摘要
采用误差反向传播人工神经网络模型(BP网络模型),以建筑特征参数为输入变量,通过实际资料对网络进行训练和模拟,并用贡献分析法筛选输入变量,对网络结构进行优化,结果显示了该模型在建筑工程造价预测中的有效性。
In the constantly changing marketing economy, it has become an urgent task for construction industry to find a rapid, simple and practical way to organize construction project budget. To solve this problem, this paper adopts the model of the back-propagation neural network, takes the features of construction as input variables, trains the network using actual data as samples and optimizes the network structure by contribution analysis. It shows the validity of the model in the forecast of construction project budget.
出处
《中国工程科学》
2005年第10期56-59,共4页
Strategic Study of CAE
关键词
BP神经网络
建筑造价
预测
BP neural network
building budget
forecast
