MATLAB,波形簇分类源码程序 - matlab算法设计

下载频道> 资源分类> matlab源码> 算法设计> MATLAB,波形簇分类源码程序

标题：MATLAB,波形簇分类源码程序
分享到：

所属分类: 算法设计资源类型：文件大小: 90.96 KB 上传时间: 2016-01-27 18:10:17 下载次数: 7 资源积分：1分提供者: 马云点击下载

内容：
MATLAB,波形簇分类源码程序，程序员在编程的过程中可以参考学习使用，希望对IT程序员有用，此源码程序简单易懂、方便阅读，有很好的学习价值！

% Demonstration of the Modified Dunn's cluster validity index.

clc;

fprintf(1,'Cluster validation using modified Dunn''s index.\n');

fprintf(1,'--------------------------------------------------------\n');

%% Load dataset.

dataName = 'Wave';

load(['datasets\',dataName,'.mat']);

[n,d]=size(data);

fprintf(1,'Dataset %s loaded!\nSamples: %d, features: %d.\n',dataName,n,d);

fprintf(1,'--------------------------------------------------------\n');

%% Cluster the data.

% Utilize hierarchical single-linkage algorithm with Euclidean distance.

% Number of clusters: from 2 to maxK.

% maxK = ceil(sqrt(n));

maxK = 9;

labels = zeros(n,maxK-1);

% Single-linkage clustering algorithm (from the Statistics toolbox).

fprintf(1,['Performing single-linkage clustering for k=2-',num2str(maxK),...

' [']);

D=pdist(data,'euclidean');

htree=linkage(D,'single');

for k = 2:maxK

% Store labels.

labels(:,k-1)=cluster(htree,k);

fprintf(1,'.');

end

fprintf(1,']\n');

% Plot clustering results.

fprintf(1,'Clustering results are shown in figure.\n');

fprintf(1,'--------------------------------------------------------\n');

options_SP.title=['Clusterings of ', dataName, ' with SL'];

options_SP.subtitle=num2cell(2:maxK);

options_SP.markerSize = 5;

scatterPlot(data,labels,options_SP);

%% Validate the clusterings with internal indices.

% Compare Modified Dunn's index to generalized version (Pal & Biswas, 1997)

% Construct a neighborhood graph on data points

% Choose type of graph: 'rng', 'gabriel', 'directedKnn', 'mutualKnn',

% 'symKnn', 'EMST' or 'epsilon'.

graph_type = 'gabriel';

isDirected = 0; % set to 1 when using directed graph, e.g. directedKnn

options.show = 0; % set it to 1 to enable plotting of a graph

fprintf(1,'Creating %s graph on data points ... ', graph_type);

[G,d,uniqueInd]=graph_create(data,[],graph_type,options);

if ~isempty(uniqueInd)

data = data(uniqueInd,:);

labels = labels(uniqueInd,:);

end

fprintf('done!\n');

fprintf(1,'Validating results with internal indices [');

DNg = zeros(1,maxK-1);

DNs = zeros(1,maxK-1);

for k = 1:maxK-1

% -> generalized Dunn's index (Pal & Biswas, 1997)

DNg(k) = indexDNg_graph(G,labels(:,k),data);

% -> modified Dunn's index (Ilc, 2012)

DNs(k) = indexDNs_graph(G,labels(:,k),isDirected);

% SLOWER: alternative calls, when graph G is not pre-computed

%DNs(k) = indexDNg(data,labels(:,k),graph_type,options);

%DNs(k) = indexDNs(data,labels(:,k),graph_type,options);

fprintf(1,'.');

end

fprintf(1,']\n');

% find maximum index value

[val_g,ind_g]=max(DNg);

[val_s,ind_s]=max(DNs);

figure();

subplot(2,1,1); plot(2:maxK, DNg,'.-k');hold on; plot(ind_g+1,val_g,'sr');

title('Generalized Dunn''s index');

xlabel('number of clusters'); ylabel('index value');

subplot(2,1,2); plot(2:maxK, DNs,'.-k');hold on; plot(ind_s+1,val_s,'sr');

title('Modified Dunn''s index');

xlabel('number of clusters'); ylabel('index value');

fprintf(1,'Values of validation indices are shown in figure.\n');

fprintf(1,'Optimal number of clusters is indicated with red square.\n');

fprintf(1,'--------------------------------------------------------\n');

文件列表（点击上边下载按钮，如果是垃圾文件请在下面评价差评或者投诉）:

关键词：波形 源码 程序

matlab源码下载排行

回到顶部

友情链接

matlab源码网站源码编程语言网络编程手机通讯编程 KUNAG 官网 KUNAG 电商更多

: 联系方式| 版权声明| 招聘信息| 广告服务| 银行汇款| 法律顾问| 兼职技术| 付款方式| 关于我们|; 网站客服程序员兼职招聘 411241577@qq.com; 沪ICP备19040327号-3; 公安备案号：沪公网安备 31011802003874号; 库纳格流体控制系统（上海）有限公司版权所有; Copyright © 1999-2014, GUSUCODE.COM, All Rights Reserved

关键词： 波形 源码 程序

相关推荐

matlab源码下载排行

友情链接

关键词：波形源码程序