论文信息 - Generalization Error of Multinomial Classifier

Generalization Error of Multinomial Classifier

Equation for generalization error of Multinomial classifier is derived and tested. Particular attention is paid to imbalanced training sets. It is shown that artificial growth of training vectors of less probable class could be harmful. Use of predictive Bayes approach to estimate cell probabilities of the classifier reduces both the generalization error and effect of unequal training sample sizes.

Sarunas Raudys

[1] B. Chandrasekaran,et al. Quantization Complexity and Independent Measurements , 1974, IEEE Transactions on Computers.

[2] G. F. Hughes,et al. On the mean accuracy of statistical pattern recognizers , 1968, IEEE Trans. Inf. Theory.

[3] Mathukumalli Vidyasagar,et al. A Theory of Learning and Generalization , 1997 .

[4] David G. Stork,et al. Pattern Classification , 1973 .

[5] Robert P. W. Duin,et al. K-nearest Neighbors Directed Noise Injection in Multilayer Perceptron Training , 2000, IEEE Trans. Neural Networks Learn. Syst..

[6] Shun-ichi Amari,et al. A universal theorem on learning curves , 1993, Neural Networks.

[7] David G. Stork,et al. Pattern Classification (2nd ed.) , 1999 .

[8] Mathukumalli Vidyasagar,et al. A Theory of Learning and Generalization: With Applications to Neural Networks and Control Systems , 1997 .

[9] Sarunas Raudys,et al. Experts' Boasting in Trainable Fusion Rules , 2003, IEEE Trans. Pattern Anal. Mach. Intell..

[10] Vladimir Vapnik,et al. Statistical learning theory , 1998 .