This feature is in the alpha tier. For more information on feature tiers, see Operations reference.
Random forest is a popular supervised machine learning method for classification and regression that consists of using several decision trees, and combining the trees' predictions into an overall prediction. To train the random forest is to train each of its decision trees independently. Each decision tree is typically trained on a slightly different part of the training set, and may look at different features for its node splits.
The idea is that the difference in how each decision tree is trained will help avoid overfitting which is not uncommon when just training a single decision tree on the entire training set. The approach of combining several predictors (in this case decision trees) is also known as ensemble learning, and using different parts of the training set for each predictor is often referred to as bootstrap aggregating or bagging.
For classification, a random forest prediction is made by simply taking a majority vote of its decision trees' predictions. The impurity criteria available for computing the potential of a node split in decision tree classifier training in GDS are Gini impurity (default) and Entropy.
For regression, a random forest prediction is made by simply taking the average of its decision trees' predictions. The impurity criterion used for computing the potential of a node split in decision tree regressor training in GDS is Mean squared error.
Random forest regression is available for the training of node regression pipelines.
In order to balance matters such as bias vs variance of the model, and speed vs memory consumption of the training, GDS exposes several hyperparameters that one can tune. Each of these are described below.
This parameter sets the number of decision trees that will be part of the random forest.
Having a too small number of trees could mean that the model will overfit to some parts of the dataset.
A larger number of trees will in general mean that the training takes longer, and the memory consumption will be higher.
For each node split in a decision tree, a set of features of the feature vectors are considered.
The number of such features considered is the
maxFeaturesRatio multiplied by the total number of features.
If the number of features to be considered are fewer than the total number of features, a subset of all features are sampled (without replacement).
This is sometimes referred to as feature bagging.
A high (close to 1.0) max features ratio means that the training will take longer as there are more options for how to split nodes in the decision trees. It will also mean that each decision tree will be better at predictions over the training set. While this is positive in some sense, it might also mean that each decision tree will overfit on the training set.
This parameter sets the maximum depth of the decision trees in the random forest.
A high maximum depth means that the training might take longer, as more node splits might need to be considered. The memory footprint of the produced prediction model might also be higher since the trees simply may be larger (deeper).
A deeper decision tree may be able to better fit to the training set, but that may also mean that it overfits.
This parameter sets the minimum number of training samples required to be present in a leaf node of a decision tree.
A large leaf size means less specialization on the training set, and thus possibly worse performance on the training set, but possibly avoiding overfitting. It will likely also mean that the training and prediction will be faster, since probably the trees will contain fewer nodes.
This parameter sets the minimum number of training samples required to be present in a node of a decision tree in order for it to be split during training. To split a node means to continue the tree construction process to add further children below the node.
A large split size means less specialization on the training set, and thus possibly worse performance on the training set, but possibly avoiding overfitting. It will likely also mean that the training and prediction will be faster, since probably fewer node splits will be considered, and thus the trees will contain fewer nodes.
Each decision tree in the random forest is trained using a subset of the training set.
This subset is sampled with replacement, meaning that a feature vector of the training may be sampled several times for a single decision tree.
The number of training samples for each decision tree is the
numberOfSamplesRatio multiplied by the total number of samples in the training set.
A high ratio will likely imply better generalization for each decision tree, but not necessarily so for the random forest overall. Training will also take longer as more feature vectors will need to be considered in each node split of each decision tree.
The special value of 0.0 is used to indicate no sampling. In this case all feature vectors of the training set will be used for training by every decision tree in the random forest.
When deciding how to split a node in a decision tree, potential splits are evaluated using an impurity criterion.
The lower the combined impurity of the two potential child nodes, the better the split is deemed to be.
For random forest classification in GDS, there are two options, specified via the
criterion configuration parameter, for such impurity criteria:
A measure of how often a randomly chosen element from the set would be incorrectly labeled if it was randomly labeled according to the distribution of labels in the set
Selected to use via the string
An information theoretic measure of the amount of uncertainty in a set
Selected to use via the string
It’s hard to say apriori which criterion is best for a particular problem, but in general using Gini impurity will imply faster training since using Entropy requires computing logarithms.
Was this page helpful?