You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
I think there are many instances when a user doesn't know which specific outcome they are looking for, e.g. they don't know the exact Survival Time threshold value to pick for the query. Instead, I think they would like information about the entire posterior distribution of the outcome.
I also link a simple boolean linking and updating result is too course in returning probabilities. Returning the complete posterior of the target outcome would allow for a finer grain analysis.
Describe the solution you'd like
Allow an option to be set or a query node attribute so that the full posterior distribution of the target outcome is returned. If this outcome is a continuous variable like Survival Time, create a discretized outcome in the patient data to a given bin size (specified by the user, defaulted or learned). Link the BKB based on this new discretized target and return the entire posterior distribution in the knowledge and result query graphs.
Describe alternatives you've considered
Run multiple queries each a part of the posterior like it is now and then have the user aggrate these results appropriately. However, this approach is not equivalent to what is proposed above.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
I think there are many instances when a user doesn't know which specific outcome they are looking for, e.g. they don't know the exact Survival Time threshold value to pick for the query. Instead, I think they would like information about the entire posterior distribution of the outcome.
I also link a simple boolean linking and updating result is too course in returning probabilities. Returning the complete posterior of the target outcome would allow for a finer grain analysis.
Describe the solution you'd like
Allow an option to be set or a query node attribute so that the full posterior distribution of the target outcome is returned. If this outcome is a continuous variable like Survival Time, create a discretized outcome in the patient data to a given bin size (specified by the user, defaulted or learned). Link the BKB based on this new discretized target and return the entire posterior distribution in the knowledge and result query graphs.
Describe alternatives you've considered
Run multiple queries each a part of the posterior like it is now and then have the user aggrate these results appropriately. However, this approach is not equivalent to what is proposed above.
The text was updated successfully, but these errors were encountered: