Re: unexpected failure on all formulas

[Nick Dokos <nicholas.dokos@hp.com>]

Christian Moe <mail@christianmoe.com> wrote:


> I think your vmode() needs to subtract one from the result. The mode
> in the "systolic" column of Jude's dataset should be 124 (f=12), not
> 125 (f=2).
> 

Yes, this is an off-by-one error: in vfreqs, I should cons prev onto
freqs, not (nth 0 vec). In addition to this and the multimodal problem
that both you and Lawrence Mitchell pointed out, there is another
off-by-one error when vfreqs gets to the end of the list.

All in all, a disaster: can a posting be deleted? or at least marked
with a big, red X so that it won't mislead in the future? I'll try to
post a corrected version later on.

OTOH, my main interest in this was the hooking up of a user function
onto calc and the example, though deeply flawed, does illustrate that.

Thanks to both Christian and Lawrence for the whacks in the head
(although now I have a headache...)

Nick


> For Jude's limited purposes I got by with less code (but note the
> caveats):
> 
> #+begin_src emacs-lisp
>   (defmath vmode (vec)
>     "Returns the mode (most frequent value) of a unimodal vector of
>   integers VEC.  WARNING: Will not work with non-integer values or
>   multiple vectors. Will only return the lowest mode for multimodal
>   data."
>     (let ((freq (histogram vec (1+ (vmax vec)))))
>             (1- (find freq (vmax freq)))))
> #+end_src
> 
> Using :=vmode(@I..@II), I get 124 (systolic), 80 (diastolic), 68 (pulse).
> 
> Yours,
> Christian
> 
> On 9/20/11 10:26 AM, Nick Dokos wrote:
> > Nick Dokos<nicholas.dokos@hp.com>  wrote:
> >
> >> Jude DaShiell<jdashiel@shellworld.net>  wrote:
> >>
> >>
> >>> |------------------------+----------+-----------+-------|
> >>> | mode                   |          |           |       |
> >>> | Min                    |          |           |       |
> >>> | Max                    |     127. |       81. |   74. |
> >>> #+TBLFM: @>>>$2..@>>>$4=vmode(@i..@ii) :: @>>$2..@>>$4=vmin(@i..@ii) :: @>$2..@>$4=vmax(@i..@ii)
> >>>
> >>>
> >>> # Local Variables:
> >>> # kept-new-versions: 100
> >>> # version-control: t
> >>> # End:
> >>> cut here.
> >>> What did I do wrong with this #+TBLFM: line?
> >>>
> >>
> >> Two things:
> >>
> >> o as Christian Moe pointed out, calc does not provide a vmode() function.
> >>
> >> o the separator lines in the formulas have to be specified using capital
> >> I letters.
> >>
> >
> > Here is an implementation of vmode - afaict, it works correctly but
> > that may be because I've only tested it on a single set of data. If
> > you find something wrong, please let me know.
> >
> > --8<---------------cut here---------------start------------->8---
> > | Date Stamp             | Systolic | Diastolic | Pulse |
> > |------------------------+----------+-----------+-------|
> > | [2011-07-19 Tue 02:26] |      138 |        92 |    74 |
> > | [2011-07-20 Wed 04:03] |      130 |        85 |    74 |
> > | [2011-07-21 Thu 03:50] |      128 |        79 |    76 |
> > | [2011-07-22 Fri 02:33] |      121 |        80 |    79 |
> > | [2011-07-23 Sat 02:52] |      118 |        75 |    68 |
> > | [2011-07-24 Sun 04:13] |      126 |        83 |    78 |
> > | [2011-07-25 Mon 02:12] |      122 |        79 |    76 |
> > | [2011-07-26 Tue 01:57] |      124 |        79 |    72 |
> > | [2011-07-27 Wed 04:03] |      137 |        74 |    76 |
> > | [2011-07-28 Thu 03:47] |      136 |        80 |    76 |
> > | [2011-07-29 Fri 04:06] |      120 |        72 |    76 |
> > | [2011-07-30 Sat 03:18] |      128 |        92 |    80 |
> > | [2011-07-31 Sun 07:11] |      124 |        84 |    72 |
> > | [2011-08-01 Mon 04:55] |      124 |        77 |    69 |
> > | [2011-08-02 Tue 02:47] |      128 |        80 |    67 |
> > | [2011-08-03 Wed 06:00] |      132 |        89 |    70 |
> > | [2011-08-04 Thu 04:30] |      137 |        82 |    77 |
> > | [2011-08-05 Fri 06:10] |      124 |        86 |    77 |
> > | [2011-08-06 Sat 04:04] |      139 |        88 |    78 |
> > | [2011-08-07 Sun 04:02] |      124 |        80 |    68 |
> > | [2011-08-08 Mon 06:36] |      137 |        93 |    79 |
> > | [2011-08-10 Wed04:03]  |      120 |        80 |    68 |
> > | [2011-08-11 Thu 04:30] |      128 |        78 |    77 |
> > | [2011-08-12 Fri 06:36] |      143 |        82 |    76 |
> > | [2011-08-13 Sat 06:18] |      125 |        76 |    69 |
> > | [2011-08-14 Sun 04:04] |      123 |        70 |    69 |
> > | [2011-08-15 Mon 04:04] |      135 |        83 |    76 |
> > | [2011-08-16 Tue 04:03] |      128 |        80 |    69 |
