Hi all

 

Based on the feedback from my last e-mail to the forum, I put together the
attached script. 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

 

 

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to