Paper Info
Title
Tracking news article evolution by dense subgraph learning
Abstract
As an important and challenging problem, effective knowledge discovery from news event evolution over massive news articles plays a critical role in public opinion analysis and social information security. Most existing methods for news knowledge discovery resort to news event detection based on connecting a sequence of news articles over time. However, they usually have to predetermine the temporal path length of news event evolution, which leads to the algorithmic inflexibility in practice. Moreover, they are incapable of well capturing the intrinsic contextual information among the news events, resulting in the performance degradation in noisy data. To address these issues, we propose a context-dependent news knowledge discovery method based on temporally successive news article connection using subgraph learning. The proposed method is able to adaptively construct a cross-article link network along the temporal dimension, and effectively discovers the news event pattern by dense subgraph learning using the contextual news connection structures. Based on the learning structures, we present a fast and accurate link path inference method (i.e., maximum-flow rule and minimal-connection rule). Experimental results on three benchmark datasets demonstrate the effectiveness of the proposed method.
Year
DOI
Venue
2015
10.1016/j.neucom.2015.05.016
Neurocomputing
Keywords
Field
DocType
News tracking,News article evolution,Dense subgraphs
Data mining,Noisy data,Contextual information,Computer science,Inference,Event evolution,Knowledge extraction,Artificial intelligence,Social information,Machine learning
Journal
Volume
Issue
ISSN
168
C
0925-2312
Citations 
PageRank 
References 
3
0.39
14
Authors
5
Name
Order
Citations
PageRank
Shengkang Yu130.39
Xi Li21850137.71
Xueyi Zhao3173.04
Zhongfei (Mark) Zhang42451164.30
Fei Wu52209153.88