-
Notifications
You must be signed in to change notification settings - Fork 1
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
Improve documentation and introduce simple synthetic examples #4
Conversation
Thanks, @slwu89, for the fantastic tutorials. Two reasons explain why other designs were not displayed: Firstly, even after running experiments Secondly, a filter for "degenerate" designs with too high cost existed in: https://github.com/Merck/CEED.jl/blob/6d932f6a69bdb7906ae8c1dd82c9a1008c336010/src/fronts.jl#L108-L110 I have now removed this condition. |
It would be great to eventually feature the Mahalanobis distance (#7) in the synthetic examples as well. |
Remove the condition on the maximum design cost in `plot_front`. Fix `SimpleGenerative.jl` tutorial.
@thevolatilebit, I've included Mahalanobis distance (#7) in the generative tutorial now. Still need to spend more time on the explanatory text before it's ready for your review. |
@thevolatilebit can you help me take a look at tutorials/SimpleGenerative.jl and tutorials/SimpleStatic.jl and make sure that they are set up correctly with the rest of the Project.toml files and things, for building docs? When I run the
|
@slwu89 Yeah, it is necessary to add all the dependencies to |
Thanks. |
Optimizes code in `distancebased.jl`. Changes the way that an "uncertainty" is provided. That is, while "distances" and "similarities" are provided by a user through a function call, an "uncertainty" was originally given as a type. Fixes #13
- Implemented a variant of sq. Mahalanobis distance with missing entries, see https://www.jstor.org/stable/3559861 on page 285, fixes #12 - Renamed `MahalanobisDistance` to `SquaredMahalanobisDistance` - Minor adjustments to tutorials - Removed extra dependencies - Incremented version Fixes #11 and #12
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Excellent job, thank you, @slwu89 !
Hi @thevolatilebit this is a draft PR to address some of the documentation topics we discussed in the enterprise internal repo. As a reminder, what I would eventually like this PR to include is:
The synthetic example for static designs works great and is producing expected results with the input data.
That all being said, I am having trouble getting synthetic examples with known properties to work for the generative designs. In this file https://github.com/Merck/CEED.jl/blob/8786d768198bad2f1d5b77013019c93291760c05/tutorials/SimpleGenerative.jl I have attempted to set up an example with both a continuous and discrete outcome. However, for both cases, the MDP solution is always the empty set, despite the features contributing information to predicting the outcome, by construction. @thevolatilebit I am wondering if you can help me look at those 2 examples to figure out why this is the case, I am very curious why I cannot get different actions proposed.