Rudi Cilibrasi writes: "In recent years, it has become possible to get tremendous amounts of
computational power cheaply on home desktops. This has opened up
the field of artificial intelligence and pattern recognition research
to a much wider audience. Another important synergy is the rapid
growth of open source scientific software in the past five years. Only
recently has the confluence of these two forces combined to create
wonderfully easy new machine learning opportunities for developers everywhere.
Over a year ago, we reported on Slashdot the results of new
involving data compression programs. Scientists had been
experimenting with language analysis, spam detection, genetic analysis,
and much more. Now, the software used to do the machine learning
experiments and pattern recognition has been refined to version 0.9.6 and is
very near its 1.0 release. The CompLearn development team would like to take
this opportunity to welcome the Slashdot community to have a look at the
libcomplearn software library and see if it suits your expanding needs
as software developers. Does it allow you to quickly solve the
data mining and pattern recognition applications that interest you?
As we get ready to release 1.0 we are particularly in need of API feedback.
If you have ever been interested in artificial intelligence,
statistics without calculus, or foundational math, then please enjoy
our software library offering and research. We have taken great pains to
make the software as easy to install as possible; if you are
familiar with programs like bzip2 or gzip then you can already combine
data compression programs with Neural Networks, Support Vector Machines,
or other machine learning algorithms to create powerful recognition systems
quickly and easily. We have used CompLearn with the open source LIBSVM
library to create powerful recognizers in a variety of languages and invite
you to experiment with this new type of algorithm.
I look forward to seeing what sorts of applications develop."