Knowledge Discovery on Structure System Selection of High-rise Buildings Based on Bayesian Networks

Abstract

The design information of the high-rise buildings in Shanghai was collected. The tables and database that can indicate these design information was constructed. Considered the affected factors of structure system selection, the Bayesian networks were introduced to make knowledge discovery with the database as the foundation. According to the relationship between the affected factors and the optimum structure system, the Bayesian networks model was constructed. The prior probability could be calculated by the buildings data had been collected. The posterior probability of the optimum structure system could be calculated by the Bayesian networks model based on the prior probability of the affected factors. The practice proved that the Bayesian network provides one new method for structural system selection. The research results have certain practical significance to the design improvement and codes supplement for out-of-codes high-rise buildings.