基于极限学习机的建筑工程施工成本预测研究  

Research on Construction cost Prediction of Construction Project Based on Extreme Iearning Machine

作　　者：李普祥 LI Pu-xiang(Weifang Sanjian Group Co.,Ltd.Weifang 261100,China)

机构地区：[1]潍坊市三建集团有限公司,山东潍坊261100

出　　处：《全面腐蚀控制》2024年第8期84-86,共3页Total Corrosion Control

摘　　要：随着建筑工程行业的发展,施工成本预测越来越受到重视。为了提高预测准确性,我们提出了一种基于极限学习机(ELM)和支持向量机(SVM)的融合预测模型。首先,我们分别构建了基于ELM和SVM的子预测模型,然后通过加权融合策略将两种子模型结合在一起。实验结果表明,所提出的融合预测模型在建筑工程施工成本预测方面表现优越,具有较高的预测精度和稳定性。我们对比了ELM、SVM和多层感知机(MLP)等方法的性能,并对融合预测模型进行了详细的分析。此外,我们还讨论了未来研究方向,以期推动建筑工程施工成本预测领域的发展。With the development of the construction engineering industry,construction cost prediction has been paid more and more attention.To improve the prediction accuracy,we propose a fusion prediction model based on extreme learning machine(ELM)and support vector machine(SVM).First,we construct a sub-prediction model based on ELM and SVM respectively,and then combine the two seed models by a weighted fusion strategy.The experimental results show that the proposed fusion prediction model is superior in the construction cost prediction,and has high prediction accuracy and stability.We compare the performance of ELM,SVM and multi-layer perceptron(MLP),and analyze the fusion prediction model in detail.In addition,we also discussed the future research direction,with a view to promoting the development of the construction cost forecasting field of construction engineering.

关 键 词：建筑工程 施工成本预测 极限学习机 支持向量机 融合预测模型 

分 类 号：TU723[建筑科学—建筑技术科学]