Re: ANN Advice

Thanks,

I am actually looking into a Bayesian routine (been using it on my spam for
a year or so and thought it would be good for this kind of thing), but was
also hoping to find a way to show potential documents that are not
necessarily "similar" in terms of word frequency. In the future, the library
will include audio, images and video which may not be classified so easily.

I am especially interested in having a routine which can gain knowledge
about which documents people read, possibly in which order, and how they
browse to them and then with the gained knowledge make predictions for other
users.

I have found:

http://freebsd.mu/freebsd/archives/000039.html

Which has an ANN implemented in PHP - could I tailor this to reach my goals?

Any further thoughts?

Jamie.

> Perhaps you would be also interested in looking into "bayesian learning"
> for "learning how to classify text". Try looking for this keywords on the
> internet.
>
>  cheers, nick
>