Re: [Jprogramming] A Neural Network in 11 lines of Python (Part 1)

2017-11-17 Thread Raul Miller
The 11 line fragment at the top is probably something like tiny_toy_network=:3 :0 X=: #: 1 3 5 7 Y=: ,. 0 1 1 0 syn0=: (2*?3 4$0)-1 syn1=: (2*?4 1$0)-1 for_j. i.6 do. l1=: %1+^-X +/ .* syn0 l2=: %1+^-l1 +/ .* syn1 l2_delta=: (Y - l2)*(l2*(1-l2)) l1_delta=: (l2_delta +

[Jprogramming] A Neural Network in 11 lines of Python (Part 1)

2017-11-17 Thread 'Skip Cave' via Programming
​All, I found this article on building a Neural Network in 11 lines of Python. Does anyone know enough Python to translate this into J? ​http://iamtrask.github.io/2015/07/12/basic-python-network/ Skip -- For information about J

Re: [Jprogramming] parsing fixed-width files

2017-11-17 Thread Rob Hodgkinson
Thanks Gilles, my typo indeed ! > On 18 Nov 2017, at 4:57 am, Gilles Kirouac wrote: > > 1(0,+/\}.20 10 12)}42$0 NB. mask for length 10 12 20 > > 1(0,+/\}:20 10 12)}42$0 NB. Drop last, not first; for 20 10 12 > > > Le 2017-11-17 à 00:24, Daniel Lyons a écrit : >> Thanks Rob! >> >> I did feel

Re: [Jprogramming] Partitions

2017-11-17 Thread Raul Miller
blockmake=: (-@[ {."1&.> ] <@({.,"0 1 }.)"0 1~ 1+])i. (5 blockmake 4) -: 4 makeblock 5 1 Thanks, -- Raul On Fri, Nov 17, 2017 at 1:14 PM, 'mike_liz@tiscali.co.uk' via Programming wrote: > Erling Helenas, Raul Miller, and others have come up with various > methods to generate subse

Re: [Jprogramming] parsing fixed-width files

2017-11-17 Thread chris burke
a variation on Rob's method: dat=: 3 42$'John Smith WA 418-Y11-4111' (restore the blanks) msk=: (i.42) e. +/\0,20 10 12 ('';msk) <;.1 dat On Thu, Nov 16, 2017 at 9:24 PM, Daniel Lyons wrote: > Thanks Rob! > > I did feel that the general idea of the rolling sum generating the indic

Re: [Jprogramming] Partitions

2017-11-17 Thread 'mike_liz....@tiscali.co.uk' via Programming
Erling Helenas, Raul Miller, and others have come up with various methods to generate subsets of “restricted generating functions” (RGFs) suitable for the production of partitions of sets. Several of these have used Ruskey’s algorithm. I’ve found a fairly simple approach which has the benefi

Re: [Jprogramming] parsing fixed-width files

2017-11-17 Thread Gilles Kirouac
1(0,+/\}.20 10 12)}42$0 NB. mask for length 10 12 20 1(0,+/\}:20 10 12)}42$0 NB. Drop last, not first; for 20 10 12 Le 2017-11-17 à 00:24, Daniel Lyons a écrit : > Thanks Rob! > > I did feel that the general idea of the rolling sum generating the indices of > the frets was not terrible, it ju

Re: [Jprogramming] geometric mean, and rank

2017-11-17 Thread Henry Rich
&.: Henry Rich On Nov 17, 2017 11:32 AM, "Raul Miller" wrote: > ^. is rank 0, which means you would be taking the geometric mean of > each number individually. > > You could do E&.(^."1) or E&.(^."2) or probably E&.(^."_) depending on > what you are trying to accomplish. > > Since E&.:^. is equ

Re: [Jprogramming] geometric mean, and rank

2017-11-17 Thread Raul Miller
^. is rank 0, which means you would be taking the geometric mean of each number individually. You could do E&.(^."1) or E&.(^."2) or probably E&.(^."_) depending on what you are trying to accomplish. Since E&.:^. is equivalent to E&.(^."E) which, in turn, is equivalent to E&.(^."_) you probably d

[Jprogramming] geometric mean, and rank

2017-11-17 Thread 'Bo Jacoby' via Programming
The arithmetic mean is    E=.+/ % # For example   E 2 3 2.5 The geometric mean is    E&.(2&^.) 2 3 2.44949 or E&.:(^.) 2 3 2.44949 but (&.) does not work with (^.) E&.(^.) 2 3 2 3 How come? Thank! Bo.  -- For information about J f