Thanks to David! Your trick is immediately applicable.
Thanks to Raul! I do not require comb-examples in order to conjecture that 'x T
y' is independent of sample size 'x' (for 2<:x and x<:#y), but I need an
algebraic proof. (I have not found it in the litterature)./ Bo.
E =.+/%# NB. mean
Va =. E@:*:-*:@:E NB. Variance
Bv =.-.&.:%@#*Va NB. Bessel variance
S=.((((=+/"1)#])[:#:[:i.2&^)#)#] NB. (x S y) = x-samples from population y
T=.E@Bv@|:@S NB. (x T y) = mean of bessel variances from (x S y). ]n=.?8$100
NB. some random population
34 40 46 84 29 8 75 97
2 T n
935.125
3 T n
935.125
4 T n
935.125
5 T n
935.125
6 T n
935.125
7 T n
935.125
8 T n
935.125
Den søndag den 1. august 2021 22.10.29 CEST skrev David Lambert
<b49p23t...@gmail.com>:
Noting that binary is the base for monadic antibase,
h=: ((= +/"1) # ]) ([: #: [: i. 2&^)
3 (0 -.@:e. (Smps -: h)&>) #\i.9 NB. a test
1
And Raul already arrived here. Can we further eliminate ]same?
Date: Sat, 31 Jul 2021 15:52:03 +0000 (UTC)
From: "'Bo Jacoby' via Programming"<programm...@jsoftware.com>
To: Programming Forum<programm...@jsoftware.com>
Subject: [Jprogramming] Tacit hook
Message-ID:<1298132844.2251818.1627746723...@mail.yahoo.com>
Content-Type: text/plain; charset=UTF-8
Consider this code
f=.#&2#:[:i.2&^ NB. 0-1 rows
Smps=. 4 : '(#~x=+/"1)f y' NB. select rows
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
