| 宁志杰.基于灰色支持向量机残差修正模型的地面沉降量预测[J].地质与勘探,2021,57(3):614-620 |
| 基于灰色支持向量机残差修正模型的地面沉降量预测 |
| Prediction of land subsidence based on a residual correction model of the grey support vector machine |
| 投稿时间：2019-12-02 修订日期：2021-04-08 |
| DOI：10.12134/j.dzykt.2021.03.014 |
| 中文关键词: 地面沉降 沉降量预测 GM(1,1)模型 GM(1,1)-SVM模型 |
| 英文关键词: land subsidence, land subsidence prediction, GM (1,1) model, GM (1,1) -SVM model |
| 基金项目:国家自然科学基金资助项目(编号：41807231)、河北省自然科学基金项目(编号：D2019403182)和河北省教育厅青年基金项目(编号：QN2019196) 联合资助 |
|
| 摘要点击次数: 1621 |
| 全文下载次数: 607 |
| 中文摘要: |
| 沉降现象在各地区普遍发生，地面沉降量预测越来越受到重视。本文通过结合灰色 (GM(1,1))预测模型和支持向量机(SVM)模型各自的优点，建立灰色支持向量机（GM(1,1)-SVM）残差修正模型，突出时间序列发展趋势影响的同时降低序列中异常值的消极作用。以某高层建筑的18次地面沉降量数据为实例，检验GM(1,1)-SVM模型的预测效果。结果表明：相对单一的GM(1,1)沉降量预测模型，GM(1,1)-SVM模型相对误差小，预测精度高，对地面沉降量预测有一定指导意义。 |
| 英文摘要: |
| Land subsidence is a common geologic hazard in many regions. The prediction of the amount of this surface process has received increasing attention. In this work, a residual correction model of the gray support vector machine (GM(1,1)-SVM) is established by combining the advantages of the gray prediction (GM(1,1)) model and support vector machine (SVM) model. It highlights the influence of time series development trend and lowers the negative effect of outliers in the sequence. Taking the data of 18 times of land subsidence of a high building as an example, the prediction effect of the GM(1,1)-SVM model is tested. The results show that the GM(1,1)-SVM model has a relatively small relative error and high prediction accuracy compared with the single GM(1,1) model, which has certain guiding significance for the prediction of land subsidence. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
