New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Largest connected set returns a state when it should return an empty set #175
Comments
Yeah I think you are right. This is an error in the |
Well... so one could say that the set of states consists of 4 independent, disconnected sets ( Some thoughts: I could imagine that returning empty sets may break something downstream. It's not technically wrong to have a 1-state model, so I wouldn't necessarily raise an error. |
I would agree except for the connectivity threshold - if it is set to a value that is greater than what can be found on the count matrix diagonal then the one-state set should be discarded I think. |
Like think about the |
As far as I remember, connectivity is only defined by transitions between different states, excluding self-transitions. In the graph sense: Two nodes are connected if there is an edge between them - the node does still exist in the graph if it's disconnected from everything else. So in the above trajectory, the connected sets with a connectivity threshold of >1 would still be |
Yeah to confirm: I just looked this up, singleton nodes are by definition connected components. Whether it is desirable to eliminate these if the weight of its loop is below connectivity threshold is a philosophical (and practical) consideration. Perhaps this should be documented somewhere more clearly. |
In that case, would it make sense to at least return the singleton set that contains the most samples? |
If we define it to be that way then yes it would make sense, I don't see the the immediate benefit from it though, can you elaborate? |
If I want to restrict my samples to the largest connected set, I can imagine I would want to 'keep' as many samples as possible. But I guess it doesn't really matter because this is not a realistic use case. Just something I came accross while testing. So just updating the docs sounds good. |
Alright, docs updated. In case of sets with more than one element it's not so clear how to order them anyways. |
The following should print the empty set
[]
but instead prints[3]
.The text was updated successfully, but these errors were encountered: