18.453 Shlomo Argamon on analysis of style, at Haifa

               Humanist Discussion Group, Vol. 18, No. 453.
       Centre for Computing in the Humanities, King's College London
                   www.kcl.ac.uk/humanities/cch/humanist/
                        www.princeton.edu/humanist/
                     Submit to: humanist_at_princeton.edu

         Date: Mon, 03 Jan 2005 06:32:17 +0000
         From: Shuly Wintner <shuly_at_cs.haifa.ac.il>
         Subject: Shlomo Argamon at the University of Haifa

You are cordially invited to attend the following talk.

Time: Tuesday, January 4th, 16:00
Place: University of Haifa, CRI, Education Building, 5th floor
Speaker: Shlomo Argamon, Illinois Institute of Technology.
Title: COMPUTATIONAL ANALYSIS OF MEANINGFUL STYLISTIC VARIATION.

Begin forwarded message:

>COMPUTATIONAL ANALYSIS OF MEANINGFUL STYLISTIC VARIATION
>Shlomo Argamon, Illinois Institute of Technology
>
>ABSTRACT:
>Reliable automated analysis of the linguistic style of a document is
>of manifold potential use for authorship attribution and profiling,
>improved information retrieval, sentiment analysis, and more. Recent
>machine learning techniques have shown reasonable effectiveness in a
>variety of stylistic text classification tasks using features such as
>frequencies of various function words ('to', 'and', 'about', etc.) or
>parts of speech (nouns, verbs, etc.). However, little linguistic or
>cognitive insight is gained through such techniques. We have recently
>been developing linguistically relevant feature for stylistic text
>classification, and analyzing the results for a deeper understanding
>of stylistic variation in different contexts. Our work is within the
>framework of Systemic Functional Linguistics (SFL), which we find
>particularly useful for this task.
>
>In this talk, I will give a brief, non-technical, overview of the
>current state-of-the-art in using machine learning for computational
>stylistics, as well as SFL theory. I will then describe two studies
>we have done in applying SFL to automated stylistic text
>categorization. One is in "sentiment analysis", where we seek to
>classify movie reviews as positive or negative. The other is in the
>analysis of scientific prose in different scientific fields, where we
>show clear differences in language use that can be linked to different
>methodologies in the various fields.
>
>This work is in collaboration with Jeff Dodick (Hebrew University),
>Casey Whitelaw (University of Sydney), and Navendu Garg (Illinois
>Institute of Technology).
>
>SPEAKER:
>Shlomo Argamon (B.Sc. 1988 CMU; M.Phil. 1991, Ph.D. 1994 Yale) is
>currently Associate Professor of Computer Science at the Illinois
>Institute of Technology, where his research focuses mainly on
>automated methods for extracting and analysing stylistic features of
>natural language texts.

_______________________________________________
Iscol mailing list
Iscol_at_cs.haifa.ac.il
https://cs.haifa.ac.il/mailman/listinfo/iscol
Received on Mon Jan 03 2005 - 01:54:08 EST