Filters








56,067 Hits in 6.2 sec

Active learning of label ranking functions

Klaus Brinker
2004 Twenty-first international conference on Machine learning - ICML '04  
Considering both the pairwise decomposition and the constraint classification technique to represent label ranking functions, we introduce a novel generalization of pool-based active learning to address  ...  While this already constitutes a major issue in classification learning, it becomes an even more serious problem when dealing with the more complex target domain of total orders over a set of alternatives  ...  Acknowledgments We would like to thank Eyke Hüllermeier and Chih-Len Lin for helpful suggestions and comments.  ... 
doi:10.1145/1015330.1015331 dblp:conf/icml/Brinker04 fatcat:fe6astoaxvcjpinbabbb52zw3q

MiSC: Mixed Strategies Crowdsourcing [article]

Ching-Yun Ko, Rui Lin, Shu Li, Ngai Wong
2019 arXiv   pre-print
Popular crowdsourcing techniques mostly focus on evaluating workers' labeling quality before adjusting their weights during label aggregation.  ...  In this work, we propose MiSC (Mixed Strategies Crowdsourcing), a versatile framework integrating arbitrary conventional crowdsourcing and tensor completion techniques.  ...  We exemplify through a case study and show that for a perfectly labeled matrix, its corresponding label tensor will have an intrinsically low-rank Tucker decomposition.  ... 
arXiv:1905.07394v1 fatcat:2dgl3dssnrdzhaxyo2ze3jmt3y

Binary Decomposition Methods for Multipartite Ranking [chapter]

Johannes Fürnkranz, Eyke Hüllermeier, Stijn Vanderlooy
2009 Lecture Notes in Computer Science  
Moreover, to learn multipartite ranking functions, we propose methods on the basis of binary decomposition techniques that have previously been used for multi-class and ordinal classification.  ...  Bipartite ranking refers to the problem of learning a ranking function from a training set of positively and negatively labeled examples.  ...  This work was supported by the German Science Foundation (DFG) and the Dutch Organisation for Scientific Research (NWO).  ... 
doi:10.1007/978-3-642-04180-8_41 fatcat:qyngpojjnzcihmtqnshi2xeedm

A Tutorial on Multi-label Classification Techniques [chapter]

André C. P. L. F. de Carvalho, Alex A. Freitas
2009 Studies in Computational Intelligence  
This tutorial presents the most frequently used techniques to deal with these problems in a pedagogical manner, with examples illustrating the main techniques and proposing a taxonomy of multi-label techniques  ...  This group of problems represents an area known as multi-label classification.  ...  Two alternatives can be followed for the label selection.  ... 
doi:10.1007/978-3-642-01536-6_8 fatcat:dmoem7uzfzhfbkjjol3436tslq

A Composite Likelihood View for Multi-Label Classification

Yi Zhang, Jeff G. Schneider
2012 Journal of machine learning research  
In this paper we show the connection between composite likelihoods [17] and many multilabel decomposition methods, e.g., one-vs-all, one-vs-one, calibrated label ranking, probabilistic classifier chain  ...  Pairwise label comparisons, which seem to be a natural choice for subproblems, are replaced by bivariate label densities, which are more informative and natural components in a composite likelihood.  ...  Calibrated Label Ranking (CLR) is a strategy that combines both one-vs-one and one-vs-all decompositions for multi-label classification [9] .  ... 
dblp:journals/jmlr/ZhangS12 fatcat:73qrvww4kjailkftkgtjnabejm

Multilabel classification via calibrated label ranking

Johannes Fürnkranz, Eyke Hüllermeier, Eneldo Loza Mencía, Klaus Brinker
2008 Machine Learning  
to the pairwise decomposition technique.  ...  Label ranking studies the problem of learning a mapping from instances to rankings over a predefined set of labels.  ...  We would like to thank the anonymous reviewers and the editor for their helpful suggestions. We also thank Janez Demšar for an interesting discussion on significance tests.  ... 
doi:10.1007/s10994-008-5064-8 fatcat:ehzskcl6gjfhphk3kd2mkm57cm

Probability Estimation for Multi-class Classification Based on Label Ranking [chapter]

Weiwei Cheng, Eyke Hüllermeier
2012 Lecture Notes in Computer Science  
Exploiting the close connection between probability estimation and ranking, our idea is to solve the former on the basis of the latter, taking advantage of recently developed methods for label ranking.  ...  The latter provides the basis of pairwise coupling techniques, which arguably constitute the state-of-the-art in multi-class probability estimation.  ...  The most common alternative to the pairwise (all pairs) decomposition scheme is one-vs-rest (OVR) decomposition [19] : One model is trained for each class label y i , using this label as positive and  ... 
doi:10.1007/978-3-642-33486-3_6 fatcat:f2fbtwsaozazpco5vi6ktpmyom

Case-Based Label Ranking [chapter]

Klaus Brinker, Eyke Hüllermeier
2006 Lecture Notes in Computer Science  
We approach this setting from a case-based perspective and propose a sophisticated k-NN framework as an alternative to previous binary decomposition techniques.  ...  Label ranking studies the problem of learning a mapping from instances to rankings over a predefined set of labels.  ...  An alternative pairwise approach to label ranking learning has been introduced in [6] .  ... 
doi:10.1007/11871842_53 fatcat:zcms462hgrezva6wsuao672bgu

