Comments (3)
Thanks for the bug report and the detailed examples. Can confirm that I can reproduce the effect. WIll investigate.
from umap.
Thanks again for spotting this bug. I just committed an update; it fixes the issue as far as I can tell. Would you be able to try it out too? Many thanks.
Here's more background. min_dist
is used together with spread
to tune how close together items should appear in the final embedding. An alternative way to achieve the same tuning is via parameters a
and b
. The latter pair actually take precedence. By default, a
and b
are None
in umap-learn
, so changing min_dist
(and/or spread
) is an effective way of altering the embedding. But if a
and b
are set, values assigned to min_dist
or spread
become irrelevant.
The bug in the R package arose because NA
s for a
and b
in the configuration object were mistakenly translated into python number-like values. Those number-like values for a
and b
made min_dist
inactive; calculations with different min_dist
produced equivalent coordinates. The fix avoids sending NA
values to umap-learn
, thus avoids setting a
and b
, and thus enables min_dist
to take effect.
from umap.
Looking great also in my code! Thank you!
from umap.
Related Issues (20)
- predict() on a umap object with n_components=1 gets two errors -- Looks like missing drop=F HOT 2
- Failed creating initial embedding; using random embedding instead HOT 3
- Intel MKL FATAL ERROR HOT 3
- Differences with Python version? HOT 4
- Add support for umap-learn 0.5 HOT 4
- Sparse Matrix support HOT 4
- is there any spark version implementations? HOT 2
- missing value where TRUE/FALSE needed HOT 3
- Problem with using custom metric HOT 2
- umap() produces matrix instead of S3 object HOT 2
- method = "python" does not work HOT 1
- when random_state is set automatically in config, it is not sufficient for reproducibility HOT 1
- Citing the package HOT 1
- Type error in optimize_embedding HOT 3
- transforming new data to an embedding HOT 3
- Error with n_components=1 HOT 3
- Number of threads HOT 3
- Allow for supervised/semi-supervised dimension reduction with labels HOT 1
- predict() generates different predictions if called with multiple points at once versus called with each point individually HOT 7
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from umap.