Closed aiday-mar closed 2 years ago
Merging #37 (8489ef4) into master (229e93a) will decrease coverage by
18.36%
. The diff coverage is11.47%
.:exclamation: Current head 8489ef4 differs from pull request most recent head 8a26a53. Consider uploading reports for the commit 8a26a53 to get more accurate results
@@ Coverage Diff @@
## master #37 +/- ##
===========================================
- Coverage 91.38% 73.02% -18.37%
===========================================
Files 4 4
Lines 209 278 +69
Branches 36 63 +27
===========================================
+ Hits 191 203 +12
- Misses 17 61 +44
- Partials 1 14 +13
Impacted Files | Coverage Δ | |
---|---|---|
src/RandomForestClassifier.js | 95.45% <ø> (+0.10%) |
:arrow_up: |
src/RandomForestRegression.js | 87.50% <ø> (+0.54%) |
:arrow_up: |
src/utils.js | 90.90% <ø> (-2.75%) |
:arrow_down: |
src/RandomForestBase.js | 55.55% <11.47%> (-33.20%) |
:arrow_down: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 229e93a...8a26a53. Read the comment docs.
The variable this.gainThreshold in the TreeNode implementation was not initialized properly and this was leading to an incorrect creation of regression tree. This will be adressed in a separate pull request for the ml-cart module.
Added the possiblity to change the number of samples one wants to use in the training phase
Added a function featureImportance in order to compute the feature importance as was suggested by @xdadda