To improve the performance of support vector machines(SVM),a hybrid kernel is constructed from the existing common kernels,and the hyper-parameters are optimized by using a quasi-Newton method.
为了提高支持向量机(SVM)的识别性能,提出了在常用内核的基础上构造一个组合内核函数,然后用拟牛顿算法对其超参数进行优化的方法。
参考来源 - 组合内核与优化算法在羽绒识别系统中的应用·2,447,543篇论文数据，部分数据来源于NoteExpress
With maximizing the marginal likelihood function of hyper-parameters, the optimal weights are acquired, i. e. the reconstructed image.
最大化超参数的边缘对数似然函数求得权值参数的最优估计即待重建图像。
Due to the non-convex of the prior function and hyper-parameters, we use the dynamic posterior simulation rather than the general optimization methods to get reconstruction image.
由于采用的先验函数是非凸的并包含超验参数，一般的优化方法难以处理，采用动态后验模拟的方法可以很好地解决这些问题。
To improve the performance of support vector machines (SVM), a hybrid kernel is constructed from the existing common kernels, and the hyper-parameters are optimized by using a quasi-Newton method.
为了提高支持向量机(SVM)的识别性能，提出了在常用内核的基础上构造一个组合内核函数，然后用拟牛顿算法对其超参数进行优化的方法。
应用推荐