mtd

说明：  贝叶斯网络的matlab实现。可以创建贝叶斯网络、训练模型，以及实现统计推理(Bayesian network Matlab achieve. Bayesian networks can create, model training, and the realization of statistical reasoning)

afsk1200 decoder for windows

介绍了simulink 构建BP神经网络模型(Introduced the BP neural network model building simulink)

说明：  正交匹配追踪，对MP的改进，实现信号的稀疏分解(Orthogonal matching pursuit, the improvement on the MP and the signal is sparse decomposition)

低通、点阻和梳子滤波器的实现，作为作业提交的，用于低通去除音频中的噪声(Low pass, point barrier and realization of the comb filter, as the job submission, for low pass removal of noise in the audio)

逐步判别分析的主函数 用于分类 [sel,c,c0,re,P]=StepclaassV2(data,[50 50 50],data,10) 输入变量 x为训练集.每行为一个样本，每列为一个变量. Class_x,为训练集的分类情况，一行，为各类样本数量,例如[5 6 9] Test为待分类样本. 输出变量：sel为选择的变量序号，c,c0为拟合出的判别函数.re为对Test的判别结果,P为其后验概率. author 王新 2012.4.8 例子Sample: load fisheriris x=meas Class_x=[50,50,50] Test=meas(1:5,:) Fa1=4 [sel,c,c0,re,P]=StepclaassV2(x,Class_x,Test,Fa1)( Stepwise discriminant analysis for the classification of the main function [sel, c, c0, re, P] = StepclaassV2 (data, [50 50 50], data, 10) input variables x as the training set, each conduct a sample of each as a variable. Class_x, the classification for the training set, a row, the number of samples for all types, for example, [5 6 9] Test sample to be classified. output variable: sel for the selected variables serial number, c, c0 is fitted discriminant function. re discrimination results for the Test, P its posterior probability. author Wang 2012.4.8 examples Sample: load fisheriris x = meas Class_x = [50,50,50] Test = meas (1:5, :) Fa1 = 4 [sel, c, c0, re, P] = StepclaassV2 (x, Class_x, Test, Fa1))

此代码主要是用于在MATLAB中的图像叠加，现实中主要应用与全景图片。(this code is used mainly in MATLAB superimposed images of the reality that the main application and panoramic images.)

说明：  利用mean5—3平滑对曲线进行拟合，可用于摩擦系数曲线，非线性数据进行拟合(Fit the curve with mean5-3 smoothing, which can be used for friction coefficient curve and nonlinear data for fitting)

说明：  有关于系统呼损率丢包率随新呼叫接入门限的变化规律，用于判定新呼叫的介入指标(Call loss rate on the system packet loss rate with the new call admission threshold variation, used to determine the involvement of the new call indicator)

局部线性嵌入算法，由laurance等人于2000年science上发表，广泛应用于图像处理模式识别等领域，是一种流形学习算法。(the algorithm of locally linear embedding)