Re: [Haskell-cafe] ANNOUNCE: repr-0.3.2
Did you consider using the traced package? 2009/12/24 Bas van Dijk : > On Thu, Dec 24, 2009 at 1:00 AM, Andrey Sisoyev > wrote: >> Where do you make use of it? :) > > A few months ago I was working on 'levmar'[1] a Levenberg-Marquardt > data fitting library in Haskell. If you want your data fitting to be > really fast you need to supply a Jacobian of the model function you > want to fit. A Jacobian describes the partial derivatives of the > parameters of the model function. I used Conal Elliott's > vector-space[2] library to automatically derive a Jacobian from the > model function. > > I was interested in the derivatives vector-space would come up with. A > derivative however, is just a function so the only thing you can do > with it is apply it to a value. This then yields a result which is > usually just a Double. I wasn't really interested in this actual > numeric result but more in the underlying numeric expressing that > generated that result. > > So I wrote repr to visualize this numeric expression. However, after I > wrote repr, the work on levmar stalled a bit and I did not actually > get around to applying repr to the derivatives. I plan to make a new > release of levmar in the coming weeks or so and I think I will use > repr then. > > regards, > > Bas > > [1] http://hackage.haskell.org/package/levmar > [2] http://hackage.haskell.org/package/vector-space > ___ > Haskell-Cafe mailing list > Haskell-Cafe@haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] ANNOUNCE: repr-0.3.2
On Thu, Dec 24, 2009 at 1:00 AM, Andrey Sisoyev wrote: > Where do you make use of it? :) A few months ago I was working on 'levmar'[1] a Levenberg-Marquardt data fitting library in Haskell. If you want your data fitting to be really fast you need to supply a Jacobian of the model function you want to fit. A Jacobian describes the partial derivatives of the parameters of the model function. I used Conal Elliott's vector-space[2] library to automatically derive a Jacobian from the model function. I was interested in the derivatives vector-space would come up with. A derivative however, is just a function so the only thing you can do with it is apply it to a value. This then yields a result which is usually just a Double. I wasn't really interested in this actual numeric result but more in the underlying numeric expressing that generated that result. So I wrote repr to visualize this numeric expression. However, after I wrote repr, the work on levmar stalled a bit and I did not actually get around to applying repr to the derivatives. I plan to make a new release of levmar in the coming weeks or so and I think I will use repr then. regards, Bas [1] http://hackage.haskell.org/package/levmar [2] http://hackage.haskell.org/package/vector-space ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: [Haskell-cafe] ANNOUNCE: repr-0.3.2
Where do you make use of it? :) Andrey -- View this message in context: http://old.nabble.com/ANNOUNCE%3A-repr-0.3.2-tp26908749p26909005.html Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com. ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
[Haskell-cafe] ANNOUNCE: repr-0.3.2
Hello, Some months ago I uploaded a little package called 'repr' to hackage. I've now updated the package to work with ghc-6.12.1 and its new base library 4.2.0.0. Back then I forgot to make a proper announcement so I will do that now: 'repr' allows you to render overloaded expressions to their textual representation. For example: *Repr> let rd = 1.5 + 2 + (3 + (-4) * (5 - pi / sqrt 6)) :: Repr Double *Repr> show rd "fromRational (3 % 2) + 2 + (3 + negate 4 * (5 - pi / sqrt 6))" See: http://hackage.haskell.org/package/repr-0.3.2 regards, Bas ___ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
