Estimation and learning of network parameters in semiparametric stochastic perceptron

Author

Kawanabe, Motoaki ; Amari, Shun-ichi

Author_Institution

Dept. of Math. Eng. & Inf. Phys., Tokyo Univ., Japan

Volume

2

fYear

1994

fDate

27 Jun-2 Jul 1994

Firstpage

787

Abstract

Kabashima and Shinomoto found (1992) that estimators of a binary decision boundary show asymptotically strange behaviors when the probability model is ill-posed or semiparametric. We give a rigorous analysis of this phenomenon in a stochastic perceptron by using the estimating function method. It is shown that there exists no estimator of threshold parameter h which converges to the true value in the order of 1/√n as the number n of observations increases. Instead we propose asymptotic estimating functions and analyze the asymptotics of the estimators therefrom

Keywords

decision theory; learning (artificial intelligence); parameter estimation; perceptrons; stochastic processes; asymptotic estimating functions; asymptotically strange behaviors; binary decision boundary; ill-posed probability model; neural network parameter estimation; neural network parameter learning; semiparametric probability model; semiparametric stochastic perceptron; Equations; Intelligent networks; Mathematical model; Maximum likelihood detection; Maximum likelihood estimation; Neurons; Parameter estimation; Physics; Probability distribution; Stochastic processes;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on

Conference_Location

Orlando, FL

Print_ISBN

0-7803-1901-X

Type

conf

DOI

10.1109/ICNN.1994.374278

Filename

374278