岩体变形模量偏最小二乘回归及神经网络关联性研究.pdfVIP

第23 卷 第22 期 岩石力学与工程学报 23(22) ：3770～3774 2004 年11月 Chinese Journal of Rock Mechanics and Engineering Nov.，2004 岩体变形模量偏最小二乘回归与神经 网络关联性研究* 马 莎 姜 彤 黄志全 李华晔 (华北水利水电学院岩土工程系 郑州 450008) 摘要 偏最小二乘回归不直接考虑因变量与自变量回归问题，而直接提取与系统有关的新的综合变量，并能利用 交叉原理确定成分个数，尤其在分析单因变量与多自变量间关系上，其所得结果更为满意；人工神经网络具有自 适应学习和记忆能力，尤其是3 层BP 网络模型。把这两者相关联，以岩体变形模量为因变量，岩石比重等7 个 因素为自变量，分析研究了焦作—晋城高速公路试验资料，所得拟合值与实测值误差最大为 5.43% ，较偏最小二 乘法回归分析 19.07% 的误差好了许多。该方法为参数选取找到了一条新的途径。 关键词 岩石力学，偏最小二乘回归，人工神经网络，岩体变形模量 分类号 TU 452 文献标识码 A 文章编号 1000-6915(2004)22-3770-05 STUDY ON PARTIAL LEAST SQUARE REGRESSION ASSOCIATED WITH NEURAL NETWORK FOR DEFORMATION MODULUS OF ROCK MASS Ma Sha ，Jiang Tong，Huang Zhiquan，Li Huaye (North China Institute of Water Conservancy and Hydroelectric Power ， Zhengzhou 450008 China) Abstract Partial least square regression method does not directly consider the regression problem of dependent variables and independent variables ，but extract the new integral variables with relation to the system. It can determine the number of components ，and especially in analysis of the correlation between the single dependent variable and multi-independent-variables ，the results are more satisfactory. The neural network has the ability of self adaptive learning and remembering，and the three-layer BP network has wide application. By associating the two methods ，with the deformation modulus of rock mass being the dependent variable and seven factors of rock mass being the independent variables ，the data from the observation of the Jiaozuo-Jincheng highway are analyzed and the maximum error between calculation and observation is 5.43%，which is better than the error of 19.07% by the p