search

crowdfightcovid19 / req-550-Syria

Repository to host the code of request 550 coming from the Pax Syriana Foundation
Other
3 stars 2 forks source link

Files for the null model #17

Closed apascualgarcia closed 4 years ago

apascualgarcia commented 4 years ago

@jordan-klein @Jennifer-Villers @ecam85

Since the null model has only five classes the files should contain only those. There is no problem for the script to run with any number of classes, the idea is that it automatically adjusts the number of variables needed reading the information contained in the files, computes the simulations and generates the figures, so if the input files already contain informative labels for the classes no action is required from the user.

I see two reasons why we should not run the null model with the same number of classes than with the interventions. Firstly, because having more variables we will introduce unnecessary non-linearities in the simulations. Secondly, because as I said it will directly generate plots with the classes included in the input files, so it will be misleading to present orange/green classes when they don't exist or to modify the code to create the plots when now it is flexible.

jordan-klein commented 4 years ago

Fixed so the null model has only 5 classes.

apascualgarcia commented 4 years ago

Hi @jordan-klein I was creating the same directory at the same time than you but with another name and after pulling your changes I've lost the old null model (the one with 7 classes). I think that to compute the contact matrix of the interventions the easiest way to go is to consider the null model with 7 classes, even if we do the simulations with the model with 5 classes only. Would be possible for you to recover the one with 7 classes? In the README file I document how the final structure of directories could look like.

apascualgarcia commented 4 years ago

Both null models are present in the directories now. Thanks Jordan for the help.