procedure LatinHyperCubeDesign(const X: TMtx; const n: Integer; const p: Integer; const fp: TMtxFloatPrecision; const smooth: Boolean; const IterCriteria: TLHCImprove; const MaxIter: Integer);
Latin Hyper-Cube design.
| # | Name | Description |
|---|---|---|
| 1 | X | Stores experimental design values (rows representing values, columns variables). |
| 2 | n | Defines the number of values for specific variable. |
| 3 | p | Defines the number of variables. |
| 4 | smooth | If true, generate random values for n random (permuted) values are distributed per one point in intervals (0,1/n),(1/n,2/n)... ((n-1)/n,n). If false, produce the points in the middle of define intervals i.e. 0.5/n, 1.5/n, ... . |
| 5 | IterCriteria | The criteria for design improvement. |
| 6 | MaxIter | Maximum number of iterations used for improving the design. |
| 7 | fp | Floating point precision to be used for the design. |
Result: stored in self (calling object)
Remarks:
Generates n x p samples for Latin Hypercube experimental design.
Examples
Use MtxExpr, Statistics;
procedure Example;
var e: Matrix;
begin
// 50 values for each of 3 variables
LatinHyperCubeDesign(e,50,3);
end;