摘要
为了探索超限学习机在路基沉降预测应用中的潜力和优势,以湖南省某高速公路为研究对象,通过超限学习机算法对路基多个断面的实测沉降数据进行了预测建模,并与BP神经网络和支持向量机进行了对比分析。研究结果表明,采用超限学习机对K4+300断面和K20+840断面的预测值的最大相对误差分别为0.199%和0.176%,精度明显优于BP神经网络和支持向量机。故超限学习机能够对路基沉降做出较为科学、合理的预测。
In order to explore the potential and advantages of extreme learning machine in the application of subgrade settlement prediction, a highway in Hunan Province it taken as the research object and the extreme learning machine is introduced to forecast the measured settlement data of multiple sections of subgrade, BP neural network (backtracking neural network) and support vector machine (SVM) for comparative analysis. The results show that the maximum relative error of K4 + 300 section and K20 -? 840 section is 0. 199% mm and 0. 176% mm, and the accuracy is better than BP neural network and support vector ma- chine. The extreme learning machine can make a scientific and reasonable prediction for sub- grade settlement.
出处
《长沙理工大学学报(自然科学版)》
CAS
2017年第4期44-48,89,共6页
Journal of Changsha University of Science and Technology:Natural Science
基金
国家自然科学基金资助项目(41471421
41671498)
关键词
超限学习机
路基沉降预测
BP神经网络
支持向量机
建模
预测
extreme learning machine
prediction of subgrade settlement
BP neural net-work
support vector machine
modeling
prediction