📄 数据仓库与数据挖掘--数据挖掘部分算法的matlab实现 c4_5.htm
字号:
<TABLE width="95%" border=0>
<TBODY>
<TR>
<TD><STRONG>我的分类（专题）</STRONG></TD></TR>
<TR>
<TD>
<DIV align=center><A
href="http://blogger.org.cn/blog/blog.asp?name=xueflhg">首页(85)</A><BR><A
href="http://blogger.org.cn/blog/blog.asp?subjectid=845&name=xueflhg">金融行业(4)</A><BR><A
href="http://blogger.org.cn/blog/blog.asp?subjectid=846&name=xueflhg">反洗钱专栏(9)</A><BR><A
href="http://blogger.org.cn/blog/blog.asp?subjectid=843&name=xueflhg">数据挖掘(11)</A><BR><A
href="http://blogger.org.cn/blog/blog.asp?subjectid=844&name=xueflhg">电信行业(9)</A><BR><A
href="http://blogger.org.cn/blog/blog.asp?subjectid=847&name=xueflhg">综合(19)</A><BR><A
href="http://blogger.org.cn/blog/blog.asp?subjectid=842&name=xueflhg">数据仓库(27)</A><BR></DIV></TD></TR></TBODY></TABLE><BR>
<TABLE width="95%" border=0>
<TBODY>
<TR>
<TD><STRONG>日志更新</STRONG></TD></TR>
<TR>
<TD>
<DIV align=left><A title=xueflhg发表于2006-7-6
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=16337"
15:41:02>数据仓库是干什么的，到现在，我终于看</A><BR><A title=xueflhg发表于2006-7-6
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=16335"
15:20:05>几个非常经典的对“数据仓库”的解释</A><BR><A title=xueflhg发表于2006-7-4
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=16260"
14:33:36>好久好久没有来管理了</A><BR><A title=xueflhg发表于2006-2-8
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=11570"
11:25:04>商业银行反洗钱的法律与实务分析</A><BR><A title=xueflhg发表于2006-1-16
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=11223"
10:42:17>UNIX系统操作命令</A><BR><A title=xueflhg发表于2005-12-31
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=10920"
11:06:00>AIX捉虫记之__invscoutd</A><BR><A
title=xueflhg发表于2005-12-26
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=10797"
13:00:44>AIX基础教程</A><BR><A title=xueflhg发表于2005-12-26
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=10796"
12:52:49>AIX文件系统性能调优</A><BR><A title=xueflhg发表于2005-12-26
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=10795"
11:17:11>[命令] xargs命令</A><BR><A title=xueflhg发表于2005-12-26
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=10793"
11:08:35>一个AIX操作系统中的用户信息拷贝到</A><BR></DIV></TD></TR></TBODY></TABLE><BR>
<TABLE width="95%" border=0>
<TBODY>
<TR>
<TD><STRONG>最新评论</STRONG></TD></TR>
<TR>
<TD>
<DIV align=left><A title=HJ(游客)发表评论于2007-7-26
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=16337#56247"
15:46:36>回复:数据仓库是干什么的，到现在，我终于</A><BR><A
title=薛峰(游客)发表评论于2007-7-20
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=11223#55976"
21:27:33>回复:UNIX系统操作命令</A><BR><A title=薛峰(游客)发表评论于2007-7-20
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=11223#55975"
21:25:57>回复:UNIX系统操作命令</A><BR><A title=hehe(游客)发表评论于2007-7-13
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=6198#55613"
13:15:12>回复:Essbase VS Cognos</A><BR><A
title=dagou(游客)发表评论于2007-4-26
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=16337#49519"
14:01:32>回复:数据仓库是干什么的，到现在，我终于</A><BR><A
title=fxwang(游客)发表评论于2007-4-25
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=16337#49472"
15:09:50>回复:数据仓库是干什么的，到现在，我终于</A><BR><A
title=111(游客)发表评论于2007-3-25
