assemble

NIPALS 算法 留1法 交叉检验 绘图 (NIPALS algorithm cross-validation method to stay one mapping)

说明：  基于matlab的K均值聚类程序。其中用IRIS数据进行验证，得到了很好的结果。文件中包含了演示后的结果图(Matlab-based K-means clustering procedure. Which use IRIS data verification, have had good results. File contains the results of the demonstration plan)

视觉注意力重要文献，基于该文献Dirk Walthera等人编制了可以在MATLAB环境下运行的图像显著区提取工具(Modeling attention to salient proto-objects)

神经网络进行语音识别：/enframe.m，该函数将输入向量分为固定长度固定重叠量的帧。/SampleCreate.m,将取所有音频的mfcc系数处理成神经网络函数所需的输入格式(Neural networks for speech recognition)

用数值方法实现该fsk信号的相关型解调器，完成一个二进制fsk通信系统的仿真(Numerical methods used in the implementation of the relevant type fsk signal demodulator, to complete a binary fsk communication system simulation)

说明：  PSO算法，通用性较强，只要换一下fitness就可以了(PSO algorithm, universal strong, as long as a change in fitness can be a)

MATLAB采用LMS,MMSE,RLS,SINR,SMI算法实现波束形成(MATLAB uses LMS, MMSE, RLS, SINR and SMI algorithm to realize beamforming.)

这是一道很基本的程序，一方面希望它对初学遗传算法的兄弟们有用（包含了遗传算法实现的大部分步骤，而且有详细的注释），另一方面旨在抛砖引玉， 在下浅见： 1、遗传算法在进行大规模计算时，数据交换量大，速度确实是一个瓶颈，就如这道程序，在种群规模50，进化代数1000时运行需要40秒左右(cpu:duron 1G),当然，小弟编程能力弱，望各位大侠指教啊，或帮我修改一下这道程序 (This is a very basic procedure, While they want it beginner GA brothers useful (including the genetic algorithm to achieve the most steps and a detailed note), on the other hand aims to attract the next suggestions : 1. GA conducting large-scale calculations, the data exchange capacity, speed is a bottleneck, as this practice. 50 in population size, the evolution of algebra required for the operation of 1000 when about 40 seconds (cpu : duron 1G), Of course, the younger programming weak, and hopes to enlighten ah Shanhaiguan, or help me rephrase this practice)

电力系统分析软件，可对特高压系统电压控制域进行分析(Power system analysis software, can UHV analysis system voltage control domain)

MIMO—ofdM无线通信仿真源代码，有利于加深对MIMO各种检测算法的理解(MIMO-ofdM wireless communication simulation source code, is conducive to deepening the understanding of MIMO detection algorithms of various)