✦ LIBER ✦

An efficient kernel matrix evaluation measure

✍ Scribed by Canh Hao Nguyen; Tu Bao Ho

Book ID: 103878980
Publisher: Elsevier Science
Year: 2008
Tongue: English
Weight: 287 KB
Volume: 41
Category: Article
ISSN: 0031-3203
DOI: 10.1016/j.patcog.2008.04.005

No coin nor oath required. For personal study only.

✦ Synopsis

We study the problem of evaluating the goodness of a kernel matrix for a classification task. As kernel matrix evaluation is usually used in other expensive procedures like feature and model selections, the goodness measure must be calculated efficiently. Most previous approaches are not efficient except for kernel target alignment (KTA) that can be calculated in O(n 2 ) time complexity. Although KTA is widely used, we show that it has some serious drawbacks. We propose an efficient surrogate measure to evaluate the goodness of a kernel matrix based on the data distributions of classes in the feature space. The measure not only overcomes the limitations of KTA but also possesses other properties like invariance, efficiency and an error bound guarantee. Comparative experiments show that the measure is a good indication of the goodness of a kernel matrix.

📜 SIMILAR VOLUMES

An efficient Kernel-based matrixized lea

An efficient Kernel-based matrixized least squares support vector machine

✍ Zhe Wang, Xisheng He, Daqi Gao, Xiangyang Xue 📂 Article 📅 2011 🏛 Springer-Verlag 🌐 English ⚖ 341 KB

Matrix Kernels for Measures on Partition

Matrix Kernels for Measures on Partitions

✍ Eugene Strahov 📂 Article 📅 2008 🏛 Springer 🌐 English ⚖ 507 KB

An extension of Mercer theorem to matrix

An extension of Mercer theorem to matrix-valued measurable kernels

✍ De Vito, Ernesto; Umanità, Veronica; Villa, Silvia 📂 Article 📅 2013 🏛 Elsevier Science 🌐 English ⚖ 255 KB

Evaluation measures for kernel optimizat

Evaluation measures for kernel optimization

✍ Paweł Chudzian 📂 Article 📅 2012 🏛 Elsevier Science 🌐 English ⚖ 375 KB

An efficient kernel discriminant analysi

An efficient kernel discriminant analysis method

✍ Juwei Lu; K.N. Plataniotis; A.N. Venetsanopoulos; Jie Wang 📂 Article 📅 2005 🏛 Elsevier Science 🌐 English ⚖ 167 KB

Small sample size and high computational complexity are two major problems encountered when traditional kernel discriminant analysis methods are applied to high-dimensional pattern classification tasks such as face recognition. In this paper, we introduce a new kernel discriminant learning method, w

Efficient full-kernel evaluation for thi

Efficient full-kernel evaluation for thin wire analysis

✍ A. Heldring; J. M. Rius 📂 Article 📅 2005 🏛 John Wiley and Sons 🌐 English ⚖ 88 KB

An accurate and fast approximation of the full cylindrical kernel in the thin-wire electric-field integral equation is proposed. The presented formulation is accurate for any ratio of the method of moments wire segment length ⌬ to the wire radius a, and anywhere in space, including on the wire surfa