Title :
The finite-sample risk of the k-nearest-neighbor classifier under the Lp metric
Author :
Snapp, Robert R. ; Venkatesh, Santosh S.
Author_Institution :
Dept. of Comput. Sci. & Electr. Eng., Vermont Univ., Burlington, VT, USA
Abstract :
The finite-sample risk of the k-nearest neighbor classifier that uses an L2 distance function is examined. For a family of classification problems with smooth distributions in Rn, the risk can be represented as an asymptotic expansion in inverse powers of the n-th root of the reference-sample size. The leading coefficients of this expansion suggest that the Euclidean or L2 distance function minimizes the risk for sufficiently large reference samples
Keywords :
pattern classification; random processes; signal sampling; smoothing methods; Euclidean distance function; Lp metric; asymptotic expansion; classification problems; distance function; finite-sample risk; inverse powers; k-nearest-neighbor classifier; leading coefficients; pattern classification; random sample; reference-sample size; smooth distributions; Computer science; Euclidean distance; Laboratories; Nearest neighbor searches; Random processes; Testing; USA Councils;
Conference_Titel :
Information Theory and Statistics, 1994. Proceedings., 1994 IEEE-IMS Workshop on
Conference_Location :
Alexandria, VA
Print_ISBN :
0-7803-2761-6
DOI :
10.1109/WITS.1994.513925
