基于大数据与机器学习的配电网电缆线路工程造价预测  

Project Cost Prediction of Distribution Network Cable Line Based on Big Data and Machine Learning

作　　者：李谟兴 何永秀[1] 柳洋[1] 陈威君 李存斌[1] LI Moxing;HE Yongxiu;LIU Yang;CHEN Weijun;LI Cunbin(North China Electric Power University,Beijing 102206,China)

机构地区：[1]华北电力大学,北京102206

出　　处：《山东电力技术》2023年第1期40-46,共7页Shandong Electric Power

基　　金：国家自然科学基金项目(71671065)。

摘　　要：为了解决目前配电网工程造价影响因素繁多且复杂,工程造价难以准确预测问题,提出一种基于大数据与机器学习算法的配电网电缆线路工程造价组合预测模型。首先基于灰色关联分析法从工程造价的大数据中选取重要造价影响因素,其次基于交叉验证与网格搜索算法对最小二乘向量机算法进行关键参数寻优,最后利用寻优之后的最小二乘支持向量机算法进行造价预测。通过不同预测方法结果对比,验证了所构建的造价预测模型能有效提升预测速度和精度,为实现配电网电缆线路工程造价精准预测提供参考。In order to solve the problems of many and complex factors affecting the project cost of distribution network,and it is difficult to accurately predict the project cost,a combined prediction model for the project cost of distribution network cable lines based on big data and machine learning algorithm was proposed.Firstly,important cost influencing factors were selected from the big data of engineering cost based on the gray correlation analysis method.Then,the key parameters of the least squares vector machine algorithm were optimized based on cross-validation and grid search algorithm.Finally,the least squares support vector machine algorithm after optimization was used to predict the cost.Through the comparison of the results of different prediction methods,it was verified that the project cost prediction model can effectively improve the prediction speed and accuracy,and provide a reference for realizing the accurate prediction of cable line project cost in distribution network.

关 键 词：机器学习 最小二乘支持向量机 造价预测 

分 类 号：C931.3[经济管理—管理学]