From Eigenvector Research Documentation Wiki
Jump to navigation Jump to search


Output a fractional factorial design matrix.


desgn = ffacdes1(k,p)
[desgn, col_ID, alias_ID, res] = ffacdes1(k,p,options)


FFACDES1 outputs a 2(k-p) fractional factorial design of experiments. The design is constructed such that the highest order interaction term is confounded. This is one way to select a fractional factorial. Input k is the total number of factors in the design and p is the number of confounded factors {default: p = 1}. Note that it is required that p < k. Output desgn is the experimental design matrix.


  • k = total number of factors in the design.

Optional Inputs

  • p = numerical indicator of the fraction desired (default = 1)
1 : Half Fraction
2 : Quarter Fraction
3 : Eighth Fraction
4 : Sixteenth Fraction
Note: Only fractionation up to 1/16 presently supported


  • desgn = experimental design matrix

If the dso option is false, the outputs include:

  • desgn = is a matrix of the experimental design in uncoded form
  • col_ID = is a cell array of strings describing the multiplicative origin of each column; one col for each coefficient to potentially be calculated where:
    • the first k cells describe the original main factors
    • the remaining cells describe the various interactions among main factors.
  • alias_ID = is a cell array of logicals describing the alias structure of the selected design; one row per coefficient/X-column. Multiplying a given logical by the full set of characters representing the factors yields the alias relationship for that row/coefficient.
ABCDEF .* [1 0 0 1 1 0] = ADE
  • res = resolution of the selected design.

See Also

boxbehnken, ccdface, ccdsphere, doegen, doescale, factdes