svlin.m

王小平《遗传算法——理论、应用与软件实现》随书光盘

function [out]=svlin(fname,inc,maxx,minx,maxy,miny,cfs,ply,flt)
%
%
% Saves a neural network type model to file
%
% fname	= file name
% inc	= matrix of included variables
% maxx	= maximum x-values
% minx	= minimum x-values
% maxy	= maximum y-value
% miny	= minimum y-value
% cfs	= regression coefficients
% ply	= polynomial powers
% flt	= filter constants
%

%
% Open the file. If none is supplied, then put
% up a dialog box
%
if fname == []
	[df_name df_path]=uiputfile('*.mod','Please select a model file');
	if df_name ~= 0
		%
		% Change to the directory
		%
		[D L]=size(df_path);
		if L == 3
			if df_path ~= 'C:\'
				df_path=df_path(:,1:L-1);
			end
		else
			df_path=df_path(:,1:L-1);
		end
		eval(['cd ' num2str(df_path)]);
		fname=df_name;
	else
		fname=[];
	end
end

if fname ~= []

	%
	% Calculate various parameters
	%
	% *** Output MUST be first column!
	% *** MUST be only one output
	%
	[D L]=size(inc);
	inc=inc(:,2:L);
	num_inc=sum(inc);
	tot_var=L-1;

	f=fopen(fname,'w');
	%
	% Header
	%
	fprintf(f,['[' num2str(hostid) ']']);
	fprintf(f,'\n');
	fprintf(f,'\n');

	fprintf(f,'[Linear type 1 model]');
	fprintf(f,'\n');
	
	%
	% Total variables
	%
	fprintf(f,'[Total variables]');
	fprintf(f,'\n');
	fprintf(f,num2str(tot_var));
	fprintf(f,'\n');
	fprintf(f,'\n');
	
	%
	% Included variables
	%
	fprintf(f,'[Included variables]');
	fprintf(f,'\n');
	fprintf(f,num2str(num_inc));
	fprintf(f,'\n');
	fprintf(f,'\n');
	
	%
	% Minimum x-values for scaling
	%
	count=1;
	fprintf(f,'[Minimum values]');
	fprintf(f,'\n');
	for i=1:tot_var
		if inc(i)==1
			fprintf(f,num2str(minx(count)));
			fprintf(f,'\n');
			count=count+1;
		end
	end
	fprintf(f,'\n');

	%
	% Maximum x-values for scaling
	%
	count=1;
	fprintf(f,'[Maximum values]');
	fprintf(f,'\n');
	for i=1:tot_var
		if inc(i)==1
			fprintf(f,num2str(maxx(count)));
			fprintf(f,'\n');
			count=count+1;
		end
	end
	fprintf(f,'\n');
	
	%
	% Included variables
	%
	fprintf(f,'[Included variables]');
	fprintf(f,'\n');
	for i=1:tot_var
		fprintf(f,num2str(inc(i)));
		fprintf(f,'\n');
	end
	fprintf(f,'\n');

	%
	% Regression constants
	%
	fprintf(f,'[Regression constants]');
	fprintf(f,'\n');
	for i=1:num_inc
		fprintf(f,num2str(cfs(i)));
		fprintf(f,'\n');
	end
	fprintf(f,'\n');

	%
	% Filter constants
	%
	fprintf(f,'[Filter constants]');
	fprintf(f,'\n');
	for i=1:num_inc
		fprintf(f,num2str(flt(i)));
		fprintf(f,'\n');
	end
	fprintf(f,'\n');

	%
	% Polynomial powers
	%
	fprintf(f,'[Polynomial powers]');
	fprintf(f,'\n');
	for i=1:num_inc
		fprintf(f,num2str(ply(i)));
		fprintf(f,'\n');
	end
	fprintf(f,'\n');

	%
	% Maximum y-value for re-scaling
	%
	fprintf(f,'[Maximum output]');
	fprintf(f,'\n');
	fprintf(f,num2str(maxy));
	fprintf(f,'\n');
	fprintf(f,num2str(miny));
	fprintf(f,'\n');
	fprintf(f,'\n');
	%
	% Close the file
	%
	fprintf(f,'[END]');
	fclose(f);
	out=1;
else
	out=0;
end