Kalman-PROGAM

一个非常好的粒子群算法实现，该算法为matlab实现(A very good PSO implementation, the algorithm for the matlab implementation)

说明：  多元线性回归的MatLAB代码，在MatLAB7.0运行成功！(Multiple linear regression MatLAB code, MatLAB7.0 runs successfully!)

lms source code for adaptive filtering the acoustic noise.

Implement loss? part of JPEG encoder for grayscale images on the first implement grayscale image quantization

GMM，LDA,PCA,GMM，LDA,PCAGMM，LDA,PCA(the code of GMM，LDA,PCA)

舵机扫频辩识，鲁棒控制器，加载系统流量阀，双阀扫频(The sweep frequency identification, robust controller, loading system flow valve, double valve sweep )

在小波域中，以图像作为水印信息，嵌入以声音为载体信号的源程序。(n wavelet domain to image as a watermark, embedded with sound as the carrier signal source program. )

软件程序按照发射端所掌握的各用户信道状态信息的程度共分为两部分：即完整信道状态信息(CSIT)和部分信道状态信息(CSIP)。其中，每一部分都包括预编码(precoding)和用户调度(scheduling) 在CSIT中，precoding又按照各用户的数据流数分为单数据流和多数据流两种情况。在每种情况下，首先考察了不同预编码算法的性能表现，包括两种ZF、MMSE、SINR、SLNR。之后又考察了功率分配算法的性能表现（文件名中含有PD表明其含有功率分配的过程）。按照不同指标进行功率分配的，在文件名中进行了区分，如PD_CN代表以信道范数为参考指标进行功率分配。Scheduling部分首先观察了RoundRobin、MaxH和MMSLNR三种算法的性能对比。之后在Kc和Round部分分别观察了不同预选用户数和不同最大替换轮数下MMSLNR算法的表现。 在CSIP中，只对各用户单数据流的情况进行了仿真。采用的预编码算法主要有DSLNR（即直接运用CSIT下的预编码算法）、ESLNR（即对SLNR进行均值计算的，在CSIP中，引入均值计算的与SLNR有关的算法，其文件名中都有modified以示区别）、EMMSE（即陈明老师那边的那篇文章中的预编码算法）。Scheduling中也只是简单的观察了RoundRobin、MaxH、DMMSLNR和EMMSLNR（前者没有均值计算，后者有）的性能对比。(The degree of channel state information for each user software program in accordance with the master transmitter consists of two parts : the complete channel state information (CSIT) and partial channel state information (CSIP). Wherein , each of which comprises a pre- coding (precoding) and the user scheduling (scheduling) In CSIT in , precoding and each user according to the number of data streams into a single data stream and multiple data streams in both cases . In each case , first examine the performance of different pre- coding algorithm, including two ZF, MMSE, SINR, SLNR. After they inspected the power allocation algorithm performance ( file name contains PD showed that it contained power allocation process ) . According to different indicators for power distribution , make a distinction in the file name , such as PD_CN representative channel norm-referenced index for power distribution . Scheduling section first observed the performance comparison RoundRobin, MaxH and MMSL)

传统的短时谱估计语音增强算法通常假设语音谱分量相互独立，没有考虑语音谱分量间的相关性。针对这 一问题，该文提出一种新的基于多元Laplace分布模型的短时谱估计算法。首先，假设语音的离散余弦变换(DCT) 系数服从多元Laplace分布，以此利用谱分量间的相关性；在此基础上，利用多元随机矢量的高斯尺度混合模型表 示，推导得到语音DCT系数矢量的最小均方误差(MMSE)估计的解析表达式；并进一步推导了基于该分布模型的 语音存在概率，对最小均方误差估计子进行修正。实验结果表明，该算法在抑制背景噪声和减少语音失真等方面优 于传统的语音增强方法。(The spectral components of speech are usually assumed to be independent in traditional short-time spectrum estimation, which is not the case in practice. Tosolve this problem, a new speech enhancement algorithm with multivariate Laplace speech model is proposed in this paper. Firstly, the speech Discrete Cosine Transform (DCT) coefficients are modeled by a multivariate Laplace distribution, so the correlations between speech spectral components can be exploited. And then a Minimum-Mean-Square-Error (MMSE) estimator based on the proposed model is derived using a Gaussian scale mixture representation of random vectors. Furthermore, the speech presence uncertainty with the new model is derived to modify the MMSE estimator. Experimental results show that the developed method has better noise suppression performance and lower speech distortion compared to the traditional speech enhancement method. )

基于matlab的网络神经源程序，用法简单，详细见打包下的readme (Matlab neural network based on source, usage is simple, in detail, see the readme packaged under)