[ML] Line search feature bag fraction for classification and regression model training #1746
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
CI on macos is expected to fail at the moment so I'll go ahead and merge. |
tveasey
added a commit
to tveasey/ml-cpp-1
that referenced
this pull request
Feb 19, 2021
…on model training (elastic#1746) Following on from elastic#1733, we can get further speedups by line searching for the best feature bag fraction for data sets where we only need a fraction of the features per tree. For example, training time on Higgs 1M drops from 2585s to 1742s and we actually get a small improvement in accuracy because our hyperparameter search region is better initialised. This makes three changes: 1. Adds a line search for the best initial feature bag fraction to use. 2. Adds a small linear penalty at most 1% minimum loss to encourage larger down sample factors and smaller feature bag fractions. 3. Handles better the case we have many features and relatively few training examples.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Following on from #1733, we can get further speedups by line searching for the best feature bag fraction for data sets where we only need a fraction of the features per tree. For example, training time on Higgs 1M drops from 2585s to 1742s and we actually get a small improvement in accuracy because our hyperparameter search region is better initialised.
This makes three changes:
Some of the variable naming in
CBoostedTreeFactoryis also now misleading. I've split this change out into a non-functional commit. The functional commit is f9eacc2.