In recent years, political methodologists, have produced innumerable automated
document classification systems. Many of these ignore word sequence
information, treating entire documents as mere collections of words. A subset
of these, including those based on the well-known Naive Bayes algorithm, assume
that word frequencies are independent. In this paper, we investigate the success
of such algorithms by comparing Wordscores, a Naive Bayes derivative, to several
well-known algorithms and a new classification system based on vanilla recursive
heirarchical Dirichlet-multinomial mixture models, pointing out avenues for
future advancement. Surprisingly, we find that its assumptions notwithstanding,
Wordscores shows dramatically increased performance, comparable to some of the
latest developments in document classification, at carrying out a small number
of carefully selected classifications on meticuously arranged collections of
political documents, and discuss its use in practical applications.
Quoting Gary King <king(a)harvard.edu>du>:
this sounds good. i would go a little bit farther in explaining some of
the terms to political scientists who don't know what naive bayes anything
is.
Gary
On Mon, 24 Apr 2006, ghumphr(a)fas.harvard.edu wrote:
Geoff Humphreys and Chris Long
Classfying Political Documents
In recent years, political methodologists, have produced innumerable
automated
document classification systems. Many of these
systems, such as those based
on
the well-known Naive Bayes algorithm, treat each
word as a distinct entity,
ignoring complex interactions between them. While for some applications
this
approach may appear reasonable, the precise
arrangements of words in
political
documents often convey meanings which cannot be
captured so easily. In this
paper, we investigate the success of such naive algorithms by comparing
Wordscores, a Naive Bayes derivative, to several well-known algorithms and
a
new classification system based on vanilla
recursive heirarchical
Dirichlet-multinomial mixture models, pointing out avenues for future
advancement. Surprisingly, we find that the assumptions of Wordscores
notwithstanding, it shows dramatically increased performance, comparable to
some of the latest developments in document classification, at carrying out
a
small number of carefully selected
classifications on meticuously arranged
collections of political documents, and discuss its use in practical
applications.
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l