Sorry, wrong scale direction, with increase of radius is should become more tolerant (returning 0).
Correct is tcmp=: 1 : '- ([EMAIL PROTECTED] * >&|) m*>.' 1134000 (0.001 tcmp) 1138000 _1 1134000 (0.001 tcmp)~ 1138000 1 1134000 (0.01 tcmp) 1138000 0 2 2.01 2.02 2.03 (0.01 tcmp) 2 0 0 0 1 2 2.01 2.02 2.03 (0.01 tcmp)&- 2 0 0 0 _1 _1 1 (0.01 tcmp) 1 _1 _1 1 2 2.001 2.01 2.1 2.5 (0.0001 tcmp) 2 0 1 1 1 1 2 2.001 2.01 2.1 2.5 (0.001 tcmp) 2 0 0 1 1 1 2 2.001 2.01 2.1 2.5 (0.01 tcmp) 2 0 0 0 1 1 2 2.001 2.01 2.1 2.5 (0.1 tcmp) 2 0 0 0 0 1 2 2.001 2.01 2.1 2.5 (0.5 tcmp) 2 0 0 0 0 0 --- Bj�rn Helgason <[EMAIL PROTECTED]> wrote: > Right to the point as usual! > > 2007/3/29, Oleg Kobchenko <[EMAIL PROTECTED]>: > > > > Maybe you need to convert the timestamps to > > integers and use a scaled tolerant comparison: > > > > tcmp=: 1 : '(m*>.) (>&| * [EMAIL PROTECTED]) -' > > 1134000 (0.01 tcmp) 1138000 > > _1 > > 1134000 (0.001 tcmp) 1138000 > > 0 > > 1134000 (0.01 tcmp)~ 1138000 > > 1 > > > > This is by extended definition of ''tolerance''. > > > > http://www.google.com/search?q=tolerance&sitesearch=jsoftware.com > > > > > > > > --- Bj�rn Helgason <[EMAIL PROTECTED]> wrote: > > > > > Lets say that the problem is to do a fuzzy match > > > Not exact match. > > > As you see the time difference is sometimes 18 56, 18 55, 18 57 so a > > good > > > match is something like 18 56 +/- 1 or 2 seconds > > > It would not be any problem if the difference is always exactly 18 56 > > > The problem is also finding what the time difference is and in what > > > direction. > > > Time difference by some measurements outside the few second span would > > not > > > be good. > > > > > > So getting a 0/1 answer would rather be something like an interval of a > > few > > > seconds for the timedifferences in each case > > > > > > This problem is fairly trivial to do manually. > > > > > > Finding the probe by looking at the image was also trivial manually. > > > Seeing if the probe was there and if it was in the center or in the > > > periphery. > > > Doing it by a program seemed to me like an impossibility a month ago. > > > That part is the easy part now. > > > > > > Doing this fuzzy matching is probably easy as well even if it is not > > crystal > > > clear to me right now. > > > The differences in time are probably triggered by that the probe is > > > triggered by receiving exposure and it seems to vary a second or two. > > > Also the clock in each equipment is not in sync. > > > > > > The correct match could be within quite a span if the number is coded > > > > > > 24 60 60 1000 #.0 18 54 0 > > > 1134000 > > > 24 60 60 1000 #.0 18 58 0 > > > 1138000 > > > > > > 2007/3/29, Devon McCormick <[EMAIL PROTECTED]>: > > > > > > > > So, taking your guess here of the time offset of 18 minutes, 56 > > seconds= 0 > > > > 18 56 0, I can > > > > adjust c: > > > > > > > > 3{.adjc=. 24 60 60 1000 #: (24 60 60 1000 #.c)-24 60 60 1000 #. 0 > > 18 56 > > > > 0 > > > > 10 57 51 450 > > > > 10 57 59 150 > > > > 10 58 6 950 > > > > > > > > and come up with a boolean to interpolate it into a,b: > > > > > > > > (#adjc)>gv=. /: adjc,a,b > > > > 1 1 1 1 1 0 1 1 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 > > 1 0 > > > > 0 > > > > 1 0 1 1 0 0 1 0 1 0 1 0 1 1 0 0 1 0 1 0 1 1 0 1 0 0 1 > > > > > > > > where the 1s mark successive elements of c and the 0s mark successive > > > > elements of a,b. Is this any help? > > > > > > > > > > > > On 3/29/07, Bj�rn Helgason <[EMAIL PROTECTED]> wrote: > > > > > > > > > > All of c come from a CT or similar equipment. > > > > > a and b come from a instrument that are put into the CT. > > > > > > > > > > The CT gives exposure and in c come the times - and lots of other > > > > > information when the exposure is made. > > > > > > > > > > a and b come from the instrument - a probe in a phantom - measure > > doses > > > > > and > > > > > such and also register time. > > > > > > > > > > I have been making program to find the place where the probe was on > > the > > > > > image. > > > > > > > > > > I have gathered the information from different files I get. > > > > > > > > > > I may not be getting the correct files, files may be missing. > > > > > > > > > > I know a,b and c came from the same session. > > > > > > > > > > The person did some exposures before the probe was connected so some > > > > > values > > > > > are missing. > > > > > > > > > > It is quite obvious that the clock in the probe (a and b) and in the > > CT > > > > > (c) > > > > > are not in sync. > > > > > The time difference is ca 18 minutes and 56 seconds as you can see > > doing > > > > > the > > > > > calculations. > > > > > > > > > > I have a gap at the beginning introducing d to fill the gap. > > > > > The gap could have been in the middle. > > > > > I have no way of knowing in advance. > > > > > > > > > > I am trying to make a reasonable guess which times fit so I can > > match > > > > the > > > > > exposure doses gathered from the probe (a and b) and not having to > > do it > > > > > manually. > > > > > > > > > > The time difference is not exactly 18 minutes and 56 seconds between > > all > > > > > corresponding lines in c and the lines in a and b. I do not know in > > > > > advance > > > > > if I do have a time difference and I do not know which way the > > > > difference > > > > > is > > > > > going to be. > > > > > > > > > > It is fair to assume all the times in c to be sorted. > > > > > It is also fair to assume the the individual files a and b are > > sorted. > > > > > > > > > > I have to find where lines in a matches lines in c > > > > > I have to find where lines in b matches lines in c > > > > > > > > > > The time difference in all cases should be roughly the same. > > > > > > > > > > I guess if I get a file - lets call it e - I would have to try to be > > > > able > > > > > to > > > > > make an educated guess that I could not find any match and it would > > be > > > > > from > > > > > some other test. > > > > > > > > > > Hope that explains the task at hand fairly well. > > > > > > > > > > Up to now I have done this part manually but would like to be able > > to > > > > see > > > > > if > > > > > I could get this to work. > > > > > I have never tried to match something like this with a program > > before. > > > > > Most of the time there is an exact match or there is not. > > > > > Not such a fuzzy match of matching roughly. > > > > > > > > > > Obviously I have to match the dates as well and both the formats of > > the > > > > > date > > > > > and the time were different. > > > > > I also added 000 in both a and b > > > > > If I am not lucky the test person could be sending in files from > > > > different > > > > > tests. > > > > > I am working on what requirements the test persons are supposed to > > have. > > > > > Try to exclude complete morons somehow politely. > > > > > > > > > > 2007/3/29, Devon McCormick <[EMAIL PROTECTED]>: > > > > > > > > > > > > Gosi - I don't understand what you're trying to do here. > > > > > > Are these timestamps that need to be matched? If so, it looks > > like > > > > > > all of "c" comes after everything in "a" and "b". > > > > > > > > > > > > gv=. /:c,a,b > > > > > > gv<#c > > > > > > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 > > 1 1 > > > > 1 > > > > > 1 > > > > > > 1 > > > > > > 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 > > > > > > > > > > > > Also, is there any reason to complicate the problem by having a > > > > separate > > > > > > "a" > > > > > > and "b"? > > > > > > Aren't you dealing with ab=. a,b versus c? > > > > > > > > > > > > Maybe I need either more context or a simpler version of this. > > > > > > > > > > > > On 3/29/07, Bj�rn Helgason <[EMAIL PROTECTED]> wrote: > > > > > > > > > > > > > > Does anyone have a good way of matching times? > > > > > > > > > > > > > > Here I have an example they happen to be somewhat in order. > > > > > > > c contains 35 lines that need to be matched with a and b. > > > > > > > There is an excess of 5 lines in c compared to a+b so d contains > > 5 > > > > > > > dummies. > > > > > > > > > > > > > > The times need not be sorted like this and they may be missing > > in > > > > > either > > > > > > > direction. > > > > > > > > > > > > > > The task is to get a best match which gives the least time > > > > difference > > > > > > and > > > > > > > get the indexes for c that give the best match for a+b > > > > > > > > > > > > > > There are a lot of fields in each line but only connection > > between > > > > > them > > > > > > is > > > > > > > the time field. > > > > > > > In this example the time difference between the equipment is > > roughly > > > > > 19 > > > > > > > minutes > > > > > > > > > > > > > > > > > > > > > > > > > > > > NB. -----------------head > > > > > > > > > > > > > > a=:".;._2 (0 : 0) > > > > > > > 11 00 29 000 > > > > > > > 11 00 38 000 > > > > > > > 11 00 44 000 > > > > > > > 11 00 52 000 > > > > > > > 11 01 00 000 > > > > > > > 11 03 20 000 > > > > > > > 11 03 28 000 > > > > > > > 11 03 35 000 > > > > > > > 11 03 43 000 > > > > > > > 11 03 51 000 > > > > > > > 11 05 09 000 > > > > > > > 11 05 17 000 > > > > > > > 11 05 25 000 > > > > > > > 11 05 32 000 > > > > > > > 11 05 40 000 > > > > > > > ) > > > > > > > > > > > > > > NB. -------------------- abdomen > > > > > > > > > > > > > > b=:".;._2 (0 : 0) > > > > > > > 11 11 04 000 > > > > > > > 11 11 11 000 > > > > > > > 11 11 19 000 > > > > > > > 11 11 27 000 > > > > > > > 11 11 34 000 > > > > > > > 11 12 46 000 > > > > > > > 11 12 54 000 > > > > > > > 11 13 02 000 > > > > > > > 11 13 10 000 > > > > > > > 11 13 17 000 > > > > > > > 11 14 36 000 > > > > > > > 11 14 44 000 > > > > > > > 11 14 52 000 > > > > > > > 11 15 00 000 > > > > > > > 11 15 07 000 > > > > > > > ) > > > > > > > > > > > > > > NB. ----------------- dcm > > > > > > > > > > > > > > c=:".;._2 (0 : 0) > > > > > > > 11 16 47 450 > > > > > > > 11 16 55 150 > > > > > > > 11 17 02 950 > > > > > > > 11 17 10 600 > > > > > > > 11 17 18 300 > > > > > > > 11 19 25 350 > > > > > > > 11 19 33 200 > > > > > > > 11 19 41 000 > > > > > > > 11 19 48 700 > > > > > > > 11 19 56 450 > > > > > > > 11 22 16 250 > > > > > > > 11 22 24 050 > > > > > > > 11 22 31 750 > > > > > > > 11 22 39 500 > > > > > > > 11 22 47 200 > > > > > > > 11 24 05 650 > > > > > > > 11 24 13 450 > > > > > > > 11 24 21 150 > > > > > > > 11 24 28 850 > > > > > > > 11 24 36 600 > > > > > > > 11 29 59 550 > > > > > > > 11 30 07 350 > > > > > > > 11 30 15 050 > > > > > > > 11 30 22 750 > > > > > > > 11 30 30 450 > > > > > > > 11 31 42 600 > > > > > > > 11 31 50 400 > > > > > > > 11 31 58 100 > > > > > > > 11 32 05 850 > > > > > > > 11 32 13 550 > > > > > > > 11 33 32 400 > > > > > > > 11 33 40 200 > > > > > > > 11 33 47 900 > > > > > > > 11 33 55 600 > > > > > > > 11 34 03 300 > > > > > > > ) > > > > > > > > > > > > > > NB. ------------------ auka > > > > > > > d=:".;._2 (0 : 0) > > > > > > > 0 0 0 0 > > > > > > > 0 0 0 0 > > > > > > > 0 0 0 0 > > > > > > > 0 0 0 0 > > > > > > > 0 0 0 0 > > > > > > > ) > > > > > > > > > > > > > > NB. ------------------- example calculations - the last one > > gives > > > > the > > > > > > > correct lines hooked together > > > > > > > c-a,b,d > > > > > > > 24 60 60 1000 #:((24 60 60 1000 #.c)-24 60 60 1000 #.a,b,d) > > > > > > > > > > > > > > c-~a,b,d > > > > > > > 24 60 60 1000 #:((24 60 60 1000 #.c)-~24 60 60 1000 #.a,b,d) > > > > > > > > > > > > > > c-d,a,b > > > > > > > 24 60 60 1000 #:((24 60 60 1000 #.c)-24 60 60 1000 #.d,a,b) ____________________________________________________________________________________ It's here! Your new message! Get new email alerts with the free Yahoo! Toolbar. http://tools.search.yahoo.com/toolbar/features/mail/ ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
