Skip to main navigation Skip to search Skip to main content

Decision trees based knowledge discovery in databases for high-rise structures intelligent form selection

  • Shi Hai Zhang*
  • , Xiao Yan Liu
  • , Qing Tu
  • , Jin Ping Ou
  • *Corresponding author for this work
  • School of Civil Engineering, Harbin Institute of Technology
  • Nanyang Institute of Technology

Research output: Contribution to journalArticlepeer-review

Abstract

First, the settlement course, tactics, and technological key for KDD based on decision tree are addressed. The method of optimizing structure that makes use of weighting entropy is put forward. And with characteristic of structure form selection, the basic algorithm of ID3 is improved. Secondly, based on 20 high-rise structure cases, the cases of high-rise structures intelligent form selection assorted case-mining and part of selection intelligent from which are addressed. Finally, 1008 high-rise building cases are collected. The KDD system which is based on case-base and improves the method of decision tree is set up. And it will help the methods and measures sustain for mining structure form-selection from building cases. The given examples prove that the decision tree methods can exploit cryptic structure form-selection knowledge from building cases effectively, which has significant meaning for utilizing abundant case resources fully, improving the design of quality and efficiency structure form-selection, and offering a new approach solving the problem of bottle-neck structure intelligent form selection.

Original languageEnglish
Pages (from-to)451-454+567
JournalHarbin Gongye Daxue Xuebao/Journal of Harbin Institute of Technology
Volume37
Issue number4
StatePublished - Apr 2005
Externally publishedYes

Keywords

  • Decision tree
  • High-rise structures
  • Intelligent form selection
  • Knowledge discovery in databases

Fingerprint

Dive into the research topics of 'Decision trees based knowledge discovery in databases for high-rise structures intelligent form selection'. Together they form a unique fingerprint.

Cite this