摘要
选择δ学习规则、有监督学习方式的BP神经网络模型作为本次应用的模型.对30个案例进行分析研究,挖掘有用的数据,找出影响工程造价的主要特征,并对这些特征按其对应分类定性处理,模拟从定性分类处理后的工程特征到工程造价的非线性映射关系.模型估算误差率在±10%以内,满足投资估算阶段的精度要求.
The BP neural network model based onδlearning rule and supervised learning style is set up in this paper.The main characters affecting construction cost were obtained through analysis of 30 cases,form which to exploit useful data.Then the main characters are classified and qualitative disposaled.Finally,the nonlinear mapping relationship of construction characteristics and construction cost is simulated.The evaluation error rate of this model is in±10%,which can satisfy the precision requirement of investment estimation stage.
出处
《数学的实践与认识》
CSCD
北大核心
2010年第21期62-67,共6页
Mathematics in Practice and Theory
基金
福建省自然科学基金项目(2008J0196)
泉州市科技计划项目(2009Z52)
关键词
BP神经网络
投资估算
数据挖掘
聚类分析
BP neural network
investment estimation
data mining
cluster analysis
