利用主成分分析改进SVM模型在大坝变形分析中的应用

Improvement of the Application of SVM Model in Dam Deformation Based on Principal Component Analysis

传统的大坝变形方法主要是基于大量数据统计分析得到,而大量数据的获得显得异常困难,同时变量与变量之间会存在相关性,针对此,本文将主成分分析引入到大坝变形分析的数据预处理中,利用该方法对变形分析的因子进行分析,使得这些因子之间不存在相关性,同时又保持原有的信息要素,再将这些变量作为大坝的影响因子。利用经过主成分分析的因子对大坝变形进行分析的过程中,保证了分析的准确性,同时又提高了分析效率。

Traditional method of dam deformation is mainly based on statistical analysis of large amounts of data which are very difficult to acquire and there exist correlations between variables. This paper introduces the principal component analysis into the preprocessing of data, and uses it to analyze the influencing factors of dam deformation, which enables these factors without correlation but with all the original information at the same time.Then we take them as the influencing factors of dam deformation and use the factors analyzed by principal component to analyze dam deformation, which ensures the accuracy and improves the efficiency of the analysis.