Bayes Error Estimation Using Parzen and k-NN Procedures
- 1 September 1987
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. PAMI-9 (5) , 634-643
- https://doi.org/10.1109/tpami.1987.4767958
Abstract
The use of k nearest neighbor (k-NN) and Parzen density estimates to obtain estimates of the Bayes error is investigated under limited design set conditions. By drawing analogies between the k-NN and Parzen procedures, new procedures are suggested, and experimental results are given which indicate that these procedures yield a significant improvement over the conventional k-NN and Parzen procedures. We show that, by varying the decision threshold, many of the biases associated with the k-NN or Parzen density estimates may be compensated, and successful error estimation may be performed in spite of these biases. Experimental results are given which demonstrate the effect of kernel size and shape (Parzen), the size of k (k-NN), and the number of samples in the design set.Keywords
This publication has 21 references indexed in Scilit:
- Consistent Cross-Validated Density EstimationThe Annals of Statistics, 1983
- On the Almost Everywhere Convergence of Nonparametric Regression Function EstimatesThe Annals of Statistics, 1981
- Choosing the window width when estimating a densityBiometrika, 1978
- Consistent Nonparametric RegressionThe Annals of Statistics, 1977
- The Strong Uniform Consistency of Nearest Neighbor Density EstimatesThe Annals of Statistics, 1977
- Consistency Properties of Nearest Neighbor Density Function EstimatorsThe Annals of Statistics, 1977
- Nonparametric estimates of probability densitiesIEEE Transactions on Information Theory, 1975
- Optimization of k nearest neighbor density estimatesIEEE Transactions on Information Theory, 1973
- A Nonparametric Estimate of a Multivariate Density FunctionThe Annals of Mathematical Statistics, 1965
- On Estimation of a Probability Density Function and ModeThe Annals of Mathematical Statistics, 1962