Generalized Maximum Entropy for Supervised Classification

被引:12
|
作者
Mazuelas, Santiago [1 ,2 ]
Shen, Yuan [3 ,4 ]
Perez, Aritz [1 ]
机构
[1] Basque Ctr Appl Math BCAM, Bilbao 48009, Spain
[2] IKERBASQUE Basque Fdn Sci, Bilbao 48009, Spain
[3] Tsinghua Univ, Dept Elect Engn, Beijing 100190, Peoples R China
[4] Beijing Natl Res Ctr Informat Sci & Technol, Beijing 100084, Peoples R China
基金
中国国家自然科学基金;
关键词
Entropy; Uncertainty; Probability distribution; Training; Codes; Channel coding; Decision making; Supervised classification; minimax risk classifiers; maximum entropy; generalized entropy;
D O I
10.1109/TIT.2022.3143764
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The maximum entropy principle advocates to evaluate events' probabilities using a distribution that maximizes entropy among those that satisfy certain expectations' constraints. Such principle can be generalized for arbitrary decision problems where it corresponds to minimax approaches. This paper establishes a framework for supervised classification based on the generalized maximum entropy principle that leads to minimax risk classifiers (MRCs). We develop learning techniques that determine MRCs for general entropy functions and provide performance guarantees by means of convex optimization. In addition, we describe the relationship of the presented techniques with existing classification methods, and quantify MRCs performance in comparison with the proposed bounds and conventional methods.
引用
收藏
页码:2530 / 2550
页数:21
相关论文
共 50 条
  • [41] Generalized Core Functions of Maximum Entropy Theory of Ecology
    Corcino, Cristina B.
    Corcino, Roberto B.
    Picardal, Jay P.
    [J]. EUROPEAN JOURNAL OF PURE AND APPLIED MATHEMATICS, 2024, 17 (03): : 1674 - 1684
  • [42] GENERALIZED VERSUS CLASSICAL MAXIMUM-ENTROPY FOR IMAGING
    BAJKOVA, AT
    [J]. VERY HIGH ANGULAR RESOLUTION IMAGING, 1994, (158): : 215 - 217
  • [43] Maximum cross-entropy generalized series reconstruction
    Hess, CP
    Liang, ZP
    Webb, AG
    Lauterbur, PC
    [J]. 1998 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING - PROCEEDINGS, VOL 1, 1998, : 19 - 23
  • [44] GENERALIZED MAXIMUM-ENTROPY DECONVOLUTION OF SPECTRAL SEGMENTS
    MAZZEO, AR
    DELSUC, MA
    KUMAR, A
    LEVY, GC
    [J]. JOURNAL OF MAGNETIC RESONANCE, 1989, 81 (03): : 512 - 519
  • [45] Statistical Distributions "generalized" from the principle of maximum entropy
    Sotolongo-Costa, O.
    Gonzalez, A.
    Brouers, F.
    [J]. REVISTA CUBANA DE FISICA, 2009, 26 (2B): : 262 - 266
  • [46] Maximum cross-entropy generalized series reconstruction
    Hess, CP
    Liang, ZP
    Lauterbur, PC
    [J]. INTERNATIONAL JOURNAL OF IMAGING SYSTEMS AND TECHNOLOGY, 1999, 10 (03) : 258 - 265
  • [47] Generalized maximum entropy approach to unreplicated factorial experiments
    Arat, Mustafa Murat
    Aktas, Serpil
    [J]. STATISTICS AND ITS INTERFACE, 2016, 9 (03) : 295 - 302
  • [48] CLASSIFICATION OF EXTENDED GENERALIZED QUADRANGLES WITH MAXIMUM DIAMETER
    DELFRA, A
    GHINELLI, D
    [J]. DISCRETE MATHEMATICS, 1992, 105 (1-3) : 13 - 23
  • [49] Maximum Approximate Entropy for Normal and Pathological Voices Classification
    Restrepo, Juan F.
    Schlotthauer, Gaston
    Torres, Maria E.
    [J]. VI LATIN AMERICAN CONGRESS ON BIOMEDICAL ENGINEERING (CLAIB 2014), 2014, 49 : 548 - 551
  • [50] Scalable packet classification through maximum entropy hashing
    Choi, L
    Heo, J
    Kim, H
    Joung, J
    Kim, S
    [J]. NETWORKING 2004: NETWORKING TECHNOLOGIES, SERVICES, AND PROTOCOLS; PERFORMANCE OF COMPUTER AND COMMUNICATION NETWORKS; MOBILE AND WIRELESS COMMUNICATIONS, 2004, 3042 : 296 - 307