On using Bayesian networks for complexity reduction in decision trees

被引：0

作者：

Adriana Brogini

Debora Slanzi

机构：

[1] University of Padova,Department of Statistics

[2] University Ca’ Foscari of Venice,Department of Statistics

来源：

Statistical Methods and Applications | 2010年 / 19卷

关键词：

Bayesian networks; Decision trees; Markov blanket; Complexity reduction; Classification;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

In this paper we use the Bayesian network as a tool of explorative analysis: its theory guarantees that, given the structure and some assumptions, the Markov blanket of a variable is the minimal conditioning set through which the variable is independent from all the others. We use the Markov blanket of a target variable to extract the relevant features for constructing a decision tree (DT). Our proposal reduces the complexity of the DT so it has a simpler visualization and it can be more easily interpretable. On the other hand, it maintains a good classification performance.

引用

页码：127 / 139

页数：12

共 17 条

[1] Acid S(2003)Searching for Bayesian network structures in the space of restricted acyclic partially directed graphs J Artif Intell Res 18 445-490
[2] de Campos L(1997)Simplifying decision trees: a survey Knowl Eng Rev 12 1-40
[3] Breslow LA(1968)Approximating discrete probability distributions with dependence trees IEEE Trans Inf Theory 14 462-467
[4] Aha DW(1992)A Bayesian method for the induction of probabilistic networks from data Mach Learn 9 309-347
[5] Chow CK(1995)Learning Bayesian networks: the combinations of knowledge and statistical data Mach Learn 20 197-243
[6] Liu CN(1994)Learning Bayesian belief networks. An approach based on the MDL principle Comput Intell 10 269-293
[7] Cooper G(1986)Induction of decision trees Mach Learn 1 81-106
[8] Herskovits E(2007)A review of feature selection techniques in bioinformatics Bioinformatics 23 2507-2517
[9] Heckerman D(undefined)undefined undefined undefined undefined-undefined
[10] Geiger D(undefined)undefined undefined undefined undefined-undefined

← 1 2 →