基于粗糙集和粒子群优化支持向量机的滑坡变形预测

来源期刊:中南大学学报(自然科学版)2015年第6期

论文作者:牛瑞卿 赵艳南 彭令 程温鸣

文章页码:2324 - 2333

关键词:滑坡变形预测;粗糙集;粒子群优化;支持向量机

Key words:landslide deformation prediction; rough sets; particle swarm optimization; support vector machine

摘    要:以三峡库区白水河滑坡为例,首先分析降雨量与库水位等影响因素与滑坡变形特征的响应关系,然后利用粗糙集理论对10个初始影响因子进行属性约减,筛选出影响滑坡变形的核因子集,最后基于该因子集建立粒子群优化支持向量回归模型,对滑坡位移速率进行预测。研究结果表明:测试样本的预测结果与实测值变化趋势基本一致,其平均绝对误差为0.234 mm/d,均方差和判定系数分别为0.163和0.520。粗糙集理论在分析滑坡变形特征、筛选关键因子方面的适用性与科学性,构建的粗糙集-粒子群优化支持向量机模型具有较高的泛化能力,是一种有效的滑坡变形预测方法。

Abstract: The Baishuihe landslide in the Three Gorges Reservoir region was selected as an example. By analysing the response relationships between landslide deformation and influencing factors such as the rainfall and the reservoir water level, 10 initial influencing factors were reduced by using the rough set theory(RS). Then, the nuclear factor set influencing the landslide deformation was screened out. Finally, the particle swarm optimization (PSO) - support vector regression (SVR) model was established based on the nuclear factor set to predict landslide displacement rate. The results show that the test sample predictive mean absolute error, mean squared error and determination coefficient are 0.234 mm/d, 0.163 and 0.520, respectively. And the change trends are consistent between predicted results and the measured ones. The rough set theory is scientific and applicable in analysing landslide deformation characteristics and selecting key factors. The RS-PSO-SVR model is an effective method in landslide deformation predicting with high generalization ability.

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号