A maximum entropy approach to natural language processing

被引：28

作者：

Berger, AL ^{[1
]}

DellaPietra, SA ^{[1
]}

DellaPietra, VJ ^{[1
]}

机构：

[1] RENAISSANCE TECHNOL, STONY BROOK, NY USA

来源：

COMPUTATIONAL LINGUISTICS | 1996年 / 22卷 / 01期

关键词：

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

The concept of maximum entropy can be traced back along multiple threads to Biblical times. Only recently, however, have computers become powerful enough to permit the widescale application of this concept to real world problems in statistical estimation and pattern recognition. In this paper, we describe a method for statistical modeling based on maximum entropy. We present a maximum-likelihood approach for automatically constructing maximum entropy models and describe how to implement this approach efficiently, using as examples several problems in natural language processing.

引用

页码：39 / 71

页数：33

共 23 条

[1] [Anonymous], 1980, P WORKSH PATT REC PR
[2] BAHL L, 1989, IEEE T ACOUSTICS SPE, V37
[3] BERGER A, 1994, P ARPA C HUM LANG TE
[4] BLACK E, 1992, P DARPA SPEECH NAT L
[5] Brown D. T., 1959, INFORM CONTR, V2, P386, DOI DOI 10.1016/S0019-9958(59)80016-4
[6] Brown P., 1990, P IBM NATURAL LANGUA, V18, P283
[7] Brown P. F., 1993, Computational Linguistics, V19, P263
[8] Brown P. F., 1990, Computational Linguistics, V16, P79
[9] Brown PF, 1991, P SPEECH NAT LANG WO, P146
[10] Cover T. M., 2005, ELEM INF THEORY, DOI 10.1002/047174882X

← 1 2 3 →