Estimating the Predictive Accuracy of a Classifier

By Hilan Bensusan, Alexandros Kalousis. 2001

126 citations recorded by the source corpus.

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

Citations in the corpus

Hyper-Parameter Optimization: A Review of Algorithms and Applications (2020, background): “n early works, linear regression was the most common surrogate model (Bensusan and Kalousis, 2001).”
Meta-Learning: A Survey (2018, neutral): “Disciminant Landmarker(NB) P(θNB, tj ) Feature independence [115] More models [14, 88] Relative LM Pa,j − Pb,j Probing performance [53] Subsample LM P(θi , tj , st ) Probing performance [160]”
Ranking Learning Algorithms: Using IBL and Meta-Learning on Accuracy and Time Results (2003, methodology): “Contrary to previous work using similar measures (Gama & Brazdil, 1995; Lindner & Studer, 1999; Bensusan & Kalousis, 2001; Kalousis & Theoharis, 1999; Sohn, 1999), we selected a set of measures a priori with the aim of representing some of the properties that affect the performance of algorithms.”
Meta-QSAR: a large-scale application of meta-learning to drug design and discovery (2017, unknown): “Other approaches directly estimate the performances of all algorithms (Bensusan and Kalousis, 2001), or use predictive clustering trees (Todorovski et al., 2002).”
On the predictive power of meta-features in OpenML (2017, background): “Yet, none of the studies, apart from that by Bensusan and Kalousis (2001), properly acknowledge this fact.”
Automatic classifier selection for non-experts (2014, methodology): “Bensusan and Kalousis [7] • • • • a(1)”
Predicting the Performance of Learning Algorithms Using Support Vector Machines as Meta-regressors (2008, methodology): “Although the strict Meta-Learning approach (as described above) has been applied by different authors (such as [2, 4, 9, 14, 15]), certain information loss may be introduced in the definition of the class labels associated to meta-examples.”
Combination of Task Description Strategies and Case Base Properties for Meta-Learning (2002, background): “lowest misclassification rate or by a regression approach where the error rate of each learner from the pool is to be predicted [3] [11].”
Meta-learning for evolutionary parameter optimization of classifiers (2012, background): “2003), or predicting the performance of algorithms individually (Gama and Brazdil 1995; Sohn 1999; Köpf et al. 2000; Bensusan and Kalousis 2001).”