An input-output hidden Markov model for tree transductions

被引：11

作者：

Bacciu, Davide ^{[1
]}

Micheli, Alessio ^{[1
]}

Sperduti, Alessandro ^{[2
]}

机构：

[1] Univ Pisa, Dipartimento Informat, I-56100 Pisa, Italy

[2] Univ Padua, Dipartimento Matemat, I-35100 Padua, Italy

来源：

NEUROCOMPUTING | 2013年 / 112卷

关键词：

Bottom-up hidden tree Markov model; Input-driven generative model; Learning tree transductions; Structured data processing; GENERAL FRAMEWORK; STRUCTURED DATA; EDIT DISTANCE;

D O I：

10.1016/j.neucom.2012.12.044

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

The paper introduces an input-driven generative model for tree-structured data that extends the bottom-up hidden tree Markov model to non-homogeneous state transition and emission probabilities. We show how the proposed input-driven approach can be used to realize different types of structured transductions between trees. A thorough experimental analysis is proposed to investigate the advantage of introducing an input-driven dynamics in structured-data processing. The results of this analysis suggest that input-driven models can capture more discriminative structural information than homogeneous approaches in computational learning tasks, including document classification and more general substructure categorization. (C) 2013 Elsevier B.V. All rights reserved.

引用

页码：34 / 46

页数：13

共 33 条

[1] Learning Nonsparse Kernels by Self-Organizing Maps for Structured Data [J].