Multi-label LeGo — Enhancing Multi-label Classifiers with Local Patterns [chapter]

Wouter Duivesteijn, Eneldo Loza Mencía, Johannes Fürnkranz, Arno Knobbe
2012 Lecture Notes in Computer Science  
We experimentally show that using such constructed features can improve the classification performance of decompositive multi-label learning techniques.  ...  These features are then used as input for multi-label classifiers.  ...  An alternative approach is calibrated label ranking (CLR) [11] , where the key idea is to learn one classifier for each binary comparison of labels.  ... 
doi:10.1007/978-3-642-34156-4_12 fatcat:vz7kmbbp25cn5jqjahdlnje2a4

A General Framework for Fast Co-clustering on Large Datasets Using Matrix Decomposition

Feng Pan, Xiang Zhang, Wei Wang
2008 2008 IEEE 24th International Conference on Data Engineering  
In this paper, we propose a general framework, CRD, for co-clustering large datasets utilizing recently developed samplingbased matrix decomposition methods.  ...  For a data matrix of m rows and n columns, the time complexity of these methods is usually in the order of m × n (if not higher).  ...  n number of row/column clusters,k,l (0) Low rank row/column decomposition (1) Randomly initialize the clusters for the selected rows/columns in the low rank row/column decomposition matrices.  ... 
doi:10.1109/icde.2008.4497548 pmid:20419039 pmcid:PMC2858408 dblp:conf/icde/PanZW08 fatcat:ojisgenwirhbbocrxvgaqizyee

A Novel Singing Voice Separation Method Based on Sparse Non-Negative Matrix Factorization and Low-Rank Modeling

S. Mavaddati
2019 Iranian Journal of Electrical and Electronic Engineering  
The vocal signal and music accompaniment can be considered as sparse and low-rank components of a singing voice segment, respectively. 2) An alternating factorization algorithm is used to decompose input  ...  This separation procedure is done using a decomposition model based on the spectrogram of singing voice signals.  ...  The rank value in the decomposition scheme based on SNMF technique is set to 2.  ... 
doaj:7bb13e894aa5463faf1af08330d208c5 fatcat:tt7survcijexbmt6kygg3cxqlu

Cautious Label-Wise Ranking with Constraint Satisfaction [chapter]

Yonatan-Carlos Carranza-Alarcon, Soundouss Messoudi, Sébastien Destercke
2020 Communications in Computer and Information Science  
In this paper, we consider such an approach for the label ranking problem, where in addition we allow the predictive model to produce cautious inferences in the form of sets of rankings when it lacks information  ...  More specifically, we propose to combine a rank-wise decomposition, in which every sub-problem becomes an ordinal classification one, with a constraint satisfaction problem (CSP) approach to verify the  ...  This work was carried out in the framework of the Labex MS2T and PreServe projects, funded by the French Government, through the National Agency for Research (Reference ANR-11-IDEX-0004-02 and ANR-18-CE23  ... 
doi:10.1007/978-3-030-50143-3_8 fatcat:h6psbwipandlxbclxmd356d2fq

Tensor-Based Algorithms for Image Classification

Stefan Klus, Patrick Gelß
2019 Algorithms  
for image classification.  ...  We show that tensor-based methods developed for learning the governing equations of dynamical systems from data can, in the same way, be used for supervised learning problems and propose two novel approaches  ...  Acknowledgments: We would like to thank Michael Götte and Alex Goeßmann from the TU Berlin for interesting discussions related to tensor decompositions and system identification.  ... 
doi:10.3390/a12110240 fatcat:r4oqhjdflvc2fdujcievbpmrqq

Kernel Discriminant Analysis Using Triangular Kernel for Semantic Scene Classification

Muhammad Atif Tahir, Josef Kittler, Fei Yan, Krystian Mikolajczyk
2009 2009 Seventh International Workshop on Content-Based Multimedia Indexing  
The results also indicate performance gains when compared with the state-of-the art multi-label methods for semantic scene classification.  ...  The main idea is to use LDL T decomposition instead of Cholesky decomposition. The modified SR-KDA is applied to scene database involving 6 concepts.  ...  Regularization techniques [2] or generalised singular value decomposition [19] can handle singularities while greedy approximation [18] or QR decomposition [26] can speed-up eigen-decomposition  ... 
doi:10.1109/cbmi.2009.47 dblp:conf/cbmi/TahirKYM09 fatcat:2cr3h6zdv5gkvbzclfvjplpaeu

Graded Multilabel Classification by Pairwise Comparisons

Christian Brinker, Eneldo Loza Mencia, Johannes Furnkranz
2014 2014 IEEE International Conference on Data Mining  
The task in multilabel classification is to predict for a given set of labels whether each individual label should be attached to an instance or not.  ...  Graded multilabel classification generalizes this setting by allowing to specify for each label a degree of membership on an ordinal scale.  ...  We would like to thank Weiwei Cheng and Eyke Hüllermeier for fruitful discussions and making their data and algorithms available.  ... 
doi:10.1109/icdm.2014.102 dblp:conf/icdm/BrinkerMF14 fatcat:4uwbai635zhzdbnqj4h36eifxy
« Previous Showing results 1 — 15 out of 56,067 results