Abstract:Nowadays most of the current kernel learning approaches are showing good results in small datasets and fail to scale to large ones. As such, it is necessary to develop faster kernel optimization algorithms that perform better with larger datasets, especially, for the “Big Data” applications. This paper presents a novel fast method to optimize the Gaussian kernel function for two-class pattern classification tasks, where it is desirable for the kernel machines to use an optimized kernel that adapts well to the input data and the learning tasks. We propose to optimize the Gaussian kernel function by using the formulated kernel target alignment criterion. By adopting the Euler–Maclaurin formula and the local and global extremal properties of the approximate kernel separability criterion, the approximate criterion function can be proved to have a determined global minimum point. Thus, when the approximate criterion function is a sufficient approximation of the criterion function, through using a Newton-based algorithm, the proposed optimization is simply solved without being repeated the searching procedure with different starting points to locate the best local minimum. The proposed method is evaluated on thirteen data sets with three Gaussian-kernel-based learning algorithms. The experimental results show that the criterion function has the determined global minimum point for the all thirteen data sets, the proposed method achieves the best high time efficiency performance and the best overall classification performance.

Optimizing the Kernel in the Empirical Feature Space

Learning with the Optimized Data-Dependent Kernel

A Kernel Function Optimization And Selection Algorithm Based On Cost Function Maximization

Similarity Learning Via Optimizing the Data-Dependent Kernel

Feature Selection with Kernel Class Separability

Kernel optimization-based discriminant analysis for face recognition

Feature Selection in a Kernel Space

A Unified Framework for Kernelization: the Empirical Kernel Feature Space

Optimizing Kernel Function with Applications to Kernel Principal Analysis and Locality Preserving Projection for Feature Extraction.

Optimal Gaussian Kernel Parameter Selection for SVM Classifier

Adaptive kernel subspace method for speeding up feature extraction

Kernel-Based Enhanced Maximum Margin Criterion Algorithm for High-Dimensional Feature Extraction

Optimizing Kernel Pca Using Sparse Representation-Based Classifier for Mstar Sar Image Target Recognition

Kernel Optimization Using Nonparametric Fisher Criterion in the Subspace

Optimizing the Gaussian Kernel Function with the Formulated Kernel Target Alignment Criterion for Two-Class Pattern Classification.

Optimizing the Data-Dependent Kernel under a Unified Kernel Optimization Framework

Gaussian kernel optimization: Complex problem and a simple solution

An Efficient Kernel Evaluation Criterion for Multiclass Classification

Kernel Optimisation for KPCA Based on Gaussianity Estimation.

TAKES: a fast method to select features in the kernel space.

Kernel based enhanced maximum margin criterion for feature extraction