On the Asymptotic Probability of Error in Nonparametric Discrimination
Devroye, Luc
Ann. Statist., Tome 9 (1981) no. 1, p. 1320-1327 / Harvested from Project Euclid
Let $(X, Y), (X_1, Y_1), \cdots, (X_n, Y_n)$ be independent identically distributed random vectors from $R^d \times \{0, 1\}$, and let $\hat{Y}$ be the $k$-nearest neighbor estimate of $Y$ from $X$ and the $(X_i, Y_i)$'s. We show that for all distributions of $(X, Y)$, the limit of $L_n = P(\hat{Y} \neq Y)$ exists and satisfies $\lim_{n\rightarrow\infty} L_n \leq (1 + a_k)R^\ast, a_k = \frac{\alpha \sqrt k}{k - 3.25}\big(1 + \frac{\beta}{\sqrt{k-3}}\big), k \text{odd}, k \geq 5,$ where $R^\ast$ is the Bayes probability of error and $\alpha = 0.3399 \cdots$ and $\beta = 0.9749 \cdots$ are universal constants. This bound is shown to be best possible in a certain sense.
Publié le : 1981-11-14
Classification:  Nonparametric discrimination,  pattern recognition,  inequality of Cover and Hart,  nearest neighbor rule,  probability of error,  62G05
@article{1176345648,
     author = {Devroye, Luc},
     title = {On the Asymptotic Probability of Error in Nonparametric Discrimination},
     journal = {Ann. Statist.},
     volume = {9},
     number = {1},
     year = {1981},
     pages = { 1320-1327},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1176345648}
}
Devroye, Luc. On the Asymptotic Probability of Error in Nonparametric Discrimination. Ann. Statist., Tome 9 (1981) no. 1, pp.  1320-1327. http://gdmltest.u-ga.fr/item/1176345648/