SVM-multiple-class

this util for voice analize

源码给出了非线性规划（NLP）的梯度法，其中gradf.m为梯度法算法，下载设置参数既能运行。(Given nonlinear programming source (NLP) in the gradient method, which gradf.m as gradient algorithm, download the set parameters not only run.)

根据小波的分解算法与重构算法来实现小波的分解与重构(Algorithm based on wavelet decomposition and reconstruction algorithm to achieve the wavelet decomposition and reconstruction)

对数据先进行经验模态分解得到不同特征尺度的信号分量，然后对所有的信号分量用AR模型进行预测得到预测后的分量，然后将预测后的分量进行重构得到最终的预测信号。(The data first empirical mode decomposition characteristics of the signal components of different scales, and then for all signal components obtained by AR model to predict the component forecasts, and then predict the components are reconstructed to obtain final predicted signal.)

《matlab程序设计教程》里的程序代码，源程序，每一章节都有例子()

LS-SVM Leave-One-Out Cross-Validation Demo G. C. Cawley, "Leave-one-out cross-validation based model selection criteria for weighted LS-SVMs", Proceedings of the International Joint Conference on Neural Networks (IJCNN-2006), pages 1661-1668, Vancouver, BC, Canada, July 16-21 2006. (pdf)theoval.cmp.uea.ac.uk/~gcc/matlab/

direct kinematics neural network in matlab

通过计算机MATLAB仿真分析和研究了维纳滤波器的阶数、信号的噪声方差、随机信号的采样点数、经过维纳滤波的均方误差之间的关系(MATLAB simulation by computer analysis and research of Wiener filter order, the signal of the noise variance, random signal sampling points, after Wiener filter mean square error of the relationship between)

说明：  一个感知器神经网络用于分类，有需要的人可以下载啊(A percep used in classification)

Computes the modified hausdorff distance