【摘要】 针对水工监测数据不可避免地存在多重共线性,而最小二乘法又难以有效识别多重共线性并消除其对模型精度的影响的问题,初步分析了多重共线性产生的原因及其对回归模型的不良影响,简要介绍了其诊断和处理方法。引进岭回归法对共线性监测数据进行分析,其核心思想简单,算法易于编程实现,回归结果实际意义明确。某工程实例分析结果表明,相对于经典的最小二乘回归法,岭回归的结果更为合理,更符合实际,在水工监测数据统计模型分析中有较好的应用价值。更多还原

【Abstract】 Multicolinearity of hydraulic monitoring data is inevitable,and the traditional least square can not effectively identify the multicolinearity of independent variables and eliminate its negative effects on model precision.The causes of multicolinearity and its negative effects on regression models are analyzed,and its diagnosis methods are introduced.Ridge regression is introduced to analyze monitoring data,which is easy to program and the regression results are practical.A project example shows the rationality and practicability of ridge regression in hydraulic monitoring data analysis.更多还原