急!! 请大家帮我调调这个程序

2-31 · 发表于 2006-5-24 11:47

马上注册，结交更多好友，享用更多功能，让你轻松玩转社区。

您需要登录才可以下载或查看，没有账号？我要加入

x

请各位帮忙调试这段程序 数据: train_features 
<TABLE cellSpacing=0 cellPadding=0 width=288 border=0 x:str>
<COLGROUP>
<COL style="WIDTH: 54pt" span=4 width=72>

<TR height=19>
<TD align=right width=72 height=19 x:num>1</TD>
<TD align=right width=72 x:num>1</TD>
<TD align=right width=72 x:num>1</TD>
<TD align=right width=72 x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>3</TD>
<TD align=right x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>3</TD>
<TD align=right x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD>
<TD align=right x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD>
<TD align=right x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>2</TD></TR></TABLE>targets 
<TABLE cellSpacing=0 cellPadding=0 width=72 border=0 x:str>
<COLGROUP>
<COL style="WIDTH: 54pt" width=72>

<TR height=19>
<TD align=right width=72 height=19 x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD></TR></TABLE>prediction_features 
<TABLE cellSpacing=0 cellPadding=0 width=288 border=0 x:str>
<COLGROUP>
<COL style="WIDTH: 54pt" span=4 width=72>

<TR height=19>
<TD align=right width=72 height=19 x:num>3</TD>
<TD align=right width=72 x:num>3</TD>
<TD align=right width=72 x:num>2</TD>
<TD align=right width=72 x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>2</TD>
<TD align=right x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>2</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>1</TD>
<TD align=right x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD></TR>
<TR height=19>
<TD align=right height=19 x:num>3</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>2</TD>
<TD align=right x:num>1</TD></TR></TABLE>
程序: function D=dan(train_features,train_targets,prediction_features)
%输入:train_features,train_targets,prediction_features %输出:D [Ni,M]=size(train_features); Nbins=length(unique(train_features)); tree=build_tree(train_features, train_targets, Nbins); %使用树 targets=use_tree(prediction_features, 1:size(prediction_features,1), tree, Nbins, unique(train_targets));
D=reshape(targets,[],1); %END
function targets=use_tree(features, indices, tree, Nbins, Uc) %使用树：如果模型已经得到，对于预测集里的每一条记录，从树的根结点开始，向下找出该记录所属的类 targets=zeros(1,size(features,1)); if (size(features,2)==1), % for i=1:Nbins, in=indices(find(features(indices)==i)); if ~isempty(in), if isfinite(tree.child(i)), targets(in)=tree.child(i); else n=1+floor(rand(1)*length(Uc)); targets(in)=Uc(n); end end end return end dim=tree.split_dim; dims=find(~ismember(1:size(features,2),dim)); %从树的根结点开始,向下找出记录所属的类 for i=1:Nbins, in=indices(find(features(indices,dim)==i)); targets=targets+use_tree(features(:,dims),in,tree.child(i),Nbins,Uc); end %END use_tree
function tree=build_tree(features,targets,Nbins)
[Ni,L]=size(features); Uc=unique(targets); U=length(unique(targets)); if (length(unique(targets))==1) tree.split_dim=0; tree.child=targets(U); %??? return end if (L == 1), for a = 1:Nbins, tree.split_dim = 0; indices = find(features == a); if ~isempty(indices), if (length(unique(targets(indices))) == 1), tree.child(a) = targets(indices(1)); else H = hist(targets(indices), Uc); [m, T] = max(H); tree.child(a) = Uc(T); end else tree.child(a) = inf; end end return end
for a=1:U, Pnode(a)=length(find(targets==a))/Ni; end Inode=-sum(Pnode.*log(Pnode)/log(2)); 
delta_Ib=zeros(1,L); P=zeros(1,U); for a=1:L, for k=1:Nbins, f=find(features(:,a)==k); V(k)=length(f); Y(k)=V(k)/Ni; if (V(k)~=0), for b=1:U, s=length(find(targets(f)==b)); P(b)=s/V(k); %P(s/v) end end Q=sum(-P.*log(eps+P)/log(2)); E(k)=Y(k).*Q; end info=sum(E); delta_Ib(a)=Inode-info; %H(U)-H(U|V) end
 [m,dim]=max(delta_Ib); tree.split_dim=dim; dims=find(~ismember(1:L,dim)); for a=1:Nbins, indices=find(features(:,dim) ==a); % if (~isempty(indices)), tree.child(a)=build_tree(features(indices,dims),targets(indices),Nbins); end end [em06]

lxq · 发表于 2006-5-24 14:00

[Ni,M]=size(train_features); ??? Input argument "train_features" is undefined. 没定义变量名?

[此贴子已经被作者于2006-5-24 14:01:14编辑过]

2-31 · 发表于 2006-5-24 14:31

定义了呀,train_features就是上面给的那个数据集
在命令窗口输入:

dan(train_features,train_targets,prediction_features)

提示
??? Subscript indices must either be real positive integers or logicals.
Error in ==> D:\matlab\work\dan.m (use_tree)
On line 38 ==> in=indices(find(features(indices,dim)==i));
Error in ==> D:\matlab\work\dan.m (use_tree)
On line 39 ==> targets=targets+use_tree(features(:,dims),in,tree.child(i),Nbins,Uc);
Error in ==> D:\matlab\work\dan.m
On line 12 ==> targets=use_tree(prediction_features, 1:size(prediction_features,1), tree, Nbins, unique(train_targets));

怎么处理啊??
循环变量和向量索引值都没有给负值,为什么还会出现
??? Subscript indices must either be real positive integers or logicals.

[ 本帖最后由 ChaChing 于 2009-11-3 15:10 编辑 ]

账号		自动登录	找回密码
密码			我要加入