Data-dependent k/sub n/-NN and kernel estimators consistent for arbitrary processes

IEEE Transactions on Information Theory - Tập 48 Số 10 - Trang 2785-2788 - 2002
S.R. Kulkarni1, S.E. Posner2, S. Sandilya3,1
1Department of Electrical Engineering, Princeton University, Princeton, NJ, USA
2Goldman Sachs and Company, New York, NY, USA
3Siemens Corporate Research, Inc., Princeton, NJ, USA

Tóm tắt

Let X/sub 1/, X/sub 2/,... be an arbitrary random process taking values in a totally bounded subset of a separable metric space. Associated with X/sub i/ we observe Y/sub i/ drawn from an unknown conditional distribution F(y|X/sub i/=x) with continuous regression function m(x)=E[Y|X=x]. The problem of interest is to estimate Y/sub n/ based on X/sub n/ and the data {(X/sub i/, Y/sub i/)}/sub i=1//sup n-1/. We construct appropriate data-dependent nearest neighbor and kernel estimators and show, with a very elementary proof, that these are consistent for every process X/sub 1/, X/sub 2/,.

Từ khóa

#Stochastic processes #Parameter estimation #Set theory

Tài liệu tham khảo

10.1080/02331889908802686 10.1214/aos/1018031110 10.1109/18.661503 10.1007/BF02532233 10.1007/978-94-011-3222-0 10.1214/aos/1176343886 10.1016/0047-259X(89)90091-2 10.1016/0304-4149(93)90050-E 10.1016/0047-259X(87)90105-9 dudley, 1989, Real Analysis and Probability 10.1007/BF00531618 györfi gyorfi, 1989, Nonparametric Curve Estimation from Time Series, 10.1007/978-1-4612-3686-3 györfi gyorfi, 1981, the rate of convergence of <formula><tex>$k_n$</tex></formula>-nn regression estimates and classification rules, IEEE Trans Information Theory, it 27, 362, 10.1109/TIT.1981.1056344 10.1109/18.391248 10.1109/TIT.1986.1057226 cover, 1968, rates of convergence for nearest neighbor procedures, Proc 1st Annu Hawaii Conf Systems Theory, 413 beck, 1979, the exponential rate of convergence of error for <formula><tex>$k_n$</tex></formula>-nn nonparametric regression and decision, Prob Contr Inform Theory, 8, 303 10.1109/18.650998