Hyperparameter tuning#
In the previous section, we did not discuss the hyperparameters of random forest and histogram gradient-boosting. This notebook gives crucial information regarding how to set them.
Caution
For the sake of clarity, no nested cross-validation is used to estimate the variability of the testing error. We are only showing the effect of the parameters on the validation set.
We start by loading the california housing dataset.
from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import train_test_split
data, target = fetch_california_housing(return_X_y=True, as_frame=True)
target *= 100 # rescale the target in k$
data_train, data_test, target_train, target_test = train_test_split(
data, target, random_state=0
)
Random forest#
The main parameter to select in random forest is the n_estimators parameter.
In general, the more trees in the forest, the better the generalization
performance would be. However, adding trees slows down the fitting and prediction
time. The goal is to balance computing time and generalization performance
when setting the number of estimators. Here, we fix n_estimators=100, which
is already the default value.
Caution
Tuning the n_estimators for random forests generally result in a waste of
computer power. We just need to ensure that it is large enough so that doubling
its value does not lead to a significant improvement of the validation error.
Instead, we can tune the hyperparameter max_features, which controls the
size of the random subset of features to consider when looking for the best
split when growing the trees: smaller values for max_features lead to
more random trees with hopefully more uncorrelated prediction errors. However
if max_features is too small, predictions can be too random, even after
averaging with the trees in the ensemble.
If max_features is set to None, then this is equivalent to setting
max_features=n_features which means that the only source of randomness in
the random forest is the bagging procedure.
print(f"In this case, n_features={len(data.columns)}")
In this case, n_features=8
We can also tune the different parameters that control the depth of each tree
in the forest. Two parameters are important for this: max_depth and
max_leaf_nodes. They differ in the way they control the tree structure.
Indeed, max_depth enforces growing symmetric trees, while max_leaf_nodes
does not impose such constraint. If max_leaf_nodes=None then the number of
leaf nodes is unlimited.
The hyperparameter min_samples_leaf controls the minimum number of samples
required to be at a leaf node. This means that a split point (at any depth) is
only done if it leaves at least min_samples_leaf training samples in each of
the left and right branches. A small value for min_samples_leaf means that
some samples can become isolated when a tree is deep, promoting overfitting. A
large value would prevent deep trees, which can lead to underfitting.
Be aware that with random forest, trees are expected to be deep since we are seeking to overfit each tree on each bootstrap sample. Overfitting is mitigated when combining the trees altogether, whereas assembling underfitted trees (i.e. shallow trees) might also lead to an underfitted forest.
import pandas as pd
from sklearn.model_selection import RandomizedSearchCV
from sklearn.ensemble import RandomForestRegressor
param_distributions = {
"max_features": [1, 2, 3, 5, None],
"max_leaf_nodes": [10, 100, 1000, None],
"min_samples_leaf": [1, 2, 5, 10, 20, 50, 100],
}
search_cv = RandomizedSearchCV(
RandomForestRegressor(n_jobs=2),
param_distributions=param_distributions,
scoring="neg_mean_absolute_error",
n_iter=10,
random_state=0,
n_jobs=2,
)
search_cv.fit(data_train, target_train)
columns = [f"param_{name}" for name in param_distributions.keys()]
columns += ["mean_test_error", "std_test_error"]
cv_results = pd.DataFrame(search_cv.cv_results_)
cv_results["mean_test_error"] = -cv_results["mean_test_score"]
cv_results["std_test_error"] = cv_results["std_test_score"]
cv_results[columns].sort_values(by="mean_test_error")
| param_max_features | param_max_leaf_nodes | param_min_samples_leaf | mean_test_error | std_test_error | |
|---|---|---|---|---|---|
| 3 | 2 | None | 2 | 34.054779 | 0.508816 |
| 0 | 2 | 1000 | 10 | 36.960128 | 0.674312 |
| 7 | None | None | 20 | 37.281536 | 0.447023 |
| 4 | 5 | 100 | 2 | 40.071118 | 0.437617 |
| 8 | None | 100 | 10 | 40.523289 | 0.568492 |
| 6 | None | 1000 | 50 | 40.876767 | 0.461232 |
| 9 | 1 | 100 | 2 | 50.031167 | 0.965327 |
| 2 | 1 | 100 | 1 | 50.248458 | 0.901105 |
| 5 | 1 | None | 100 | 54.462856 | 0.463538 |
| 1 | 3 | 10 | 10 | 54.943519 | 0.667540 |
We can observe in our search that we are required to have a large number of
max_leaf_nodes and thus deep trees. This parameter seems particularly
impactful with respect to the other tuning parameters, but large values of
min_samples_leaf seem to reduce the performance of the model.
In practice, more iterations of random search would be necessary to precisely
assert the role of each parameters. Using n_iter=10 is good enough to
quickly inspect the hyperparameter combinations that yield models that work
well enough without spending too much computational resources. Feel free to
try more interations on your own.
Once the RandomizedSearchCV has found the best set of hyperparameters, it
uses them to refit the model using the full training set. To estimate the
generalization performance of the best model it suffices to call .score on
the unseen data.
error = -search_cv.score(data_test, target_test)
print(
f"On average, our random forest regressor makes an error of {error:.2f} k$"
)
On average, our random forest regressor makes an error of 33.85 k$
Histogram gradient-boosting decision trees#
For gradient-boosting, hyperparameters are coupled, so we cannot set them
one after the other anymore. The important hyperparameters are max_iter,
learning_rate, and max_depth or max_leaf_nodes (as previously discussed
random forest).
