- Haskell-Cafe - Haskell.org

Neural networks and optimization in Haskell
by Joel Reymont 28 Sep '05

28 Sep '05

Folks, I got a project where I have a large number of variables and an outcome and I need to figure out which 20% of the variables has the largest effect on the outcome. Of course I also need to optimize the 20% of variables I end up with. This sounds like a job for a neural network to me, with arguments possibly optimized through genetic algorithms. I'm wondering, though, if I'm complicating things for myself and there's an easier approach to this. If not I'm wondering if there are ready-made NN or GA libraries for Haskell. Also, I'm curious if Haskell is really the best language for this type of problem and if lazy evaluation brings any specific advantages to the solution or would be a hindrance. I would welcome any pointers and feedback, yes, someone is actually paying me to do this :-). Thanks, Joel -- http://wagerlabs.com/

2 2

newbe question
by feucht＠uni-koblenz.de 28 Sep '05

28 Sep '05

Hi i can not load program test1 into hugs, but test2 works. Am i missing some special syntax? greetings, Philip -------------- test1 ------------------ foo :: Maybe Int -> Int foo Nothing =-1 foo (Just a)= a -------------- test2 ------------------ foo :: Maybe Int -> Int foo Nothing = -1 -- ^ -- +-- note an extra space foo (Just a)= a

4 7

Help wanted: Lazy multiway zipper with mismached intervals
by Rene de Visser 27 Sep '05

27 Sep '05

Hello, I need to zip together multiple lists. The lists are sorted by date, and each entry in the list represents data for a time interval. The time intervals between the lists may be missmatched from each other. This means that sometimes you don't need to move forward in list, while you move forward in other lists. It might also mean that you need to skip a number of entries foward. If the computation does not need part of the contents of one of the lists, then this list should be ignored until it is needed. As the lists are sorted you never need to go backwards. I found it fairly easy to write a zipper for two lists, where the items from both lists are needed (example binary operator on the payload of the list). However the combination of merging multiple lists together, and ignoring a list in the case it is not needed by the computation leads to very messy code. I though about if there was a away of encapsulating this in a monad, but haven't thought of any way to do this. Another idea is to have a mutable cursor for each list and move these forward as required. But I guess this might be worth avoiding? I think it would be good if somehow I could encapsulate each list, so on a per list basis, and can say given me the current head, or move one forward. But I haven't figured out how to pass the state of the other threads invisibly through. I guess the ony way might be to use the state monad? I guess there can be no simple recursive solution? Rene.

6 9

XMLParse module
by Huong Nguyen 27 Sep '05

27 Sep '05

Hello all, I am reading XMLParse module http://www.germane-software.com/repositories/public/StatusReport/trunk/hxml… But, I am not clear about the purpose of this module and the function of two functions: parseInstance = unfoldr (pRun instanceItem) . pcdataMode parseDTD = foldl (\a b->b a) emptyDTD . unfoldr (pRun dtdItem) . markupMode parseDocument text = case pRun prolog (pcdataMode text) of Just (_, rest) -> unfoldr (pRun instanceItem) rest Nothing -> [ErrorEvent "Error parsing prolog"] -- can't happen. If any knows the meaning of those, please let me know. Thanks a lot. Best regards,

1 0

Re: [Haskell-cafe] Typing problems with basic arithmetic - help!
by Marcin Tustin 26 Sep '05

26 Sep '05

For some reason the following code is producing an error message from ghci that the the patterns are non-exhaustive. Does anyone have any idea why that could be, given that the patterns are, at least in my meaning, provably exhaustive? choosenonuniqueset n (a:r) | (length r) > (n-1) = [ (sort (a:x)) | x <- (choosenonuniqueset (n-1) r)] `union` [ (sort (a:x)) | x <- (choosenonuniqueset n r)] | (length r) == (n-1) = [a:r] | (length r) < (n-1) = [] Error message is: *Main> :reload Compiling Main ( birthday.hs, interpreted ) Ok, modules loaded: Main. *Main> choosenonuniqueset 2 [1..5] *** Exception: birthday.hs:(22,0)-(27,29): Non-exhaustive patterns in function choosenonuniqueset *Main> choosenonuniqueset 5 [1..5] [[1,2,3,4,5]] *Main>

3 3

Re: [Haskell-cafe] Typing problems with basic arithmetic - help!
by Marcin Tustin 26 Sep '05

26 Sep '05

Thanks for this: All I have to do now is fix the fact that my maths is stupidly screwed! On Fri, Sep 23, 2005 at 02:23:10PM -0400, J. Garrett Morris wrote: > Haskell doesn't do automatic numeric conversion for you - you need to > use fromIntegral to go from an integral type (like 2) to a floating > point type (as you'd need for division) and you can use round or trunc > to go the other direction. There's similarly a fromRational function > to convert from rational numbers to just about anything else you'd > need. A couple of applications of those functions - probably mainly > around your divisions - should fix things up. > > /g

2 1

RE: [Haskell-cafe] Template Haskell and Types
by Simon Peyton-Jones 26 Sep '05

26 Sep '05

Hmm. Q is a monad, so I think fail :: Monad m => String -> m a will do the job. 'recover' should catch the exception, and let you try something else. Simon | -----Original Message----- | From: Gracjan Polak [mailto:gracjan@acchsh.com] | Sent: 20 September 2005 10:43 | To: Simon Peyton-Jones | Cc: haskell-cafe(a)haskell.org | Subject: Re: [Haskell-cafe] Template Haskell and Types | | Simon Peyton-Jones wrote: | > design note http://research.microsoft.com/~simonpj/tmp/notes2.ps | | | In the above paper there is something about 'giveUp'. Seems to quite | useful, but there is no such thing in ghc 6.4. | | Where did my giveUp go? And why? | | -- | Gracjan

3 2

Failure and comonads
by Einar Karttunen 26 Sep '05

26 Sep '05

Hello With the resent discussion on comonads on #haskell (mostly due to Uustalo's excellent paper) I am wondering is it possible to model failure with comonads? It seems to me that Reader,Writer and State can be implemented with both monads and comonads. IO can be implemented as a monad but OI wants linear types. Streams make fine comonads. Is there anything corresponding to Maybe or List? - Einar Karttunen

1 0

map in IO
by Tom Hawkins 25 Sep '05

25 Sep '05

What is the best way to perform a map inside IO? I defined the following function, but this must be common enough to warrant something in the standard library. What am I missing? -Tom mapIO :: (a -> IO b) -> [a] -> IO [b] mapIO _ [] = return [] mapIO f (x:xs) = do y <- f x ys <- mapIO f xs return (y : ys)

2 1

RE: Typing problems with basic arithmetic - help!
by Scherrer, Chad 24 Sep '05

24 Sep '05

> For some reason the following code is producing an error > message from ghci that the the patterns are non-exhaustive. > Does anyone have any idea why that could be, given that the > patterns are, at least in my meaning, provably exhaustive? > You forgot to match against the empty list. Try adding choosenonuniqueset n [] = [] -Chad Scherrer

1 0