Input space versus feature space in kernel-based methods
By Bernhard Schölkopf, Sebastian Mika, C. Burges, Phil Knirsch, K. Müller, Gunnar Rätsch, Alex Smola. 1999
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- Yoshua Bengio
- Sergey Levine
- Dacheng Tao
- Trevor Darrell
- Yi Yang
- Kristen Grauman
- Luc Van Gool
- Andrew Zisserman
Citations in the corpus
- Model learning for robot control: a survey (2011, background): “Feature selection has an inherent connection to sparse and reduced set methods, where the purpose is to filter the information tht is crucial for the model approximation (Csato and Opper 2002; Schölkopf et al. 1999).”
- Facial Expression Recognition in Image Sequences Using Geometric Deformation Features and Support Vector Machines (2007, background): “where is an arbitrary dimensional Hilbert space [43] and .”
- Gabor wavelets and General Discriminant Analysis for face identification and verification (2007, methodology): “Kernel based methods [9], exemplified by Kernel Principal Component Analysis (KPCA) [10,11], Kernel Fisher Discriminant Analysis (KFDA) [12,13] and General Discriminant Analysis (GDA) [14] have significantly outperformed PCA, LDA, ICA and neural networks in similar recognition tasks.”
- Sparse On-Line Gaussian Processes (2002, methodology): “The gap between our on-line result and the batch performance reported in Schölkopf et al. (1999) is also smaller; this might be due to the Bayesian nature of the GPs that avoids the overfitting.”
- Nonlinear gradient denoising: Finding accurate extrema from inaccurate functional derivatives (2015, background): “[25] B.”
- A Unifying Review of Deep and Shallow Anomaly Detection (2020, background): “, [210], [292], [301]) further increases the flexibility of one-class models and enables decision boundaries for more complex distributions.”
- Survey on data-driven industrial process monitoring and diagnosis (2012, methodology): “Hence, it is desirable to develop methods that reduce the dimension of the eigen-decomposition problem (Schölkopf et al., 1999; Schölkopf et al., 1998) in the kernel methods.”