In this paper,by discussing the relationship between concrete dam deformation and environmental temperature,water level change,aging deformation and other factors,support vector machine(SVM)and minimum covariance matrix(MCD)robust estimation are used to identify outliers in the monitoring data.Through a dam safety monitoring project of a reservoir,the application of SVM-MCD algorithm based outlier recognition method is studied,and the traditional regression method is combined with the 3σ criteria of outlier judgment results are compared and analyzed,which verifies the reliability of the SVM-MCD algorithm based outlier recognition method,and can effectively solve the problem of missing judgment and misjudgment of the traditional outlier recognition method.The outliers in monitoring data can be accurately identified,to improve the reliability of concrete dam monitoring data.
维普
万方数据