Sciweavers

2712 search results - page 108 / 543
» On learning algorithm selection for classification
Sort
View
ICIP
2002
IEEE
14 years 1 months ago
An iterative algorithm for context selection in adaptive entropy coders
Context-based adaptive entropy coding is an essential feature of modern image compression algorithms; however, the design of these coders is non-trivial due to the balance that mu...
Tong Jin, Jacques Vaisey
ICML
2009
IEEE
14 years 9 months ago
Optimal reverse prediction: a unified perspective on supervised, unsupervised and semi-supervised learning
Training principles for unsupervised learning are often derived from motivations that appear to be independent of supervised learning. In this paper we present a simple unificatio...
Linli Xu, Martha White, Dale Schuurmans
ICML
2006
IEEE
14 years 9 months ago
Learning a kernel function for classification with small training samples
When given a small sample, we show that classification with SVM can be considerably enhanced by using a kernel function learned from the training data prior to discrimination. Thi...
Tomer Hertz, Aharon Bar-Hillel, Daphna Weinshall
TIP
2008
79views more  TIP 2008»
13 years 8 months ago
An Overview and Performance Evaluation of Classification-Based Least Squares Trained Filters
An overview of the classification-based least squares trained filters on picture quality improvement algorithms is presented. For each algorithm, the training process is unique and...
Ling Shao, Hui Zhang, Gerard de Haan
ISNN
2007
Springer
14 years 3 months ago
Extensions of Manifold Learning Algorithms in Kernel Feature Space
Manifold learning algorithms have been proven to be capable of discovering some nonlinear structures. However, it is hard for them to extend to test set directly. In this paper, a ...
Yaoliang Yu, Peng Guan, Liming Zhang