# What is Ensemble Learning in Machine Learning

Ensemble learning typically refers to methods that generate several models which are combined to make a prediction, either in classification or regression problems. This approach has been the object of a significant amount of research in recent years and good results have been reported. This section introduced the basics of the ensemble learning of classification.

Overview

Ensemble learning is a machine learning paradigm where multiple learners are trained to solve the same problem. In contrast to common machine learning approaches which try to learn one hypothesis from training data, ensemble methods try to construct a set of hypotheses and combine them to use. An ensemble contains several learners which are usually called base learners. The generalization ability of an ensemble is usually much stronger than that of base learners. Ensemble learning is appealing because it can boost weak learners which are slightly better than random guesses to strong learners which can make very accurate predictions. So, “base learners” are also referred to as “weak learners”. It is noteworthy, however, that although most theoretical analyses work on weak learners, base learners used in practice are not necessarily weak since using not-so-weak base learners often results in better performance. Ensemble learning is the process by which multiple models, such as classifiers or experts, are strategically generated and combined to solve a particular computational intelligence problem. Ensemble learning is primarily used to improve the (classification, prediction, function approximation, etc.) performance of a model, or reduce the likelihood of an unfortunate selection of a poor one.

The Ensemble Framework

A typical ensemble method for classification tasks contains the following building blocks:

• The training set— a labeled dataset used for ensemble training. The training set can be described in a variety of languages. Most frequently, the instances are described as attribute-value vectors. We use the notation A to denote the input attributes containing n attributes:  and y to represent the class variable or the target attribute.
• Base Inducer— the inducer is an induction algorithm that obtains a training set and forms a classifier representing the generalized relationship between the input and target attributes. Let me represent an inducer. We use the notation M = I(S) for representing a classifier M which was induced by inducer I on a training set S.
• Diversity Generator— this component is responsible for generating the diverse classifiers.
• Combiner— the combiner is responsible for combining the classifications of the various classifiers.

Figure 1: An Ensemble Learning Classification

It is useful to distinguish between dependent frameworks and independent frameworks for building ensembles. In a dependent framework, the output of a classifier is used in the construction of the next classifier. Thus it is possible to take advantage of knowledge generated in previous iterations to guide the learning in the next iterations. Alternatively, each classifier is built independently and their outputs are combined in some fashion.

Figure 2: Ensemble Learning Model

A Classifier ensemble is a set of learning machines whose decisions are combined to improve the performance of the pattern recognition system. Much of the efforts in classifier combination research focus on improving the accuracy of difficult problems, and managing weaknesses and strengths of each model to give the best possible decision taking into account all ensembles The use of a combination of multiple classifiers was demonstrated to be effective, under some conditions, for several pattern recognition applications.

References

[1] Zhou, Zhi-Hua, “Ensemble learning”, Encyclopedia of biometrics (2015): pp. 411-416.

[2] Rokach, Lior, “Ensemble-based classifiers”, Artificial Intelligence Review 33.1 (2010): pp. 1-39.

[3] Ren, Ye, Le Zhang, and Ponnuthurai N. Suganthan, “Ensemble classification and regression-recent developments, applications and future directions.” IEEE Computational Intelligence Magazine 11.1 (2016): pp. 41-53.

Tags: , , , , , ,