摘要
针对单纯使用遗传算法处理大规模数据需要时间长和对计算机的内存等硬件要求较高的问题,将神经网络嵌入到遗传算法中构造出混合智能遗传算法用于SVM核函数的参数优化,数值试验结果表明该算法对SVM核参数优化是可行的、有效的,并能得到较好的SVM核参数组合和具有较高的分类准确率及较好的泛化能力.
For using classic genetic algorithm requires long hours and a higher demand for computer hardware, a new algorithm is applied to the parameter optimization of the SVM kernel function and combines the nonlinear fitting capabilities of the neural network with the global optimization capability of the genetic algorithm. The numerical test results show that the algorithm is feasible and effective for the SVM kernel parameter optimization, which can get better SVM kernel parameter combinations and has high classification accuracy and better generalization ability.
出处
《数学的实践与认识》
CSCD
北大核心
2014年第1期200-204,共5页
Mathematics in Practice and Theory
关键词
支持向量机
遗传算法
参数优化
神经网络
support vector machine
genetic algorithm
parameter optimization
neuralnetwork
