Base learner

As explained in Machine Learning Algorithms in Depth

A base learner is an individual predictive model that serves as a component of an ensemble.

A base learner is an individual predictive model used as a component of an ensemble. Ensemble methods combine several learners to reduce variance, reduce bias, or improve predictions; accuracy and diversity determine whether the ensemble outperforms its members. section 7.5

Base learners participate in ensembles through distinct training and aggregation patterns:

• Bagging trains learners independently on bootstrap samples, then aggregates classification by voting or regression by averaging. section 7.5.1
• Boosting trains weak learners sequentially on weighted data, emphasizing examples misclassified in earlier rounds; weighted voting or summation produces the prediction. section 7.5.2
• Stacking trains base-level models on the complete training set and supplies their outputs as features to a meta-classifier or meta-regressor. section 7.5.3
• Learners may be homogeneous, such as multiple models of one type, or heterogeneous, such as KNN, random forest, and naive Bayes. section 7.5 section 7.5.3

Decision stumps in AdaBoost

An AdaBoost ensemble uses decision trees of depth 1 as base learners. Each stump partitions the feature space with a threshold; later rounds increase weights for misclassified observations, while more accurate learners receive greater ensemble weights. section 7.5.2

Figure 7.15 AdaBoost algorithm pseudo-code

Follow the AdaBoost pseudo-code to inspect learner weighting, data-example weights, and sequential ensemble construction. section 7.5.2

How Machine Learning Algorithms in Depth covers it

How deep it goes

Explains learner diversity, bootstrap and weighted training, voting, averaging, and meta-model inputs, with implementation examples.

Examples

Uses decision stumps in AdaBoost and KNN, random forest, naive Bayes, and logistic regression in ensemble demonstrations.

What it assumes

Assumes familiarity with predictive models and classification or regression while defining how models become ensemble components.

What you can do after

Evaluate how individual models contribute to ensemble accuracy, diversity, weighting, voting, averaging, or meta-model prediction.

How it connects

Connects base learners to ensemble methods, bagging, boosting, stacking, decision trees, random forests, voting, averaging, and meta-classifiers.

liveBookmanning.com

Read Machine Learning Algorithms in Depth on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Ensemble Methods for Machine LearningA base estimator is an individual predictive model whose output contributes to an ensemble or higher-level estimator.

Compare how 2 books explain base learner →

Related concepts

• ensembleused by · An ensemble is a technique that combines or considers outputs from multiple models or component methods collectively to produce or assess a joint result.
• decision treeincludes · A decision tree is a branching structure that represents sequential tests or choices, with paths leading to outcomes, predictions, or actions.
• base-learning algorithmproduced by · A base-learning algorithm is an algorithm used to train an individual base estimator within an ensemble.
• heterogeneous ensemblepart of · A heterogeneous ensemble is an ensemble whose component learners use different model types or learning algorithms.
• parallel ensemble methodpart of · A parallel ensemble method combines outputs from multiple independently trained base learners that perform the same task.

This explainer was generated by Manning's knowledge graph from the text of Machine Learning Algorithms in Depth. The section links point to the parts of the book it draws on.