[SPARK-4711] [mllib] [docs] Programming guide advice on choosing optimizer #3569
Conversation
…osing an optimization method
|
Test build #24073 has started for PR 3569 at commit
|
|
Test build #24073 has finished for PR 3569 at commit
|
|
Test PASSed. |
| However, different optimization methods can have different convergence guarantees depending on the properties of the objective function, and we cannot cover the literature here. | ||
|
|
||
| * L-BFGS is recommended since it generally converges faster (in fewer iterations) than SGD. | ||
| * SGD can be faster for datasets with a very large number of instances (rows), especially when using a small `miniBatchFraction`. |
There was a problem hiding this comment.
This part might not be true because we implemented mini-batch SGD but obtaining a mini-batch from an RDD is expensive, which requires one pass, while computing the gradient is not super expensive. Maybe we can also mention this trade-off.
|
@mengxr Thanks for taking a look! Updated based on your comment. |
| @@ -359,13 +362,15 @@ public class LBFGSExample { | |||
| {% endhighlight %} | |||
| </div> | |||
| </div> | |||
| #### Developer's note | |||
|
|
|||
There was a problem hiding this comment.
I think this caused a .md generation problem in the old docs.
|
Test build #24108 has started for PR 3569 at commit
|
|
Test build #24109 has started for PR 3569 at commit
|
|
Test build #24108 has finished for PR 3569 at commit
|
|
Test PASSed. |
|
Test build #24109 has finished for PR 3569 at commit
|
|
Test PASSed. |
jkbradley
changed the title
…mizer I have heard requests for the docs to include advice about choosing an optimization method. The programming guide could include a brief statement about this (so the user does not have to read the whole optimization section). CC: mengxr Author: Joseph K. Bradley <[email protected]> Closes #3569 from jkbradley/lr-doc and squashes the following commits: 654aeb5 [Joseph K. Bradley] updated section header for mllib-optimization 5035ad0 [Joseph K. Bradley] updated based on review 94f6dec [Joseph K. Bradley] Updated linear methods and optimization docs with quick advice on choosing an optimization method (cherry picked from commit 27ab0b8) Signed-off-by: Xiangrui Meng <[email protected]>
|
LGTM. Merged into master and branch-1.2. Thanks! |
I have heard requests for the docs to include advice about choosing an optimization method. The programming guide could include a brief statement about this (so the user does not have to read the whole optimization section).
CC: @mengxr