Deformation prediction of foundation pit with PCA-SVM

摘要

Using the principal component (PCA) strong ability of extracting effective features of foundation pit horizontal displacement monitoring data (monitoring, monitoring temperature, relative humidity, excavation depth) characteristics analysis extracting the effective principal component, constructing the PCA SVM regression prediction model, and the analysis result through comparing with the measured values show that: the displacement data of prediction model based on PCA and SVM was more higher accuracy than a model using SVM, higher reliability, which means has certain applicability in engineering application.