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=6839#47724"
14:55:53>回复:数据挖掘部分算法的matlab实现</A><BR><A
title=zlf(游客)发表评论于2007-3-24
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=4312#47684"
14:52:41>回复:一些编程经验，与大家共享！</A><BR><A
title=qzy2004(游客)发表评论于2007-3-9
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=5574#47129"
11:56:04>回复:Web数据挖掘的研究现状及发展</A><BR><A
title=LP(游客)发表评论于2007-1-22
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=4312#45988"
10:41:25>回复:一些编程经验，与大家共享！</A><BR></DIV></TD></TR></TBODY></TABLE><BR>
<TABLE width="95%" border=0>
<TBODY>
<TR>
<TD><STRONG>留言板</STRONG></TD></TR>
<TR>
<TD>
<DIV align=left>
<DIV align=left><A
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#message">签写新留言</A><BR><BR><A
title=阿娅(游客)发表于2006-9-6
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#13518"
14:30:59>我也想要ID3的源码 C++的,yang</A><BR><A
title=阿娅(游客)发表于2006-9-6
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#13517"
14:30:45>我也想要ID3的源码 C++的,yang</A><BR><A
title=panwei (游客)发表于2006-5-18
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#4417"
15:58:17>ID3算法的VC++源程序 </A><BR><A
title=panwei (游客)发表于2006-5-18
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#4416"
15:58:11>ID3算法的VC++源程序 </A><BR><A title=尕光(游客)发表于2006-2-21
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#2314"
9:46:31>同行</A><BR><A title=haoyanyou(游客)发表于2006-2-15
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#2271"
13:57:16>好地方</A><BR><A title=xueflhg发表于2005-8-18
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#1009"
18:18:26>忙 啊 </A><BR><A title=idmer发表于2005-6-26
href="http://blogger.org.cn/blog/message.asp?name=xueflhg#344"
17:41:14>呵呵，碰到同行了</A><BR></DIV></DIV></TD></TR></TBODY></TABLE><BR>
<TABLE height=36 width="95%" border=0>
<TBODY>
<TR>
<TD><STRONG>链接</STRONG></TD></TR>
<TR>
<TD height=15>
<DIV align=left><BR><A
href="http://www.unixblog.net/index.php?blog=6">AIX快活如意斋</A>
</DIV>
<DIV align=left><A
href="http://bidwhome.itpub.net/">BIDW之家</A></DIV></TD></TR></TBODY></TABLE><BR>
<TABLE width="95%" border=0>
<TBODY>
<TR>
<TD><STRONG>Blog信息</STRONG></TD></TR>
<TR>
<TD>
<DIV
align=left>blog名称:数据仓库与数据挖掘<BR>日志总数:85<BR>评论数量:121<BR>留言数量:8<BR>访问次数:213297<BR>建立时间:2005年3月17日
</DIV></TD></TR></TBODY></TABLE>
<P align=center><A
href="http://blogger.org.cn/blog/rss2.asp?name=xueflhg"
target=_blank><IMG height=14
src="数据仓库与数据挖掘--数据挖掘部分算法的matlab实现 C4_5.files/xml.gif" width=36
border=0></A><BR></P>
<DIV align=center></DIV>
<DIV align=center><BR></DIV></TD></TR></TBODY></TABLE></TD>
<TD vAlign=top align=middle><BR>
<DIV id=adsense1
style="DISPLAY: block; FLOAT: none; TEXT-ALIGN: left"></DIV>
<STYLE type=text/css>A.categorylink:link {
COLOR: #999999
}
A.categorylink:visited {
COLOR: #999999
}
A.categorylink:active {
COLOR: #999999
}
A.categorylink:hover {
COLOR: #ff9900
}
</STYLE>
<TABLE style="TABLE-LAYOUT: fixed; WORD-BREAK: break-all" cellSpacing=1
cellPadding=3 width="98%" bgColor=#cccccc border=0>
<TBODY>
<TR bgColor=#f8f8f8>
<TD>
<P><FONT size=4><STRONG><IMG
src="数据仓库与数据挖掘--数据挖掘部分算法的matlab实现 C4_5.files/1.gif"><A
href="http://blogger.org.cn/blog/blog.asp?name=xueflhg&subjectid=843">[数据挖掘]<A
