中文摘要：

选择最小二乘支持向量机对混凝土测量中存在的大样本数据进行回归分析以及预测。研究表明：利用LS-SVM模型进行混凝土测量当中存在的大样本数据回归分析具有精度高,速度快等优点。模型不需建立方程式,修改容易,可处理被干扰的数据,具有较强的概括性。利用LS-SVM对冻融作用下的混凝土的断裂韧度进行回归分析以及断裂预测,模型较好的反映了断裂韧度的下降趋势。

英文摘要：

LS-SVM could be choose to carry out regression analysis and forecastong big sample data that exists on concrete measurement.Regression analysis by LS-SVM has well accuracy and speed quickly.Model needn＇t built equation and easier modify, disturbed data can be disposed and have well gengralization.using LS-SVM carry out regression analysis and forecast of concrete fracture toughness subjected to freeze-thaw cycles.Model well reflects degressive trend of concrete fracture toughness.