Let’s first discuss max_iter which, similarly to the n_estimators
hyperparameter in random forests, controls the number of trees in the
estimator. The difference is that the actual number of trees trained by the
model is not entirely set by the user, but depends also on the stopping
criteria: the number of trees can be lower than max_iter if adding a new
tree does not improve the model enough. We will give more details on this in
the next exercise.
The depth of the trees is controlled by max_depth (or max_leaf_nodes). We
saw in the section on gradient-boosting that boosting algorithms fit the error
of the previous tree in the ensemble. Thus, fitting fully grown trees would be
detrimental. Indeed, the first tree of the ensemble would perfectly fit
(overfit) the data and thus no subsequent tree would be required, since there
would be no residuals. Therefore, the tree used in gradient-boosting should
have a low depth, typically between 3 to 8 levels, or few leaves (\(2^3=8\) to
\(2^8=256\)). Having very weak learners at each step helps reducing overfitting.
With this consideration in mind, the deeper the trees, the faster the
residuals are corrected and then less learners are required. Therefore,
it can be beneficial to increase max_iter if max_depth is low.
Finally, we have overlooked the impact of the learning_rate parameter
until now. This parameter controls how much each correction contributes to the
final prediction. A smaller learning-rate means the corrections of a new
tree result in small adjustments to the model prediction. When the
learning-rate is small, the model generally needs more trees to achieve good
performance. A higher learning-rate makes larger adjustments with each tree,
which requires fewer trees and trains faster, at the risk of overfitting. The
learning-rate needs to be tuned by hyperparameter tuning to obtain the best
value that results in a model with good generalization performance.
from scipy.stats import loguniform
from sklearn.ensemble import HistGradientBoostingRegressor
param_distributions = {
"max_iter": [3, 10, 30, 100, 300, 1000],
"max_leaf_nodes": [2, 5, 10, 20, 50, 100],
"learning_rate": loguniform(0.01, 1),
}
search_cv = RandomizedSearchCV(
HistGradientBoostingRegressor(),
param_distributions=param_distributions,
scoring="neg_mean_absolute_error",
n_iter=20,
random_state=0,
n_jobs=2,
)
search_cv.fit(data_train, target_train)
columns = [f"param_{name}" for name in param_distributions.keys()]
columns += ["mean_test_error", "std_test_error"]
cv_results = pd.DataFrame(search_cv.cv_results_)
cv_results["mean_test_error"] = -cv_results["mean_test_score"]
cv_results["std_test_error"] = cv_results["std_test_score"]
cv_results[columns].sort_values(by="mean_test_error")
| param_max_iter | param_max_leaf_nodes | param_learning_rate | mean_test_error | std_test_error | |
|---|---|---|---|---|---|
| 14 | 300 | 100 | 0.018640 | 31.075632 | 0.297297 |
| 6 | 300 | 20 | 0.047293 | 31.829026 | 0.257212 |
| 2 | 30 | 50 | 0.176656 | 32.645976 | 0.277494 |
| 13 | 300 | 10 | 0.297739 | 33.006395 | 0.472993 |
| 19 | 100 | 10 | 0.215543 | 33.137945 | 0.274292 |
| 9 | 100 | 20 | 0.083745 | 33.197275 | 0.234218 |
| 12 | 100 | 20 | 0.067503 | 33.871296 | 0.418139 |
| 16 | 300 | 5 | 0.059290 | 35.810194 | 0.469997 |
| 1 | 100 | 5 | 0.160519 | 36.248312 | 0.364016 |
| 0 | 1000 | 2 | 0.125207 | 40.889103 | 0.724700 |
| 7 | 1000 | 2 | 0.054511 | 42.194444 | 0.632877 |
| 8 | 3 | 5 | 0.906226 | 50.229581 | 0.764358 |
| 18 | 10 | 5 | 0.248463 | 50.256974 | 0.559465 |
| 5 | 10 | 100 | 0.061034 | 61.595093 | 0.676801 |
| 17 | 3 | 5 | 0.079415 | 81.473618 | 0.964294 |
| 4 | 10 | 2 | 0.035100 | 82.538669 | 0.931620 |
| 15 | 3 | 50 | 0.019923 | 87.608300 | 1.093378 |
| 3 | 3 | 2 | 0.039361 | 87.824971 | 1.081327 |
| 11 | 3 | 10 | 0.019351 | 88.317090 | 1.069122 |
| 10 | 3 | 5 | 0.017240 | 88.933888 | 1.074173 |
Caution
Here, we tune max_iter but be aware that it is better to set max_iter to a
fixed, large enough value and use parameters linked to early_stopping as we
will do in Exercise M6.04.
In this search, we observe that for the best ranked models, having a
smaller learning_rate, requires more trees or a larger number of leaves
for each tree. However, it is particularly difficult to draw more detailed
conclusions since the best value of each hyperparameter depends on the other
hyperparameter values.
We can now estimate the generalization performance of the best model using the test set.
error = -search_cv.score(data_test, target_test)
print(f"On average, our HGBT regressor makes an error of {error:.2f} k$")
On average, our HGBT regressor makes an error of 30.57 k$
The mean test score in the held-out test set is slightly better than the score of the best model. The reason is that the final model is refitted on the whole training set and therefore, on more data than the cross-validated models of the grid search procedure.
We summarize these details in the following table:
Bagging & Random Forests |
Boosting |
|---|---|
fit trees independently |
fit trees sequentially |
each deep tree overfits |
each shallow tree underfits |
averaging the tree predictions reduces overfitting |
sequentially adding trees reduces underfitting |
generalization improves with the number of trees |
too many trees may cause overfitting |
does not have a |
fitting the residuals is controlled by the |