XGBoost
- Trees
- 100.
- Learning rate
- 0.3.
- Depth
- 6.
Loading…
XGBoost fits 100 trees of depth 6 at learning rate 0.3. Gradient boosting fits 100 trees of depth 3 at learning rate 0.1. The published score is hold-out accuracy on a stratified 25% split.
Do you want depth 6 at learning rate 0.3?
That is the published XGBoost fit. The other published fit is depth 3 at learning rate 0.1.
Gradient boosting has the higher hold-out accuracy on Iris Flowers and Customer Segments; Hold-out accuracy is within 0.01 on Wine (UCI wine recognition).
Choose the score the table needs.
This is a choice, not a measured winner.
Iris Flowers lists 150 records and 4 inputs. XGBoost, objective multi:softprob, ran 100 rounds at learning rate 0.3 and depth 6. Its hold-out accuracy is 0.921 and its macro F1 is 0.923, measured on 38 test rows. Gradient boosting, loss log_loss and sqrt features per split, ran 100 rounds at learning rate 0.1 and depth 3. Its accuracy is 0.947 and its macro F1 is 0.949 on the same rows. The scatter for each fit is the test-row PCA view from the published run. Gradient boosting has the higher hold-out accuracy on this split.
XGBoost used 100 trees, depth 6, and learning rate 0.3, with no early stopping, on about 112 training rows. Gradient boosting used 100 trees, depth 3, and learning rate 0.1, with no early stopping. That is why 0.947 beat 0.921 on this split.


The link loads the sample when the classification tool has it. Run is still required.
Wine (UCI wine recognition) lists 178 records and 13 inputs. XGBoost, objective multi:softprob, ran 100 rounds at learning rate 0.3 and depth 6. Its hold-out accuracy is 1.000 and its macro F1 is 1.000, measured on 45 test rows. Gradient boosting, loss log_loss and sqrt features per split, ran 100 rounds at learning rate 0.1 and depth 3. Its accuracy is 1.000 and its macro F1 is 1.000 on the same rows. The scatter for each fit is the test-row PCA view from the published run. Hold-out accuracy is within 0.01 on this split.


The link loads the sample when the classification tool has it. Run is still required.
Customer Segments lists 200 records and 5 inputs. XGBoost, objective multi:softprob, ran 100 rounds at learning rate 0.3 and depth 6. Its hold-out accuracy is 0.860 and its macro F1 is 0.858, measured on 50 test rows. Gradient boosting, loss log_loss and sqrt features per split, ran 100 rounds at learning rate 0.1 and depth 3. Its accuracy is 0.880 and its macro F1 is 0.879 on the same rows. The scatter for each fit is the test-row PCA view from the published run. Gradient boosting has the higher hold-out accuracy on this split. The classification tool does not load this sample; the published runs used the segment column that ships with the table.
XGBoost used 100 trees, depth 6, and learning rate 0.3, with no early stopping, on about 150 training rows. Gradient boosting used 100 trees, depth 3, and learning rate 0.1, with no early stopping. That is why 0.880 beat 0.860 on this split.


The link loads the sample when the classification tool has it. Run is still required.
| Dataset | XGBoost accuracy | XGBoost macro F1 | Gradient boosting accuracy | Gradient boosting macro F1 | Test rows |
|---|---|---|---|---|---|
| Iris Flowers | 0.921 | 0.923 | 0.947 | 0.949 | 38 |
| Wine (UCI wine recognition) | 1.000 | 1.000 | 1.000 | 1.000 | 45 |
| Customer Segments | 0.860 | 0.858 | 0.880 | 0.879 | 50 |
A new row is labeled by the fitted XGBoost or the fitted gradient boosting.
XGBoost fits 100 trees of depth 6 at learning rate 0.3. Gradient boosting fits 100 trees of depth 3 at learning rate 0.1. The published score is hold-out accuracy on a stratified 25% split.
Method: StandardScaler, drop constant columns. Comparison code version 8fbe69461da4 (scripts/seo/build_compare_runs.py). scikit-learn 1.4.2. XGBoost 2.1.4. Seed 42. XGBoost and gradient boosting were fit on 2026-11-12. The class column is the training target. Scores are the stratified 25% hold-out. The shared run applies StandardScaler, and a tree split does not use that scale. Iris Flowers: UCI Machine Learning Repository, Iris (Fisher, 1936). License: CC BY 4.0. Wine (UCI wine recognition): UCI Machine Learning Repository, wine recognition. License: CC BY 4.0. Customer Segments: Synthetic customer records bundled with this site. License: Generated for demos; no external license.
Results computed, published, and last updated on 2026-11-12.
The comparison sentences are filled from these published runs by a fixed template, so the scores can be checked against the same datasets in the tool.
Maintainer: nikolaou1982.
Contact: nick1982100@gmail.com.