readme.texi

良好的代码实现

@chapter Bag Of Words Library README

@c set the vars BOW_VERSION
@include version.texi

@samp{libbow}, version @value{BOWVERSION}.

@include libbow-desc.texi


@section Rainbow

@samp{Rainbow} is a standalone program that does document
classification.  Here are some examples:

@itemize @bullet

@item

@example
rainbow -i ./training/positive ./training/negative
@end example

Using the text files found under the directories
@file{./positive} and @file{./negative},
tokenize, build word vectors, and write the resulting data structures
to disk.

@item

@example
rainbow -q ./testing/254
@end example

Tokenize the text document @file{./testing/254}, and classify it,
producing output like:

@example
/home/mccallum/training/positive 0.72
/home/mccallum/training/negative 0.28
@end example

@item

@example
rainbow -t 5
@end example

Perform 5 trials, each consisting of a test/train split, a resetting of
weights according to the new split, and outputs of the classification of
the test documents.

@end itemize

Typing @samp{rainbow --help} will give list of all rainbow options.

After you have compiled @samp{libbow} and @samp{rainbow}, you can run
the shell script @file{./demo/script} to see an annotated demonstration
of the classifier in action.

The web page
http://www.cs.cmu.edu/afs/cs.cmu.edu/project/theo-11/www/naive-bayes.html
has a pointer to a ``Gentle Introduction to Rainbow'', as well as some
sample UseNet text data.

@format
Rainbow improvements coming soon:
   Better documentation.
   Better modularily of command-line options for changing parameters
     of weight-setting methods.
   Incremental model training.
   Better smoothing.  Good-Turing estimates, etc.
@end format



@section Arrow

@samp{Arrow} is a standalone program that does document retrieval.
Sorry, there is no documentation yet.