Construction of Learning Algorithm based on SGA Bayesian Network

Author

Jia, T.J.

Author_Institution

Coll. of Electron. & Inf., Shanghai Dianji Univ., Shanghai

fYear

2008

fDate

3-5 Aug. 2008

Firstpage

37

Lastpage

40

Abstract

A typical characteristic of Bayesian network topology is dependences of each variable within the network, which makes it impossible to optimize variables. This problem is solved by the developed approach to Bayesian network construction based on self-organizing genetic algorithm (SGA) from knowledge base. The genetic algorithm (GA) is improved by self-organizing organism and an effective operator is provided to search the global optimum value in order to avoid an early convergence for a normal GA algorithm. At last the experiment results and the convergence of SGA are discussed.

Keywords

belief networks; genetic algorithms; learning (artificial intelligence); Bayesian network construction; Bayesian network topology; learning algorithm; self-organizing genetic algorithm; self-organizing organism; Bayesian methods; Convergence; Educational institutions; Electronic commerce; Genetic algorithms; Information security; Network topology; Organisms; Polynomials; Uncertainty; Bayesian learning based on Knowledge base; Bayesian network; Self-organizing Genetic Algorithm (SGA); network safety applies;

fLanguage

English

Publisher

ieee

Conference_Titel

Electronic Commerce and Security, 2008 International Symposium on

Conference_Location

Guangzhou City

Print_ISBN

978-0-7695-3258-5

Type

conf

DOI

10.1109/ISECS.2008.23

Filename

4606020