Ensemble diversity measures and their application to thinning

Banfield, Robert E.

doi:10.1016/s1566-2535(04)00038-7

Cited by 55 publications

(99 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The final subensemble is selected by stopping aggregation at a prescribed ensemble size. In (Banfield et al (2005)) the original complete ensemble is pruned (or thinned, using the term proposed by the authors) by sequential backward selection: classifiers that do not improve the classification performance are progressively removed from the ensemble. The metrics used to identify the redundant classifiers are based on ensemble accuracy and ensemble diversity.…”

Section: Introductionmentioning

confidence: 99%

Using boosting to prune bagging ensembles

Martínez-Muñoz

Suárez

2007

Pattern Recognition Letters

View full text Add to dashboard Cite

Esta es la versión de autor del artículo publicado en: This is an author produced version of a paper published in:Pattern Recognition Letters 28.1 (2007) AbstractBoosting is used to determine the order in which classifiers are aggregated in a bagging ensemble. Early stopping in the aggregation of the classifiers in the ordered bagging ensemble allows the identification of subensembles that require less memory for storage, classify faster and can improve the generalization accuracy of the original bagging ensemble. In all the classification problems investigated pruned ensembles with 20 % of the original classifiers show statistically significant improvements over bagging. In problems where boosting is superior to bagging, these improvements are not sufficient to reach the accuracy of the corresponding boosting ensembles. However, ensemble pruning preserves the performance of bagging in noisy classification tasks, where boosting often has larger generalization errors. Therefore, pruned bagging should generally be preferred to complete bagging and, if no information about the level of noise is available, it is a robust alternative to AdaBoost.

show abstract

Section: Introductionmentioning

confidence: 99%

Using boosting to prune bagging ensembles

Martínez-Muñoz

Suárez

2007

Pattern Recognition Letters

View full text Add to dashboard Cite

show abstract

“…We focus on the relative time interval segment approach (RTI) which usually outperforms other automaticsegment approaches [1] and can provide fixed number of segment-feature vectors to establish base classifiers for ensemble. On the other hand, it is well-known that diversity is closely related to ensemble systems [9] [10]. A better investigation of diversity can be expected to achieve higher performance in our ensemble scheme.…”

Section: Introductionmentioning

confidence: 94%

Decision Templates Ensemble and Diversity Analysis for Segment-Based Speech Emotion Recognition

Bi¹,

Yang²,

Yu³

et al. 2007

Proceedings on Intelligent Systems and Knowledge Engineering (ISKE2007)

View full text Add to dashboard Cite

In this paper, we propose a novel scheme for speech emotion recognition, which uses decision-templates ensemble algorithm (DT) to combine base classifiers built on segment-level feature sets. Different feature sets from segments can provide sufficient diversity among base classifiers, which is known as a necessary condition for improvement in ensemble performance. Compared with those methods of majority voting ensemble and support vector machine, our ensemble scheme can achieve the highest performance at suitable segment levels. On the other hand, we investigate which segment-level and strategy of training base classifiers can provide potential performance in speech emotion recognition, in terms of diversity analysis.

show abstract

“…Several different approaches to measuring the diversity of classifier ensemble have been proposed in the existing literature. However, most of them rely on predictions made by classifiers [3] and as a result are computationally expensive. We propose a naive, yet fast approach based on measuring evenness of the feature spread among the subspaces.…”

Section: Diversity Measurementioning

confidence: 99%