NonlControl

时间系列非线线性处理工具包，可以计算多数非线性参数，例如：李亚谱诺夫指数，相关维等(nonlinear parameter toolkit)

说明：  对离散数据运用最小二乘法进行椭圆拟合，运行过了觉得还不错就上传了。(Discrete data on the use of least squares method for ellipse fitting, run over a feel good on the uploaded.)

Matlab源程序：MMSE Receiver for DS-SS in AWGN Channel！希望对大家有帮助。(Matlab source : MMSE Receiver for DS-SS in AWGN Channel! We want to help.)

（压缩包里一共有5个代码） pca+lda+粗糙集+模糊神经网络 saveORLimage.m将ORL人脸库分为测试集ptest和训练集pstudy存为imagedata.mat 1.savelda.m将人脸库先进行pca降维，再用lda进行特征提取，得到新的测试集ldatest和训练集ldastudy存为imageldadata.mat 2.对ldastudy进行离散化(discretimage.m)，得到离散化矩阵disdata,存入到imagedisdata.mat 3.将disdata组成决策表(savers.m)，通过对disdata的条件属性进行约简，得到其一个约简，组成新的测试集rstest和训练集rsstudy存为imagersdata.mat 4.对rsstudy进行模糊神经网络训练(savecul.m)，对模糊神经网络的参数进行调整学习将其存入culdata.mat 5.用runfnn.m对rstest进行测试得到其识别率 savem.m和cm.m是用最小距离分类器对训练集和测试集进行分类.(pca+ lda+ Rough Set+ fuzzy neural network saveORLimage.m will ORL face database is divided into test set and training set ptest for pstudy keep imagedata.mat Treasury will face 1.savelda.m first dimensionality reduction pca, lda used feature extraction, a new test set and training set ldatest for ldastudy keep imageldadata.mat 2. Ldastudy carried out on the discretization (discretimage.m), to be discrete matrix of disdata, deposited to imagedisdata.mat 3. Disdata the composition of the decision table (savers.m), the conditions on the attributes disdata about Jane, has been one of its reduction to form the new test set and training set rstest for rsstudy keep imagersdata.mat 4. Rsstudy training fuzzy neural network (savecul.m), on the parameters of fuzzy neural network to learn to adjust their deposit culdata.mat 5. Rstest used to test for runfnn.m by its recognition rate cm.m is savem.m and minimum distance classifier on the training set and test set classificati)

Multirate Filtering for Digital Signal Processing MATLAB Applications

通过应用于小波变换到lena.jpg图像，简单实现图像压缩，实现一级压缩和二级压缩(Through applied wavelet transform to lena.jpg image, a simple image compression, compression and two compression to achieve a)

基于MATLAB实现的说话人识别程序，分别用bp、pnn、som、rbf、lvq等算法，对语音文件进行训练和测试，效果不错。~..~ 下面说明一下bprengong程序： 数据分别用来训练和测试两部分。 具体程序分为两部分，第一部分为：计算识别模型 变量v是mfcc处理以后的矢量。因为数据可能长短不一，所以放在同一进行截取。p的每一行代表一个语音数据（共15个）。变量Pr为每一行的最大最小值。变量T为目标值。输出神经元个数为15。 在训练阶段，如果用于训练的输入训练样本的类别标号为i（即语音数据的标号），则训练时设第i个节点的期望输出设为1。其余节点期望输出均为0。 在识别时，当一个未知类别的样本作用到输入端时，考查各输出节点的输出，并将这个样本的类别判定为输出值最大的那个节点对应的类别。 (err)

说明：  matlab中完整实现分水岭算法的分割，效果很理想(matlab full implementation in the watershed segmentation algorithm, the effect is very satisfactory)

libsvm.rar 为机器学习的开源库matlab语言(libsvm.rar open source library for machine learning matlab language)

Pump speed controller IR2130