基于支持向量机的超早强混凝土抗压强度预测研究

近年来，工程中对超早强混凝土的要求越来越严格，不仅要求在规定时间内达到一定强度，而且还要求经济效益最大化。超早强混凝土强度受多种因素的影响，其抗压强度预测是一个复杂的过程。因此，基于支持向量机理论，建立超早强混凝土抗压强度预测模型，选用不同的核函数及其参数进行预测，将预测值与实测值相对比，得出最优的核函数。经研究发现:线性核函数与RBF核函数的预测结果与实测值吻合较好，RBF核函数的预测值更加接近实测值。

Study of Compression Strength Prediction of Ultra High-early-strength Concrete by Support Vector Machine

In recent years, the criterion of ultra high-early-strength concrete in project is increasingly stricter, not only the time requirement of reaching certain strength, but also requirements of maximum economic effectiveness. However, the predicting process would be complicated because of the various factors which can have impact on its strength. In this paper, SVM (support vector machine) will be used to build a numerical model that can predict its strength via comparing different kernel functions and parameters, and the most optimized kernel function can be obtained. In conclusion, numerical models with linear kernel function and RBF kernel function can lead to reasonable results, and the one with RBF kernel function is the best.