Data-Dependent Kernels for Small-Scale, High-Dimensional Data Classification

J. Wang, J.T. Kwok, H.C. Shen, L. Quan.

Abstract: For small-scale, high-dimensional data classification, one of the most efficient classifiers is nearest neighbor (NN) classifier. What mostly decides NN classification performance is the feature extracted by some methods. Wherein kernel method is efficient to extract features. However, the selection of kernel parameters is still difficult. In this paper, we propose a so-called data dependent kernel (DDK) which is defined by generalizing the Gaussian kernel. Also an efficient and practical method is presented to calculate the DDK parameters. Moreover, one DDK based on intra-difference is given to improve the recognition performance. Experiments show that the proposed method can achieve promising classification performance in face recognition and other datasets.

Proceedings of the International Joint Conference on Neural Networks (IJCNN'05), pp.102-107, Montreal, Canada, July 2005.


Back to James Kwok's home page.