 |
 |
 |
 |
 |
|
|
|
|
|
| HAL: hal-00639117, version 1 |
| Detailed view |  Export this paper |
 |
| Pattern Recognition Letters 32, 13 (2011) 1511-1515 |
|
|
|
|
| Learning General Gaussian Kernel Hyperparameters using Optimization on Symmetric Positive-Definite Matrices Manifold |
|
|
| Fahed Abdallah 1Hicham Laanaya |
|
|
| (2011-05-09) |
|
|
|
| We propose a new method for general Gaussian kernel hyperparameter optimization for support vector machines classification. The hyperparameters are constrained to lie on a differentiable manifold. The proposed optimization technique is based on a gradient-like descent algorithm adapted to the geometrical structure of the manifold of symmetric positive-definite matrices. We compare the performance of our approach with the classical support vector machine for classification and with other methods of the state of the art on toy data and on real world data sets. |
|
|
|
|
|
|
|
|
|
|
|
| 1: | Heuristique et Diagnostic des Systèmes Complexes (HEUDIASYC) |
|
|
|
CNRS : UMR6599 – Université de Technologie de Compiègne |
|
|
| 2: | Laboratoire des signaux et systèmes (L2S) |
|
|
|
UMR8506 CNRS – SUPELEC – Université Paris XI - Paris Sud |
|
|
| 3: | University of technology de Troyes (ISTIT/M2S) |
|
|
|
University of technology of Troyes |
|
|
| 4: | Laboratoire Hippolyte Fizeau (FIZEAU) |
|
|
|
Université Nice Sophia Antipolis [UNS] – CNRS : UMR6525 – Observatoire de la Côte d'Azur – INSU |
|
|
|
|
|
|
|
|
|
| Subject | : | Statistics/Machine Learning |
| hal-00639117, version 1 | |
| http://hal.archives-ouvertes.fr/hal-00639117 | |
| oai:hal.archives-ouvertes.fr:hal-00639117 | |
| From: Fahed Abdallah | |
| Submitted on: Tuesday, 8 November 2011 11:57:18 | |
| Updated on: Tuesday, 8 November 2011 11:57:18 | |
