Artificial Intelligence Review - Special issue on lazy learning
Global and local neural network ensembles
Pattern Recognition Letters
The lack of a priori distinctions between learning algorithms
Neural Computation
Vapnik-chervonenkis generalization bounds for real valued neural networks
Neural Computation
Adaptive voting rules for k-nearest neighbors classifiers
Neural Computation
A Scalable Noise Reduction Technique for Large Case-Based Systems
ICCBR '09 Proceedings of the 8th International Conference on Case-Based Reasoning: Case-Based Reasoning Research and Development
Fast Local Support Vector Machines for Large Datasets
MLDM '09 Proceedings of the 6th International Conference on Machine Learning and Data Mining in Pattern Recognition
Fast and Scalable Local Kernel Machines
The Journal of Machine Learning Research
Operators for transforming kernels into quasi-local kernels that improve SVM accuracy
Journal of Intelligent Information Systems
Adaptive weighted fusion of local kernel classifiers for effective pattern classification
ICIC'11 Proceedings of the 7th international conference on Advanced Intelligent Computing
Local discriminative distance metrics ensemble learning
Pattern Recognition
Parallel and local learning for fast probabilistic neural networks in scalable data mining
Proceedings of the 6th Balkan Conference in Informatics
Universal consistency of localized versions of regularized kernel methods
The Journal of Machine Learning Research
| Hi-index | 0.06 |
In previous publications (Bottou and Vapnik 1992; Vapnik 1992)we described local learning algorithms, which result in performanceimprovements for real problems. We present here the theoreticalframework on which these algorithms are based. First, we present anew statement of certain learning problems, namely the local riskminimization. We review the basic results of the uniformconvergence theory of learning, and extend these results to localrisk minimization. We also extend the structural risk minimizationprinciple for both pattern recognition problems and regressionproblems. This extended induction principle is the basis for a newclass of algorithms.