SRM_DITC

基于Wiener滤波的语音消噪Matlab源程序 调用方式： output=WienerScalart96(signal,fs,IS) signal－－输入含噪语音 fs－－－－语音信号采样率 IS－－－－起始静音端，用来估计噪声方差（秒做单位，默认是0.25秒） (Wiener filter-based de-noising of the voice call Matlab source code the way: output = WienerScalart96 (signal, fs, IS) signal- voice input with noise fs---- voice signal sampling rate IS---- Start mute end used to estimate the noise variance (units of seconds to do, the default is 0.25 seconds))

本程序模拟了无线通信中常用的AWGN，瑞利信道模型！！！！(This procedure simulates the wireless communications commonly used in AWGN, Rayleigh channel model! ! ! !)

MEL 求解GMM模型的参数matlab环境下的源代码，和大家分享(MEL Solving GMM model parameters Matlab environment of the source code and share)

Optimal output feedback controller

模式识别课程中的主成分分析方法PCA的matlab实现源码(Pattern Recognition in the principal component analysis PCA matlab to achieve source)

基于改进高斯模型的车流量检测算法，用MATLAB编写(Adaptive vehicle detection and traffic statistics algorithm based on improved Gaussian model)

MATLAB实现的核估计后验概率算法，输入训练集和窗函数带宽，得到坐标的概率估计(matlab ptrain)

MATLAB与外部程序接口，直接调用matlab的算法很方便，推荐使用(MATLAB interface with external programs)

使用matlab对基于高斯模型下的获取幅度估计的语音增强算法进行的仿真，效果较维纳算法和谱减算法有一定的改善。(Use matlab Gaussian model based on the estimated rate of access to speech enhancement algorithm for the simulation, the effect is less than the Wiener algorithm and the spectral algorithm has some improvement.)

这是用身高体重数据进行性别分类的实验。 用最小错误率贝叶斯分类器决策时，首先通过比较概率大小判断一个体重身高二维向量代表的人是男是女，然后再逐一与已知性别的数据比较，就可以得到错误率的统计。然后改变先验概率，重复上面的过程，观察数据结果的变化。 用最小风险贝叶斯分类器决策时，首先求出用最小错误率贝叶斯分类器得到的条件概率；然后根据人为给定的决策表，根据公式算出条件风险；然后逐一比较条件风险，找出使条件风险最小的决策（也就是分类）。最后用分类得到的结果逐一比较已经知道的原始数据，统计处错误率。 (This is the height and weight data for gender classification experiment. With the minimum error rate Bayesian classifier decisions , first by comparing the probability of the size and weight to height to determine a person represented by two-dimensional vector is male or female , and then one by one with known gender data comparison, the statistical error rate can be . Then change the prior probability , repeat the above process , the results of the changes observed data . Bayesian classifier with the minimum risk decision-making , first find the minimum error rate using Bayesian classifier to get the conditional probability then artificially given decision table , according to the formula to calculate conditional risk and then one by one more conditional risk , to find ambassador to the conditions of minimum risk decision making (ie classification) . Finally, the results obtained with the classification by-side comparison of the raw data have been aware of SD error rate . )