Hello. I would like to write the equation used by the function ‘calibrate’ included in a fuzzy set software.
Such command does the following:
http://www.u.arizona.edu/~cragin/fsQ...alibration.pdf
(“The direct method of calibration“, p. 13-15).
From that I arrived at:
Y = exp{( X – Xh) * (X- Xh) *[ 3/(Xo – Xh )]} / {1 + exp{( X – Xh) * (X- Xh) *[ 3/(Xo – Xh )]}}
For X > Xh
and Y = exp{( X – Xh) * (X- Xh) *[ -3/(Xz – Xh )]} / {1 + exp{( X – Xh) * (X- Xh) *[ -3/(Xz – Xh)]}}
For X < Xh
Where:
Y = output (fuzzy score)
X = input (original score)
Xz = threshold of full non-membership
Xh = crossover point
Xo = threshold of full membership
Did I get it more or less right? I would be happy to receive your comments. Thank you.