Yep, I was just confused there. Thanks,
-- Raul On Tue, Jan 25, 2022 at 1:57 PM Ric Sherlock <[email protected]> wrote: > > Raul, > Not sure if your request was prompted by your thinking that H. was no > longer available in J9, but in case it's useful erf as defined in > stats/distribs/normal works on non integers. > > erf 0.5 1 1.5 > 0.5205 0.842701 0.966105 > > NB. erf v error function > NB. ref Abramovitz and Stegum 7.1.21 (right) > erf=: (*&(%:4p_1)%^@:*:)*[:1 H. 1.5*: > > > On Sun, 23 Jan 2022, 07:53 Raul Miller, <[email protected]> wrote: > > > https://www.jsoftware.com/help/dictionary/dhcapdot.htm > > > > Using > > > > H=: {{(m {{m rf %&(*/) n rf=. {{(,m) ^!.1/ i.@[}}}} n +/ .* (i.@[ ^~ > > ]) % !@i.@[)"0}} > > erf=: (1 H 1.5)@*: * 2p_0.5&* % ^@:*: > > > > or > > > > rf=: 1 : '(,m) ^!.1/ i.@[' NB. Rising factorial > > L1=: 2 : 'm rf %&(*/) n rf' > > L2=: (i.@[ ^~ ]) % !@i.@[ > > H =: 2 : '(m L1 n +/ . * L2) " 0' > > erf=: (1 H 1.5)@*: * 2p_0.5&* % ^@:*: > > > > I get > > > > erf 0.5 1 1.5 > > | domain error > > > > Apparently the problem is that this version of hypergeometric can only > > handle integer arguments (because L2 uses i. which requires integer > > arguments). > > > > Does anyone know how to implement the J version 8 of erf? > > > > Thanks, > > > > -- > > Raul > > ---------------------------------------------------------------------- > > For information about J forums see http://www.jsoftware.com/forums.htm > > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
