Re: Neural net - guessing literature kind

From: Will Dwinnell (predictr_at_bellatlantic.net)
Date: 11/29/03 

Date: 29 Nov 2003 09:12:52 -0800

Borys Musielak <michuk@poczta.onet.pl> wrote:
"I have to write an algorithm which guesses the literature kind having
100 words of text."

Does that mean the first 100 words, specifically? Also, you'll need
to be specific about what kinds (literature, technical, poetry, news
copy, etc.) and how many kinds (2, 3, 12?) of literature you want to
distinguish.

Borys Musielak <michuk@poczta.onet.pl> continues:
"Of course the natural way to design it is by using a neural net and
back propagation."

I don't see why- there are many classification systems, both neural
and non-neural. There's nothing wrong with this choice, but I don't
understand why it is "the natural way to design it".

Borys Musielak <michuk@poczta.onet.pl> continues:
"The question is - how to design the net and what function to use to
make it work well. And even more important task is: what does really
differentiate poems from fiction, or criminals in terms of words (i
think of using some syntax loops like verses with rhymes).
Enough would be to differ poetry from prose. Better but i dont know if
easy would be guessing prose kind: like love story, whodoneit, fantasy
etc. I dont expect solutions, I rather need some ideas and maybe
links to similar projects.
I'd be very grateful for any advice."

Filling in the above details would be a good start. All such
projects, though, share the common element of needing appropriate
features by which to distinguish classes. Designing these will need
to proceed from the aforementioned details.

-Will Dwinnell, MBA
http://will.dwinnell.com

Relevant Pages

  • Re: Smalltalk Processor
    ... Of course, projects such as SOAR, which spawned Sun's SPARC, suggested that an efficient implementation could be had with only minimal modifications to a general purpose RISC. ... Don't expect to find much revolutionary in the literature. ... Most design are pretty straightforward, but the good news is that you can quite easily implement your own with even an entry-level FPGA development kit. ... Smalltalk processor(a silicon implementation of the virtual machine - ...
    (comp.arch)
  • Re: measuring benefits from automating tests
    ... So you will have to design experiments to do the evaluation that are tailored to your shop. ... Any of the existing literature on process improvement will have guidelines for doing such small scale experiments. ... Here the body of literature on metrics programs should be helpful since designing good metrics requires quantification. ... doing what you want to do well requires substantial experience in at least three distinct disciplines (process improvement, experimental design, and metrics design). ...
    (comp.software.testing)
  • Re: NAT Secure?
    ... These were basic design points. ... > assumptions is why NAT causes so many protocols to fail. ... > There's actually substantial literature about this. ... and why your network equipment is ...
    (comp.security.firewalls)
  • Re: Human design and natural "design"
    ... > If you are referring to biological evolution literature, ... > around each other for lack of clear definitions (design? ... Feel free to provide definitions for design, random, feedback, model and ...
    (talk.origins)
  • Re: Genes and the Mind
    ... I really dont know what you mean by "swaying" someone with a study. ... genes or environment. ... because it is part of a body of literature ...
    (uk.philosophy.humanism)