基于IGWO-SVR的地铁车站投资预测  

Subway Station Investment Prediction Based on IGWO-SVR

作　　者：郝晶晶 段鹏鑫[2] 陈雨欣 段晓晨[3] HAO Jingjing;DUAN Pengxin;CHEN Yuxin;DUAN Xiaochen(School of Civil Engineering,Shijiazhuang Tiedao University,Shijiazhuang 050043,China;College of Mathematics and Statistics,Hebei University of Economics and Business,Shijiazhuang 050061,China;School of Management,Shijiazhuang Tiedao University,Shijiazhuang 050043,China)

机构地区：[1]石家庄铁道大学土木工程学院,河北石家庄050043 [2]河北经贸大学数学与统计学学院,河北石家庄050061 [3]石家庄铁道大学管理学院,河北石家庄050043

出　　处：《铁道学报》2024年第5期179-188,共10页Journal of the China Railway Society

基　　金：国家自然科学基金(72071133);河北省自然科学基金(G2019210226);河北省研究生创新资助(CXZZBS2020144)。

摘　　要：为快速准确地预测可行性研究阶段的地铁车站投资以给投资方提供决策支持,提出一种改进的灰狼算法,来优化支持向量回归机参数的方法。通过收集整理地铁车站案例数据并建立数据库作为预测原始样本,利用自助法扩充样本以解决小样本预测准确率低的问题;对灰狼算法的初始种群生成、收敛因子和位置更新方式进行改进以避免灰狼算法陷入局部最优的问题;使用改进的灰狼算法对支持向量回归机的参数进行优化,建立地铁车站投资预测模型;以车站实例验证预测模型有效性。结果表明,改进的灰狼算法优化参数的支持向量回归机的预测模型在测试集平均相对误差为4.33%,拟合优度为0.9440,改进的灰狼算法优化参数后预测模型优于未经优化的、粒子群优化的和灰狼算法优化的支持向量回归机预测模型。实例红高路车站相对误差为4.87%,证明提出的模型是有效的。To predict the investment of subway stations quickly and accurately in the feasibility study stage and provide decision support for investors,this paper proposed a method to optimize the parameters of support vector regression machine based on improved grey wolf optimizer.Firstly,by collecting and sorting out cases of subway stations and establishing a database as the original prediction sample,bootstrapping was adopted to expand the sample to solve the problem of low accuracy of small sample prediction.Secondly,the initial population generation,convergence factor and location update of the grey wolf optimizer were improved to avoid the problem of grey wolf optimizer being stuck in local optima.Subsequently,the improved grey wolf optimizer was used to optimize the parameters of support vector regression machine to establish a subway station investment prediction model.Finally,a station example was used to verify the validity of the prediction model.The results show that the average relative error of the prediction model of the support vector regression machine optimized by the improved grey wolf algorithm is 4.33%,with the goodness of fit of 0.9440 on the test set.The prediction model based on improved grey wolf algorithm with optimized parameters is better than the unoptimized,particle swarm optimized and grey wolf algorithm optimized support vector regression prediction models.The relative error of the case of Honggaolu station is 4.87%,proving the effectiveness of the proposed prediction method.

关 键 词：投资预测 自助法 灰狼算法 支持向量回归机 地铁车站 

分 类 号：U231.4[交通运输工程—道路与铁道工程]