STATISTICA Basic Program: Eta-Tran.stb

Products Services Solutions Support Downloads Customers Partners Company

Home I Electronic Statistics Textbook I Contact Us I Request Quote I Search

STATISTICA BASIC Program Eta-Tran.stb

This program will compute the standard (Taguchi) Signal-to-Noise ratios for a user-specified list of variables. The resulting index can be used (as the dependent or Y-variable) together with all types of designs available in the Experimental Design module of STATISTICA.

For more information about Taguchi methods, refer to the respective section in the Experimental Design electronic manual.

Program written, modified, or edited at StatSoft, Inc.}

randomaccess;
NoDataFileVariableNames;
dim matrix(1,1), VarList1(20), VarList2(1), TempMatrix(1,1), RowVector(1);
ifault:=0;

pick :=  DisplayButtonBox ('Taguchi Transformation',
	'Smaller the Better|Nominal the Best|Larger the Better|Signed Target');
if pick = 0 then stop;

okay := SelectVariables2 ('Select Variables for Transformation', 1, 20, VarList1, Count1, 'Dependent Variable(s)',
		  1, 1, VarList2, Count2, 'Result (Target) Variable');
if not okay then stop;
target :=VarList2(1);

redim matrix(ncases,Count1), VarList1(Count1), TempMatrix(ncases,1), RowVector(Count1);

{read dependent variables into matrix for transformation}
for i := 1 to Count1 do
begin
	MatrixExtract (Data, 1, VarList1(i), ?NoOfRows, 1, TempMatrix);
	MatrixSetColumn (Matrix, i, TempMatrix);
end;

if pick = 1 then
begin {smaller the better}

	for i := 1 to NCases do
	begin
		sum := 0;
		for j := 1 to Count1 do sum := sum + Matrix(i,j)^2;
     		DATA(i,target) := -10*log10(sum/Count1);
	end;

end; {smaller the better}

if pick = 2 then
begin {nominal the best}
	for i:= 1 to NCases do
	begin
		MatrixGetRow (Matrix, i, RowVector);
		ValMean (RowVector, 1, ?HowMany, Mean);
		ValVariance (RowVector, 1, ?HowMany, Variance);
	            if (Variance<=0) then ifault:=1;
		DATA(i,target) := 10*log10(Mean^2/Variance);
	end;
end;  {nominal the best}

if pick = 3 then
begin {larger the better}
	for i:= 1 to NCases do
	begin
		element := 0;
		for j:= 1 to Count1 do begin
  	              if (Matrix(i,j)^2<=0) then ifault:=1;
		  element := element+1/Matrix(i,j)^2;
	           end;
		DATA(i,target) := -10*log10(element/Count1);
	end;
end; {larger the better}

if pick = 4 then
begin {signed target}
	for i:= 1 to NCases do
	begin
		MatrixGetRow (Matrix, i, RowVector);
		ValVariance (RowVector, 1, ?HowMany, Variance);
	            if (Variance<=0) then ifault:=1;
		DATA(i,target) := -10*log10(Variance);
	end;
end; {signed target}

if (ifault=1) then DisplayMessageBox (MB_ICONEXCLAMATION, 'Some Transformations Failed',
   'The respective transformation could not be computed for some cases; the TARGET values for those cases were set to missing data.');

Back to List of Programs

[StatSoft]
2300 East 14th Street, Tulsa, OK 74104
Phone: (918) 749-1119; Fax: (918) 749-2217
e-mail: info@statsoft.com