This has been a recurring problem. The best that I have is due to
Martin Neitzel: (watch line wrap)
NB. u is gerund, v is a verb. [x] (k{u)`:0 &. v is applied to cell k of y
NB. The first cell is evaluated to get the shape of a result cell, then
all cells are reevaluated
respectively =: 2 : 0
a: (<@:((,'2')&(,&<))@:((<,']')&,) "0 u) respectively v y
:
z =. (#y) $ ,: x ({.m)`:0&.v {.y
for_y. y do. z =. (x (((#m)|y_index){m)`:0 &.v y) y_index} z end.
)
Henry Rich
Sherlock, Ric wrote:
> Can anybody provide some guidance on the following?
>
> fns=: ((-~ +/)"1)`((*~ +/)"1)`((+~ +/)"1)`((%~ +/)"1)
> rnge=. >: i. 5
> ]data=. (i.5 3);(i.5 2);(i.5 4);(i.5 3)
> +--------+---+-----------+--------+
> | 0 1 2|0 1| 0 1 2 3| 0 1 2|
> | 3 4 5|2 3| 4 5 6 7| 3 4 5|
> | 6 7 8|4 5| 8 9 10 11| 6 7 8|
> | 9 10 11|6 7|12 13 14 15| 9 10 11|
> |12 13 14|8 9|16 17 18 19|12 13 14|
> +--------+---+-----------+--------+
>
> Given the gerund "fns" (4 dyadic verbs) and the noun "data" (4 arrays which
> are the desired arguments for the corresponding verbs in the gerund), how can
> I apply the verbs to their respective arguments such that:
>
> rnge applyVerbs data
> +-------------------+--------------+--------------+--------------+
> | 3 1.5 1 0.75 0.6| 7 8 9 10 11| 1 2 3 4 5| 2 1 0 _1 _2|
> |12 6 4 3 2.4|23 24 25 26 27| 5 10 15 20 25|11 10 9 8 7|
> |21 10.5 7 5.25 4.2|39 40 41 42 43| 9 18 27 36 45|20 19 18 17 16|
> |30 15 10 7.5 6|55 56 57 58 59|13 26 39 52 65|29 28 27 26 25|
> |39 19.5 13 9.75 7.8|71 72 73 74 75|17 34 51 68 85|38 37 36 35 34|
> +-------------------+--------------+--------------+--------------+
>
> Here is a looping explicit solution, but I feel like should be a
> simpler/better way?
>
> applyVerbs=: 3 : 0
> (>: i. 4) applyVerbs y
> :
> res=. ''
> for_mod. y do.
> res=. res ;~ x [email protected]_index >mod
> end.
> }:res
> )
>
> For those interested in the actual problem - the verbs are different models
> that are being compared. Different models have different numbers of inputs
> (the columns of arrays in data) and the rows of the arrays in data are
> different suggested input combinations. I'm going to multiplot the result. Eg:
> require 'plot'
> pd 'reset'
> pd 'multi 2 2 , 2 2'
> pd (<'type line;'),.rnge makeFuncData data
> pd 'show'
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
