da_lsgo.m

王小平《遗传算法——理论、应用与软件实现》随书光盘

%
% da_lsgo.m
%
% Runs the least squares regression routines
%

%
% If there is more than one output variable defined then
% only use the FIRST one.
%
if sum(output_var) > 1
	[D L]=size(output_var);
	for i = 1 : L
		if output_var(i) == 1
			y=data(:,i);
			break;
		end
	end
else
	y=data(:,output_var);
end
x=pcincl(data,include_var,output_var);

% y=data(:,output_var);

mx=mean(x);
sdx=std(x);
my=mean(y);
sdy=std(y);


%
% Split the data correctly
%

%
% Check radio button 4 for the percentage partition
%

if get(pcr_rad4,'Value') == 1
	[tx ty vx vy]=psplit(x,y,get(pcr_rad4sl,'Value'));
end

%
% Check radio button 5 for the random partition
%

if get(pcr_rad5,'Value') == 1
	disp('Bollocks')
	[tx ty vx vy]=rsplit(x,y,get(pcr_rad5sl,'Value'));
end

%
% Check radio button 6 for the block splitting
%

if get(pcr_rad6,'Value') == 1
	[tx ty vx vy]=bsplit(x,y,ceil(get(pcr_rad6sl,'Value')));
end

%
% Check radio button 7 for alternate split
%

if get(pcr_rad7,'Value') == 1
	[tx ty vx vy]=asplit(x,y);
end


%
% Bog standard regression with no polynomials
% and no best combination regression.
%

if get(pcr_rad1,'Value') == 0
	if get(pcr_rad3,'Value') == 0
		%
		% Calculate the regression coefficients
		% based on the training set
		%

		a=inv(tx'*tx)*tx'*ty;
		impact=a;
		regv=a;
		py=vx*a;
		[D L]=size(vy);
		e=((sse(vy,py).^0.5)/D)/mean(vy);
		c=corrcoef(vy,py);
		c=c(2,1);
	end
end


%
% Include polynomial terms.
% NO best combination regression
% get the number of terms to include from sl1
%
if get(pcr_rad1,'Value') == 1
	if get(pcr_rad3,'Value') == 0
		degree=ceil(get(sl1,'Value'));
		tx=pol(tx,degree);
		vx=pol(vx,degree);
		a=inv(tx'*tx)*tx'*ty;
		impact=a;
		regv=a;
		py=vx*a;
		[D L]=size(vy);
		e=((sse(vy,py).^0.5)/D)/mean(vy);
		c=corrcoef(vy,py);
		c=c(2,1);
	end
end

%
% Now check radio button 3 to regress best combination of variables
%

if get(pcr_rad3,'Value') == 1
	[D L]=size(tx);
	%
	% Only carry out this option if there are 10 varibles or less
	%
	if L < 11
	e_min=1e100;
	axes(ax1);
	ovt=zeros(1,L);
	for i = 1:2^L
		%
		% Build a binary include vector
		%
		inc=binary(i,L);

		tx1=pcincl(tx,inc,ovt);
		vx1=pcincl(vx,inc,ovt);

		%
		% Perform regression
		%
		a=inv(tx1'*tx1)*tx1'*ty;
		py=vx1*a;
		e=sse(py,vy);

		if e < e_min
			e_min=e;
			ba=a;
			include=inc;
			plot([py vy]);
			drawnow;
			bpy=py;
		end
	end
	end
	[D L]=size(vy);
	e=((sse(vy,py).^0.5)/D)/mean(vy);
	c=corrcoef(vy,py);
	c=c(2,1);

	%
	% Calculate the impact vector
	%
	[D L]=size(vx);
	impact=zeros(L,1);
	count=1;
	for i = 1 : L
		if include(i) == 1
			impact(i)=ba(count);
			count=count+1;
		end
	end

end


%
% Print results in the little blue box
%
text1=uicontrol(w1,...
	'Style','text',...
	'Position',[275 385 130 15],...
	'BackGroundColor',[0 0 1],...
	'ForeGroundColor',[1 1 1],...
	'String',['Error =  ' num2str(e)]);

text2=uicontrol(w1,...
	'Style','text',...
	'Position',[410 385 130 15],...
	'BackGroundColor',[0 0 1],...
	'ForeGroundColor',[1 1 1],...
	'String',['Correlation =  ' num2str(c)]);

%
% More detail button
%
pca_but2=uicontrol(w1,...
	'Style','push',...
	'Position',[380 10 100 20],...
	'String','More detail',...
	'CallBack','da_lsdet;');

%
% Save model button
%
pca_but3=uicontrol(w1,...
	'style','push',...
	'position',[500 10 100 20],...
	'string','Save model',...
	'callback',[...
		'[D L]=size(tx);',...
		'out=svlin2('''',include_var,zeros(L,1),zeros(L,1),zeros(L,1),zeros(L,1),impact,ones(L,1),zeros(L,1))']);


clear x y;