Variables needed in the cateogrical method

[maude-v]

Hi everyone,
To start working on the categorical method, we need to check what parameters we need to be able to implement in the functions.
Here is the basis that we discussed, @canessas if you think of other parameters can you tell us? Thanks!

topics (third layer, i.e. mechanisms in our manuscript):

• there can be from 1 to infinite number of topics
• default names = topic 1, topic 2, topic 3, ...
• can be named by the user

sites (second layer):

• there can be from 1 to infinite number of sites
• the number or name of the sites does not have to be the same for every topic
• default names = site 1, site 2, site 3, ...
• can be named by the user

levels (first layer):

• from 2 to infinite number of levels, default is 5.
• the number or name of the levels does not have to be the same for every topic. Does it have to be same for every site inside a topic? (i would say yes)
• default names = if 5 levels: c("minimal", "minor", "moderate", "major", "massive"). if less or more levels: c(1,2,3,4,...)
• can be named by the user

experts

• there can be from 1 to infinite number of experts
• the number or name of the experts does not have to be the same for every topic. Does it have to be same for every site inside a topic?

estimates (first layer):

• probabilities (0-1, must sum to one for all the levels)
• confidence (0-100, given in percent)

votes

• for bootstrapping method, the number of times you want to multiply the weight of experts by

other potential ideas

• turn on or of dirichlet (i.e. use dirichlet priors or not)
• allow users to apply only one method (chose either basic method or bootstrapping method instead of both together)

[sgvignali]

@maude-v Should we not better name "mechanisms" instead of "topics" to be consistent with the manuscript?

[maude-v]

Yes we can, I thought topic was more general, but if we're keeping sites anyway we might as well go with mechanisms too so we can justify the names 👍