摘要
住宅工程的投资费用影响着工程决策、投资经济效益等多方面,设计方案的选择影响着工程全生命期造价。本文针对山东省潍坊市4个县市区的24个工程案例,采用MLP神经网络进行模型构建,确定了16个影响造价的主要因素作为网络输入变量,并对输入参数进行归一化处理,其中20组作为训练样本,4组作为检验样本,对建筑工程项目造价进行准确的预测。经检验,基于MLP模型进行住宅工程造价估算,其预测精度符合工程投资估算标准要求。
The investment cost of residential projects affects many aspects such as engineering decision-making and investment economic benefits. The choice of design scheme affects the whole life-cycle cost of the project. In this paper, based on 24 engineering cases in 4 counties and cities of Weifang City, Shandong Province, MLP neural network was used to construct the model, and 16 main factors affecting the cost were determined as network input variables, and the input parameters were normalized. 20 groups were used as training samples, and 4 groups were used as test samples to accurately predict the cost of construction projects. After inspection, the MRP model is used to estimate the cost of residential construction, and its prediction accuracy meets the requirements of engineering investment estimation standards.
作者
汤青慧
李盼盼
李双双
TANG Qing-hui;LI Pan-pan;LI Shuang-shuang(School of Management,Qingdao University of Technology,Qingdao 266520,China)
出处
《价值工程》
2019年第14期1-4,共4页
Value Engineering
