|
|
  Vahed Qazvinian
Citation Context Extraction
- Introduction: The citation context extraction is an unsupervised method that takes as input a scientific article and extracts the sentences that are about a secondary paper but do not contain an explitict citation to that paper.
- Description: This program models the sentences in an article and their lexical similarities as a Markov Random Field tuned to detect the patterns that context data create, and employ a Belief Propagation mechanism to detect likely context sentences. (See Qazvinian & Radev 2010 for more details).
- Download: For academic use only: ContextExtractor.v1.1.tar.gz
- Dependencies: This code depends two libraries:
- The CLAIRLIB (http://clairlib.org)
- The c_inference package by Talya Meltzer, available here
After installing this package, you'd need to add the path to c_inference by modifying the first two lines of run_infer.m accordingly:
- Usage: To run the code you need to call the following perl command:
%perl extract_context.pl Source Target Explicit_Citations
The arguments are as following:
- Source is the source paper, one sentnece per line
- Target is the cited paper one sentence per line
- Explicit_Citations is a list file in which each line is marked with a '1', if the corresponding sentence in Source is an explicit citation sentence to the Target paper, and '0' otherwise.
Example Usage:
- Please cite this paper when you use this code:
@InProceedings{qazvinian-radev:2010:ACL,
author = {Qazvinian, Vahed and Radev, Dragomir R.},
title = {Identifying Non-Explicit Citing Sentences for
Citation-Based Summarization.},
month = {July},
year = {2010},
booktitle = {Proceedings of the 48th Annual Meeting of
the Association for Computational Linguistics},
address = {Uppsala, Sweden},
publisher = {Association for Computational Linguistics},
pages = {555--564},
url = {http://www.aclweb.org/anthology/P10-1057}
}
|
|