J
Jerry W. Lewis
No, but you can use the relationship between erf and statistical
distributions to fake it.
erf(x) = GammaDist(x^2,0.5,1,True)
so
erfinv(p) = Sqrt(GammInv(p,0.5,1))
erfc(x) = ChiDist(2*x^2,1)
so
erfcinv(p) = Sqrt(ChiInv(p,1)/2)
The ChiDist function is much more accurate for large x than the ATP erfc
function. The Inv functions are useless for p's associated with x > 3.5
in Excel versions prior to 2003.
Jerry
distributions to fake it.
erf(x) = GammaDist(x^2,0.5,1,True)
so
erfinv(p) = Sqrt(GammInv(p,0.5,1))
erfc(x) = ChiDist(2*x^2,1)
so
erfcinv(p) = Sqrt(ChiInv(p,1)/2)
The ChiDist function is much more accurate for large x than the ATP erfc
function. The Inv functions are useless for p's associated with x > 3.5
in Excel versions prior to 2003.
Jerry