论文部分内容阅读
决定LS-SVM性能的因素是惩罚因子C和核函数的选取。核函数通常选用RBF核函数。本课题基于遗传模拟退火算法来优化LS-SVM的参数组合(C,γ),并通过土石坝渗水量的数据做实例分析,测试结果同没有进行参数优化的最小二乘支持向量机测试结果和BP神经网络测试结果进行对比,结果表明基于改进遗传算法优化最小二乘支持向量机参数的建模方式,建模速度和预测精度同后面两种方法相比都有所提高,有着广阔的应用前景。
The factors that determine the performance of LS-SVM are the selection of penalty factor C and kernel function. The kernel function usually chooses RBF kernel function. In this paper, genetic algorithm based on annealing algorithm to optimize the LS-SVM parameter combinations (C, γ), and through earth-rock dam water seepage data as an example, the test results with no optimization of least squares support vector machine test results and BP neural network test results are compared, the results show that based on improved genetic algorithm to optimize the least square support vector machine parameters modeling method, the modeling speed and prediction accuracy compared with the latter two methods have improved, has broad application prospects .