the last links with a supposedly civilised society are snapped

Read this book while in quarntine

I think it used computational models, These computational models use well-known statistical equations that calculate the probability of transmission of the disease from infected to others.

Dear Omar

Thank you for your answer.

best question

Dear Miss Hassan,

Thanks.

good luck

I made a lot of progress using the system dynamics approach.