Advanced nonparametric tests for multiple comparisons in the design of experiments in computational intelligence and data mining: Experimental analysis of power

By Salvador Garcia, Alberto Fernández, Julián Luengo, Francisco Herrera. 2010

2,151 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

On the characterization of noise filters for self-training semi-supervised in nearest neighbor classification (2014, methodology): “In order to test the behavior of noise filters, the experimental study will include a statistical analysis based on nonparametric statistical tests [38].”
EUSBoost: Enhancing ensembles for highly imbalanced data-sets by evolutionary undersampling (2013, methodology): “The results obtained in the comparisons are contrasted by the proper non-parametric statistical tests, as suggested in the literature [13,21,22].”
Benchmarking state-of-the-art classification algorithms for credit scoring: An update of research (2015, methodology): “Similarly, pairwise comparisons without p -value adjustment increase the actual probability of Type-I errors beyond the stated level of 𝛼 (e.g., García, et al., 2010).”
Self-labeled techniques for semi-supervised learning: taxonomy, software and empirical study (2015, methodology): “The experimental study will include a statistical analysis based on nonparametric statistical tests [27,28].”
Chaotic Local Search-Based Differential Evolution Algorithms for Optimization (2021, methodology): “The adjusted p-value is a relatively conservative value used to eliminate the family wise error rate [62].”
SMOTE-IPF: Addressing the noisy and borderline examples problem in imbalanced classification by a re-sampling method with filtering (2015, methodology): “After preprocessing these datasets, the performances of the classifiers built with C4.5 [41] will be evaluated and they will also be contrasted using the proper statistical tests as recommended in the specialized literature [14] [19] [25].”
Progressive Meta-Learning With Curriculum (2022, background): “reported in Table V (please refer to [54] for more details).”
Financial credit risk assessment: a recent review (2015, background): “The parametric tests require the independency, normality, and homoscedasticity of investigated data, which are difficultly verified in machine learning studies (Garcia et al. 2010).”
Multi-head CNN-RNN for multi-time series anomaly detection: An industrial case study (2019, methodology): “These tests are widely used in the field of machine learning as can be shown in [82, 84, 85, 86], where they recommend their use.”
Overview of Friedman’s Test and Post-hoc Analysis (2015, methodology): “Demsar (2006), Garcı́a et al. (2010) and Derrac (2011) explained Friedman’s test (Friedman, 1937), the Iman Davenport correction (Iman and Davenport, 1980) and some post-hoc procedures with adjusted p-values, such as Bonferroni−Dunn (Dunn, 1961), Holm (Holm, 1979), Hochberg (Hochberg, 1988), and…”
Ensemble learning for data stream analysis: A survey (2017, background): “Statistical tests have gained a signiﬁcant popularity in the machine learning community [66] .”
Enhancing particle swarm optimization using generalized opposition-based learning (2011, methodology): “To compare the performance of multiple algorithms on this test suite, the average ranking of the Friedman test is conducted by the suggestions considered in [12,13].”