href="http://blogger.org.cn/blog/more.asp?name=xueflhg&id=6839">数据挖掘部分算法的matlab实现
C4_5</A></STRONG></FONT><BR><A class=categorylink
href="http://blogger.org.cn/blog/list.asp?classid=46"
target=_blank>网上资源</A>, <A class=categorylink
href="http://blogger.org.cn/blog/list.asp?classid=4"
target=_blank>随笔</A></P>
<P><A href="http://blogger.org.cn/blog/blog.asp?name=xueflhg"
target=_blank>薛 峰</A> 发表于 2005-6-27 14:21:09 </P></TD></TR>
<TR bgColor=#ffffff>
<TD height=0>
<TABLE cellSpacing=0 cellPadding=0 width="100%" border=0>
<TBODY>
<TR>
<TD> function D = C4_5(train_features, train_targets,
inc_node, region)<BR><BR>% Classify using Quinlan´s C4.5
algorithm<BR>% Inputs:<BR>% features - Train features<BR>%
targets - Train targets<BR>%
inc_node - Percentage of incorrectly
assigned samples at a node<BR>% region
- Decision region vector: [-x x -y y
number_of_points]<BR>%<BR>% Outputs<BR>% D - Decision
sufrace<BR><BR>%NOTE: In this implementation it is assumed
that a feature vector with fewer than 10 unique values (the
parameter Nu)<BR>%is discrete, and will be treated as such.
Other vectors will be treated as continuous<BR><BR>[Ni, M] =
size(train_features);<BR>inc_node =
inc_node*M/100;<BR>Nu =
10;<BR><BR>%For the decision
region<BR>N
=
region(5);<BR>mx =
ones(N,1) * linspace
(region(1),region(2),N);<BR>my =
linspace (region(3),region(4),N)´ *
ones(1,N);<BR>flatxy =
[mx(:), my(:)]´;<BR><BR>%Preprocessing<BR>%[f, t, UW,
m] = PCA(train_features,
train_targets, Ni, region);<BR>%train_features = UW
* (train_features -
m*ones(1,M));;<BR>%flatxy =
UW * (flatxy - m*ones(1,N^2));;<BR><BR>%Find which of the
input features are discrete, and discretisize the
corresponding<BR>%dimension on the decision
region<BR>discrete_dim = zeros(1,Ni);<BR>for i =
1:Ni,<BR> Nb =
length(unique(train_features(i,:)));<BR> if (Nb
<= Nu),<BR> %This is a
discrete
feature<BR> discrete_dim(i)
= Nb;<BR> [H, flatxy(i,:)]
= high_histogram(flatxy(i,:), Nb);<BR>
end<BR>end<BR><BR>%Build the tree
recursively<BR>disp(´Building
tree´)<BR>tree =
make_tree(train_features, train_targets, inc_node,
discrete_dim, max(discrete_dim), 0);<BR><BR>%Make the decision
region according to the tree<BR>disp(´Building decision
surface using the tree´)<BR>targets = use_tree(flatxy, 1:N^2,
tree, discrete_dim,
unique(train_targets));<BR><BR>D =
reshape(targets,N,N);<BR>%END<BR><BR>function targets =
use_tree(features, indices, tree, discrete_dim,
Uc)<BR>%Classify recursively using a tree<BR><BR>targets =
zeros(1, size(features,2));<BR><BR>if (tree.dim ==
0)<BR> %Reached the end of the
tree<BR> targets(indices) =
tree.child;<BR>
break<BR>end<BR> <BR>%This
is not the last level of the tree, so:<BR>%First, find the
dimension we are to work on<BR>dim = tree.dim;<BR>dims=
1:size(features,1);<BR><BR>%And classify according to it<BR>if
(discrete_dim(dim) == 0),<BR> %Continuous
feature<BR> in = indices(find(features(dim,
indices) <= tree.split_loc));<BR> targets =
targets + use_tree(features(dims, :), in, tree.child(1),
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -