id: 05730842
dt: a
an: 05730842
au: Chen, Wei; Wang, Can; Chen, Chun; Zhang, Lijun; Bu, Jiajun
ti: Topic decomposition and summarization.
so: Zaki, Mohammed J. (ed.) et al., Advances in knowledge discovery and data
    mining. 14th Pacific-Asia conference, PAKDD 2010, Hyderabad, India,
    June 21‒24, 2010. Proceedings, Part I. Berlin: Springer (ISBN
    978-3-642-13656-6/pbk). Lecture Notes in Computer Science 6118. Lecture
    Notes in Artificial Intelligence, 440-448 (2010).
py: 2010
pu: Berlin: Springer
la: EN
cc: 
ut: Non-negative Matrix Factorization; Topic Decomposition; Topic
    Summarization; Singular Value Decomposition
ci: 
li: doi:10.1007/978-3-642-13657-3_47
ab: Summary: In this paper, we study topic decomposition and summarization for
    a temporal-sequenced text corpus of a specific topic. The task is to
    discover different topic aspects (i.e., sub-topics) and incidents
    related to each sub-topic of the text corpus, and generate summaries
    for them. We present a solution with the following steps: (1) deriving
    sub-topics by applying Non-negative Matrix Factorization (NMF) to
    terms-by-sentences matrix of the text corpus; (2) detecting incidents
    of each sub-topic and generating summaries for both sub-topic and its
    incidents by examining the constitution of its encoding vector
    generated by NMF; (3) ranking each sentences based on the encoding
    matrix and selecting top ranked sentences of each sub-topic as the text
    corpus’ summary. Experimental results show that the proposed topic
    decomposition method can effectively detect various aspects of original
    documents. Besides, the topic summarization method achieves better
    results than some well-studied methods.
rv: