[Jprogramming] Kalman filter in J?

[Ian Clark]

@Devon: okay, I guess the PDF is misbehaving because it uses some
other font -- a []AV-based layout -- than today's standard: "APL385
Unicode" -- and I haven't got it installed. As for the RTF, TextEdit
has used what it considers the "nearest" font and shows me a garbling
(which I just happen to recognise!) with '@' for the lamp symbol and
'„' for the left-arrow.

I've spent thousands of hours decyphering this sort of nonsense in
archived papers. "APL385 Unicode" really is the way to go. Or (of
course) J!

PM me for your mailbox, and I'll reply with my scanned PDF from an
actual Vector hardcopy as an attachment. Can't get more definitive
than that (...it was, after all, what was actually published).

Ian

On Thu, Jun 21, 2012 at 5:12 PM, Devon McCormick <devonmcc at gmail.com> wrote:
> The APL character reaches out from the distant past to continue to
> plague me - another reason to re-write this in J.
>
> When I look at both documents on my machine, I see all the APL except
> for the domino-quads in the MultiVDLM3 function: in the PDF, this is
> rendered as a bold, san-serif, capital "z" with a "caron"
> (http://en.wikipedia.org/wiki/Caron) accent.  This infernal character
> set also continues to teach me new things!
>
> I've put a picture of the final APL function from the paper here:
> http://www.jsoftware.com/jwiki/DevonMcCormick/DynamicLinearModels/BayesianFinancialDynamicLinearModel
> .
>
> On Thu, Jun 21, 2012 at 11:30 AM, Ian Clark <earthspotty at gmail.com> wrote:
>> @Devon, thank you for that ref. One more technique for me to test. I
>> guess I can handle APL.
>>
>> I've just taken a look at the link you mentioned: http://sigapl.org/articles.php
>> FYI -- in both the downloadable PDF and RTF versions of your paper,
>> the APL code is mutilated and unreadable.
>>
>> By amazing coincidence, just a month ago I was asked to supply a copy
>> of that very paper, in my capacity as BAA Vector archivist, in the
>> version you published in Vector Volume 21 No 2 (Spring 2005) p65. This
>> I did by scanning the actual printed Vector back-copy, in which of
>> course the APL code is clearly legible.
>>
>> Ian
>>
>> On Thu, Jun 21, 2012 at 4:03 PM, Devon McCormick <devonmcc at gmail.com> wrote:
>>> Ian -
>>>
>>> you might try searching for "dynamic linear model" as well.  I know of
>>> one written in J but it's proprietary.  I've been meaning to re-write
>>> my APL paper - an introduction to dynamic linear models (at
>>> http://sigapl.org/Articles/) - in J but haven't gotten around to it
>>>
>>> Regards,
>>>
>>> Devon
>>>
>>> On Thu, Jun 21, 2012 at 8:34 AM, Ian Clark <earthspotty at gmail.com> wrote:
>>>> Thanks, Bo and Ric. Yes, I'd tried searching jwiki on "Kalman" and
>>>> Pieter's page was the only instance.
>>>
>>> --
>>> Devon McCormick, CFA
>>> ^me^ at acm.
>>> org is my
>>> preferred e-mail
>>> ----------------------------------------------------------------------
>>> For information about J forums see http://www.jsoftware.com/forums.htm
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/forums.htm
>
>
>
> --
> Devon McCormick, CFA
> ^me^ at acm.
> org is my
> preferred e-mail
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm