Re^2: Reverse engineering a formula...

in reply to Re: Reverse engineering a formula...
in thread Reverse engineering a formula...

I was trying to keep my question "theoretical". But here's the numbers if you really want them:

$VAR1 = [
          '18.930167',
          '17.967469',
          '0.008720',
          '0.008720',
          '122.640000',
          '12493.320000',
          '359.520000',
          '288.700000',
          '359.520000',
          '89.880000',
          '32.960000',
          '56.920000',
          '13.470000',
          '1231.360000',
          '20587.440000',
          '359.520000',
          '792.170000',
          '629.160000',
          '972.290000',
];
[download]

- Devnul

Comment on Re^2: Reverse engineering a formula... Download Code

Replies are listed 'Best First'.
Re^3: Reverse engineering a formula... by BrowserUk (Patriarch) on Jun 13, 2006 at 12:02 UTC
How's this for a guess? A 24-month lease `An initial price of 12,493.32 A first payment of 1,231.36 + 23 monthly payements of 843.24 ===================================== Total cost (- tax relief) 20,587.44 Daily interest rate 0.008720 Annual Equivalent rate 17.967469 Annual effective rate (inc. one off payments) 18.930167` [download] Other numbers include one-off admin charges and daily/weekly effective costs? Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error. Lingua non convalesco, consenesco et abolesco. -- Rule 1 has a caveat! -- Who broke the cabal? "Science is about questioning the status quo. Questioning authority". In the absence of evidence, opinion is indistinguishable from prejudice.	[reply] [d/l]
Re^3: Reverse engineering a formula... by Gilimanjaro (Hermit) on Jun 13, 2006 at 12:27 UTC
Well, look at it this way: You have 19 floats Each float may: Need to be added (multiply by 1) Need to be subtracted (multiply by -1) Need to be ignored (multiply by 0) This means there are 3 to the power 19 different ways to combine these numbers... That's a lot of possible combinations to try, but it can be done I suppose... You'd probably want to write a recursive algorithm to try all these permutations, and see if the rounded total matches the number you specified... Update: Possible solution code posted below	[reply]
Re^3: Reverse engineering a formula... by Skeeve (Parson) on Jun 13, 2006 at 10:31 UTC
That will never give the sum you told us, because of 18.930167 and 17.96746 You will never get aything with just 2 decimal digits behind the decimal point with those 2 numbers. `s$$([},&%#}/&/]+}%&{});#$&&s&&$^X.($'^"%]=\&(\|?{%` `+`.+=%;.#_}\&"^"-+%*).}%:##%}={~=~:.")&e&&s""`$''`"e	[reply] [d/l] [select]
Re^4: Reverse engineering a formula... by Gilimanjaro (Hermit) on Jun 13, 2006 at 13:58 UTC
And as the information is financial in nature, the end-result is most probably rounded to two digits. So in fact these two numbers may well be part of the formula... In fact, without knowing for sure how exactly rounding is done, you'll have to look for a 'fuzzy' match; the last few digits may be off. Perhaps rounding the target and the tested result to integers would be the best way to go about this. On a side note. the two numbers below 0.01 are most probably percentages... So we may also need to do multiplication, which would make it an almost impossible task, mainly because this would mean we also have to take operator-precedence into account...	[reply]
Re^4: Reverse engineering a formula... by monkey_boy (Priest) on Jun 13, 2006 at 10:34 UTC
With the info we have, there is no guarantee that all the columns are used. This is not a Signature...	[reply] [d/l]
Re^4: Reverse engineering a formula... by devnul (Monk) on Jun 13, 2006 at 10:34 UTC
Not every number necessarily needs to be part of the formula (obviously its 2 or more numbers in some combination of addition and/or subtraction). Those two numbers could simply not be part of the formula. - Devnul	[reply]
Re^5: Reverse engineering a formula... by Skeeve (Parson) on Jun 13, 2006 at 10:41 UTC
As I said before: Not enough information... It seems as if your formular is something like this (untested code of course): `my @sign=( .... ); # Some combination of -1, 0 1 my @sum= map { $sign[$i] * $VAR1[$i] } (0..$#VAR1); print eval join ' + ', @sum;` [download] Now you "only" need to get the correct combination of -1, 0 and 1 for @sign, which is a "simple" combinational task but will be very lengthy... `s$$([},&%#}/&/]+}%&{});#$&&s&&$^X.($'^"%]=\&(\|?{%` `+`.+=%;.#_}\&"^"-+%*).}%:##%}={~=~:.")&e&&s""`$''`"e	[reply] [d/l] [select]

In Section Seekers of Perl Wisdom