I have a long table that calculates a column from 3 other columns.
It's not exactly a J problem, but I'm interested in J solutions or any
other ideas and advice.
I'd like to gain some intuition about how the 3 variables interact to
compute the outcome and identify any other patterns. I have another
function that works off of 4 - 5 variables. Each of the variables are
bounded by reasonable business values. I'd like a way to plot/visualize the
adta.
Ideas I considered:
1. 3d contour plot but didn't think it would resonate with the audience and
also wouldn't work with more dimensions
2. A 2d scatter plot and label the points with two of the variables
For example: plotting x=t and y=rr and labeling "a,d" as the point label.
Any other ideas?
ts=:(i.100) + 1
as=:(i.100) % 100
ds=: (i.200)
NB. Calculate cartesian product of possible values
cc=: > , { ts;ds;as
d=:(1}"1 cc) % 10000
t=:(0}"1 cc)
a=:(2}"1 cc)
V=:1000
rr=: (V * d) + ((V % t) * (a+0.0065)) % V
tbl=: (cc,.rr)
Example:
10 {. tbl
1 0 0.01 0.0165
1 0 0.02 0.0265
1 0 0.03 0.0365
1 0 0.04 0.0465
1 0 0.05 0.0565
1 0 0.06 0.0665
1 0 0.07 0.0765
1 0 0.08 0.0865
1 0 0.09 0.0965
10 {. 100000 }. tbl
6 0 0 0.00108333
6 0 0.01 0.00275
6 0 0.02 0.00441667
6 0 0.03 0.00608333
6 0 0.04 0.00775
6 0 0.05 0.00941667
6 0 0.06 0.0110833
6 0 0.07 0.01275
6 0 0.08 0.0144167
6 0 0.09 0.0160833
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
