A comparison of stacking with meta decision trees to bagging, boosting, and stacking with other methods

Ženko, Bernard; Todorovski, Ljupčo; Džeroski, Sašo

doi:10.1109/icdm.2001.989601

Cited by 37 publications

(30 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While conducting this study, the study of Džeroski andŽenko [6], and a few other recent studies [16,13], we have encountered quite a few contradictions between claims in the recent literature on stacking and our experimental results. For example, Merz [8] claims that SCANN is clearly better than the oracle selecting the best classifier (which should perform even better than SelectBest).…”

Section: Conclusion and Further Workmentioning

confidence: 56%

See 1 more Smart Citation

Stacking with an Extended Set of Meta-level Attributes and MLR

Ženko

Džeroski

2002

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. We propose a new set of meta-level features to be used for learning how to combine classifier predictions with stacking. This set includes the probability distributions predicted by the base-level classifiers and a combination of these with the certainty of the predictions. We use these features in conjunction with multi-response linear regression (MLR) at the meta-level. We empirically evaluate the proposed approach in comparison to several state-of-the-art methods for constructing ensembles of heterogeneous classifiers with stacking. Our approach performs better than existing stacking approaches and also better than selecting the best classifier from the ensemble by cross validation (unlike existing stacking approaches, which at best perform comparably to it).

show abstract

Section: Conclusion and Further Workmentioning

confidence: 56%

“…On the other hand, MDTs perform only slightly better than SCANN and selecting the best classifier with cross validation (SelectBest).Ženko et al [16] report that MDTs perform slightly worse as compared to stacking with MLR. Overall, SCANN, MDTs, stacking with MLR and SelectBest seem to perform at about the same level.…”

Section: Recent Advancesmentioning

confidence: 99%

Stacking with an Extended Set of Meta-level Attributes and MLR

Ženko

Džeroski

2002

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Therefore, it is only natural to consider that a method efficiency can be affected by lucky, or unlucky, sample choices. A large number of studies compares empirically Bagging and Boosting with some other classification methods, e.g., [18,15,19,6,22,11,14,10,13,23]. Despite of such an abundant literature about comparison, very few research papers explicit take random choices of samples into account.…”

Section: Introductionmentioning

confidence: 99%

The impact of random samples in ensemble classifiers

Fernandes

Lopes

Ruiz

2010

Proceedings of the 2010 ACM Symposium on Applied Computing

View full text Add to dashboard Cite

The use of ensemble classifiers, e.g., Bagging and Boosting, is wide spread to machine learning. However, most of studies in this area are based on empirical comparisons that suffer from a lack of care to the randomness of these methods. This paper describes the dangers of experiments with ensemble classifiers by analyzing the efficiency of Bagging and Boosting methods over 32 different data sets. The experiments show that variations due to randomness are often more relevant than the advantages among methods encountered in the literature. This paper main contribution is the claim, supported by statistical analysis, that no empirical comparison of ensemble classifiers can be scientifically done without paying attention to the random choices taken.

show abstract

“…Other types of committees-stacking and gradingwere presented in [24,22,3,25] and [21]. In the place of voting or weighting schemes these committees combine models via stacking.…”

Section: Introductionmentioning

confidence: 99%

“…Such knowledge is used to make the final decision of the committee. To approximate model certainty a linear regression [24,22] or meta-decision trees [25] may be used. Other combining schemes were proposed in [14,16].…”

Section: Introductionmentioning

confidence: 99%