The Canonical Distortion Measure in Feature Space and 1-NN Classification
暂无分享,去创建一个
We prove that the Canonical Distortion Measure (CDM) [2,3] is the optimal distance measure to use for 1 nearest-neighbour (1-NN) classification, and show that it reduces to squared Euclidean distance in feature space for function classes that can be expressed as linear combinations of a fixed set of features. PAC-like bounds are given on the sample-complexity required to learn the CDM. An experiment is presented in which a neural network CDM was learnt for a Japanese OCR environment and then used to do 1-NN classification.
[1] Jonathan Baxter. The Canonical Metric For Vector Quantization , 1995 .
[2] Jonathan Baxter,et al. Learning internal representations , 1995, COLT '95.
[3] Peter L. Bartlett,et al. Efficient agnostic learning of neural networks with bounded fan-in , 1996, IEEE Trans. Inf. Theory.
[4] Jonathan Baxter,et al. The Canonical Distortion Measure for Vector Quantization and Function Approximation , 1997, ICML.