A resend with the script listed below!
Hi all
Based on the feedback from my last e-mail to the forum, I put together the
script below. It illustrates my problem more completely. The input data are
vectors but in a real application, they would be matrices. For each read of
a list in data, I sample from a table (s_r) to obtain values of the noun,
recruit. For illustration, I NBed that line and assigned recruit to 30000.
My idea was to have each table be an independent series of calculations from
the same starting n. Thus, here, the output should be two identical tables.
You will see that the first table estimates each year's n (each row is a
year in this example) but then the calculations run over to the second table
using the n from the last row of the first table rather than the same
starting row as the first table. I have used looping and counters to get
around this in earlier versions but there should be a matrix way of doing
this.
Any insight would be greatly appreciated!
Thanks
Bob
NB. Bootstrap Projection
NB.
NB.=====================================================
NB.
NB. Input Data
n =: 14965 29010 15811 13526 5595
5147 1679 1537 0 0 0 0 0 0 0 0
waa =: 0.292 0.533 0.795 1.22 1.741
2.408 3.347 3.63 4.515 4.926 5.313 5.63 5.885 6.089 6.249 6.375
paa =: 0.38 0.91 0.98 1 1 1 1 1 1 1 1
1 1 1 1 1
maa =: 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
praa =: 0.02 0.15 0.72 1 1 1 1 1 1 1 1 1 1 1 1 1
recruits =: 80323 48380 48579 53444 43921 13033 27270
30352 28686 32316 33158 35257 38694 28258 29339 31185 29433 39386 31181
36047 25265 35505 14965
ssb =: 22582 24435 21870 19853 18391
19082 10883 7025 9940 8743 9905 12287 15100 18976 20067 20413 22245 22551
26130 33835 39051 44786 43951
breakpoint =: 15000
option =: 10700 0.21 0.3 0.4 0.25 0.15 0.20 0.25 0.3
s_r =: recruits,:ssb
recruit_l =: {.(({:s_r)< breakpoint)#"1 s_r
recruit_h =: {.(({:s_r)>:breakpoint)#"1 s_r
spawn_t =: 0.9
data =: (waa, paa, maa,:praa);" 2 0 option
NB.
NB.======================================================
NB.
NB. Functions
NB.
F_QUOTA =: 4 : 0
y*x%(+/w*(f%z)*n*1-^-z=.m+f=.pr*y)
)
PROJECT_BOOT =: 3 : 0
y1 =: >0{y
y2 =: >1{y
'w per_mat m pr' =: y1
'opt' =: y2
f =: >(opt
<2) { (pr * opt (F_QUOTA ^:_) 2); (pr * opt)
b =:
+/per_mat*w*n*^-(m+f)*spawn_t
NB. recruit =:
(b<breakpoint){((?$ recruit_h){ recruit_h),((?$ recruit_l){ recruit_l)
recruit =: 30000
n =:
recruit,}:n *^-(f+m)
)
PROJECT_BOOT"2"1 data,:data
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
