Boosted Neural Networks in Evolutionary Computation

Holeňa, Martin; Linke, David; Steinfeldt, Norbert

doi:10.1007/978-3-642-10684-2_15

Cited by 5 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance of models can often be further increased by combining or ensembling (Brazdil et al 2009;Kuncheva 2004;Wolpert 1992;Schapire 1990;Woods et al 1997;Holeňa et al 2009) base algorithms, particularly in cases where base algorithms produce models of insufficient plasticity or models overfitted to training data (Brown et al 2006).…”

Section: Ensembling Algorithmsmentioning

confidence: 99%

Discovering predictive ensembles for transfer learning and meta-learning

2017

View full text Add to dashboard Cite

Recent meta-learning approaches are oriented towards algorithm selection, optimization or recommendation of existing algorithms. In this article we show how data-tailored algorithms can be constructed from building blocks on small data sub-samples. Building blocks, typically weak learners, are optimized and evolved into data-tailored hierarchical ensembles. Good-performing algorithms discovered by evolutionary algorithm can be reused on data sets of comparable complexity. Furthermore, these algorithms can be scaled up to model large data sets. We demonstrate how one particular template (simple ensemble of fast sigmoidal regression models) outperforms state-of-the-art approaches on the Airline data set. Evolved hierarchical ensembles can therefore be beneficial as algorithmic building blocks in meta-learning, including meta-learning at scale.

show abstract

Section: Ensembling Algorithmsmentioning

confidence: 99%

Discovering predictive ensembles for transfer learning and meta-learning

2017

View full text Add to dashboard Cite

show abstract

“…The classification algorithms can be further combined in several possible ways [2], [3], [4], [5], [6], [7] in order to increase the generalization performance. This makes the problem of identifying best algorithm for given dataset even more complex.…”

Section: Introductionmentioning

confidence: 99%

On performance of meta-learning templates on different datasets

Kordík

Černý

2012

The 2012 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

Meta-learning templates are data-tailored algorithms that produce supervised models. When a template is evolved on a particular dataset, it is supposed to generate good models not only on this data set but also on similar data. In this paper, we will investigate one possible way of measuring the similarity of datasets and whether it can be used to estimate if meta-learning templates produce good models. We performed experiments on several well known data sets from the UCI machine learning repository and analyzed both the similarity of datasets and templates in the space of performance metafeatures (landmarking). Our results show that the most universal algorithms (in terms of average performance) for supervised learning are the complex hierarchical templates evolved by our SpecGen approach.

show abstract

On Scalability of Predictive Ensembles and Tradeoff Between Their Training Time and Accuracy

Kordík

Frýda

2017

Advances in Intelligent Systems and Computing II

View full text Add to dashboard Cite

Boosted Neural Networks in Evolutionary Computation

Cited by 5 publications

References 13 publications

Discovering predictive ensembles for transfer learning and meta-learning

Discovering predictive ensembles for transfer learning and meta-learning

On performance of meta-learning templates on different datasets

On Scalability of Predictive Ensembles and Tradeoff Between Their Training Time and Accuracy

Contact Info

Product

Resources

About