> > | [2011-08-17 Wed 04:04] |      124 |        80 |    68 |
> > | [2011-08-18 Thu 05:58] |      136 |        87 |    72 |
> > | [2011-08-19 Fri 05:50] |      116 |        80 |    64 |
> > | [2011-08-20 Sat 06:38] |      139 |        84 |    64 |
> > | [2011-08-21 Sun 08:20] |      124 |        84 |    73 |
> > | [2011-08-22 Mon 03:50] |      138 |        76 |    70 |
> > | [2011-08-23 Tue 05:52] |      126 |        79 |    68 |
> > | [2011-08-24 Wed 04:53] |      125 |        81 |    64 |
> > | [2011-08-25 Thu 05:37] |      124 |        79 |    68 |
> > | [2011-08-26 Fri 06:41] |      124 |        87 |    74 |
> > | [2011-08-27 Sat 05:44] |      133 |        74 |    61 |
> > | [2011-08-28 Sun 03:22] |      124 |        76 |    61 |
> > | [2011-08-29 Mon 04:06] |      124 |        76 |    65 |
> > | [2011-08-30 Tue 05:52] |      139 |        87 |    71 |
> > | [2011-08-31 Wed 06:47] |      152 |        86 |    68 |
> > | [2011-09-01 Thu 06:00] |      136 |        86 |    78 |
> > | [2011-09-02 Fri 01:22] |      142 |       100 |    78 |
> > | [2011-09-03 Sat 09:12] |      126 |        79 |    70 |
> > | [2011-09-04 Sun 11:53] |      130 |        90 |    72 |
> > | [2011-09-05 Mon 06:29] |      138 |        77 |    65 |
> > | [2011-09-06 Tue 05:42] |      121 |        76 |    68 |
> > | [2011-09-07 Wed 06:47] |      131 |        77 |    68 |
> > | [2011-09-08 Thu 06:02] |      122 |        76 |    72 |
> > | [2011-09-10 Sat 05:38] |      126 |        76 |    69 |
> > | [2011-09-11 Sun 10:12] |      137 |        87 |    64 |
> > | [2011-09-12 Mon 04:00] |      135 |        78 |    89 |
> > | [2011-09-13 Tue 06:00] |      132 |        82 |    66 |
> > | [2011-09-14 Wed 06:01] |      137 |        77 |    77 |
> > | [2011-09-15 Thu 04:08] |      141 |        87 |    64 |
> > | [2011-09-17 Sat 08:35] |      121 |        78 |    66 |
> > | [2011-09-18 Sun 07:04] |      124 |        69 |    65 |
> > | [2011-09-19 Mon 06:00] |      133 |        81 |    69 |
> > |------------------------+----------+-----------+-------|
> > | mode                   |      125 |        81 |    69 |
> > | min                    |      116 |        69 |    61 |
> > | max                    |      152 |       100 |    89 |
> > #+TBLFM: @>>>$2..@>>>$4=vmode(@I..@II) :: @>>$2..@>>$4=vmin(@I..@II) :: @>$2..@>$4=vmax(@I..@II)
> >
> > #+begin_src elisp
> >
> > (defun calcFunc-vmode (&rest vecs)
> >    (let* ((flat (copy-sequence (cdr (math-flatten-many-vecs vecs))))
> > 	 (p flat)
> > 	 (len (length flat))
> > 	 (hlen (/ len 2)))
> >      (if (= len 0)
> > 	(math-reject-arg nil "*Must be at least 1 argument")
> >        (if (and (= len 1) (memq (car-safe (car flat)) '(sdev intv)))
> > 	  (calcFunc-vmean (car flat))
> > 	(while p
> > 	  (if (eq (car-safe (car p)) 'sdev)
> > 	      (setcar p (nth 1 (car p))))
> > 	  (or (Math-anglep (car p))
> > 	      (math-reject-arg (car p) 'anglep))
> > 	  (setq p (cdr p)))
> > 	(setq flat (sort flat 'math-lessp))
> > 	(setq freqs (vfreqs flat))
> > 	(cadr (maxfreq freqs))
> > 	))))
> >
> > (defun vfreqs (v)
> >    "v is assumed to be sorted.
> > Return a list of pairs: (count element)
> > for each distinct element of v."
> >    (let ((count 0)
> > 	(prev (car v))
> > 	(vec v)
> > 	freqs)
> >      (setq freqs nil)
> >      (while vec
> >        (if (= (nth 0 vec) prev)
> > 	  (setq count (1+ count))
> > 	(setq freqs (cons (list count (nth 0 vec)) freqs))
> > 	(setq count 1 prev (nth 0 vec)))
> >        (setq vec (cdr vec)))
> >      freqs))
> >
> > (defun freq-greater (a b)
> >    (if (>  (car a) (car b))
> >        t
> >      nil))
> >
> > (defun maxfreq (f)
> >    "Sort the frequency list by count and return the first pair."
> >    (let ((fs (sort f 'freq-greater)))
> >      (nth 0 fs)))
> >
> > #+end_src
> > --8<---------------cut here---------------end--------------->8---
> >
> > I cribbed heavily from the implementation of vmedian. Although there is
> > no key binding, you can even use the above in an interactive Calc
> > session, by entering the formula in algebraic form. I yanked the vector
> > twice onto the Calc stack and then entered 'vmode($) with the following
> > result:
> >
> > ,----
> > | --- Emacs Calculator Mode ---
> > | 2:  [138, 130, 128, 121, 118, 126, 122, 124, 137, 136, 120, 128, 124, 124, 128, 132, 137, 124, 139,
> > |      124, 137, 120, 128, 143, 125, 123, 135, 128, 124, 136, 116, 139, 124, 138, 126, 125, 124, 124,
> > |      133, 124, 124, 139, 152, 136, 142, 126, 130, 138, 121, 131, 122, 126, 137, 135, 132, 137, 141,
> > |      121, 124, 133]
> > | 1:  125
> > |     .
> > `----
> >
> > Nick
> >
> >
>