DOC Added information about space complexity to docs DBSCAN#26783
DOC Added information about space complexity to docs DBSCAN#26783adrinjalali merged 4 commits intoscikit-learn:mainfrom
Conversation
|
I think you used Alt+Q to format the code (line length) and it has formatted the whole docstring, resulting in a larger than needed diff. Could you please redo those? You can also set it in the settings of your vscode that Alt+Q only formats the current paragraph instead of the whole section. |
|
Okay, did it. Thanks for your support! |
|
I'm wondering if this is good as is, or if it should be a note ( |
|
I've added a sentence to clarify what |
Co-authored-by: Guillaume Lemaitre <g.lemaitre58@gmail.com>
…earn#26783) Co-authored-by: Guillaume Lemaitre <g.lemaitre58@gmail.com>
…earn#26783) Co-authored-by: Guillaume Lemaitre <g.lemaitre58@gmail.com>
Co-authored-by: Guillaume Lemaitre <g.lemaitre58@gmail.com>
…earn#26783) Co-authored-by: Guillaume Lemaitre <g.lemaitre58@gmail.com>
|
The claim "The worst case memory complexity of DBSCAN is O(n²), which can occur when the (Original) DBSCAN has linear memory requirements, and worst-case quadratic distance computations. The O(n²) memory use is due to the sklearn implementation, as it was already documented in the "Notes" just a bit further down. In #28493 I propose a revised claim. |
Reference Issues/PRs
#26726
What does this implement/fix? Explain your changes.
Added information about space complexity to docstring; because users were wondering about the huge RAM usage if param
epsis high, while parammin_samplesis low.