NARCL 0.4.1

[ Posted by Urban Hafner Wed, 10 Aug 2005 11:47:00 GMT ]

This release just adds a better README, actually a README that should be useable.

Download: narcl-0.4.1.tar.bz2

Have fun.

Tags , , , , , ,

NARCL 0.4.0

[ Posted by Urban Hafner Sun, 07 Aug 2005 17:07:39 GMT ]

This is the release of version 0.4.0 of NARCL.

The tarball can be found here: narcl-0.4.0.tar.bz2.

The darcs repository is at http://www.cip.ifi.lmu.de/~hafner/darcs/narcl

What’s new?

This release is almost a complete rewrite of the whole library. The main purpose was the speedup of the rule mining. I have come quite far in this direction but I am by no means finished with it.

The aim is to allow association rule mining of data sets with at least 125,000 transactions and about the same number of items in reasonable time.

Have fun with the program and of course feedback is welcome.

Tags , , , , , ,

Fico 0.1.1

[ Posted by urban Mon, 11 Apr 2005 19:29:58 GMT ]

This release is just a point release that makes the following changes:

  • Added script to run tournaments of AIs
  • Fixed bug that prevented the Board class from recognizing certain finished board positions.
  • Added support for draws

As before the tarball can be gotten from bettong.net/fico/ and the latest development version from the darcs repository at http://www.cip.ifi.lmu.de/~hafner/darcs/fico .

Tags , , , , ,

Fico Version 0.1.0

[ Posted by urban Sun, 10 Apr 2005 14:17:54 GMT ]

Fico is an implementation of the board game Havannah. It is completley written in Ruby.

The aim is to write a fairly strong AI for the game. At the moment the only features are that the program can play according to the rules and that a random player exists.

How to get it

There are two ways to get the program.

  1. Get fico-0.1.0.tar.bz2, the latest release tarball (generally new releases can be found at bettong.net/fico/)
  1. Get the latest development from the darcs repository: darcs get http://www.cip.ifi.lmu.de/~hafner/darcs/fico.

How to run it

To run fico against itself on a board of size 10 run the command ./fico in the toplevel directory. Run ./fico -h to see all possible command line options.

If you want to play against the program use the -b or -w switch to select your colour. You will then be asked to enter your moves in the terminal. Be sure to enter them in the form <char><number> (e.g. a1 or c4).

Hacking

You are of course welcome to write your own AI. But don’t expect to much (yet :)) from the framework that binds your bot into the rest of the application.

How to build your own AI

Look at src/player/RandomPlayer.rb for a simple AI. The necessary steps to get your own bot running are:

  • create a sublcass of ComputerPlayer
  • include TermComputerPlayer
  • define the method internal_generate_move. This method should return the next move of your AI.
  • patch src/fico.rb so that it uses your AI instead of RandomPlayer or Simple1Ply

Contact

You can contact me at urban@bettong.net. Watch this place for announcements of new versions.

Tags , , , , ,

First release of NARCL

[ Posted by urban Sat, 02 Apr 2005 12:31:36 GMT ]

This is the announcement for NARCL version 0.1.0.

What is NARCL?

NARCL stands for Negative Association Rules in Common Lisp. As the name suggests it is a program to find association rules and it is written in Common Lisp.

Why did you write it?

At the time where I started writing it no program existed that could compute negative association rules. But I may eventually need a program like that in the future.

How mature is the program?

Well, its version is 0.1.0 so you can imagine. What works is the computation of negative and positive association rules on a given set of transactions. But it is very slow, due to the use of rather inefficient data structures. Nonetheless it might be useful to some people, so here it is.

What license does it have?

The code is licensed under the LLGPL which is essentially the LGPL with some Lisp specific additions and clarifications.

Where can I download it?

You can download the source as a compressed tar ball from my web site:

Tags , , , , , ,