Presentation 2006-05-18
A study of hot topic extraction from document stream based on probabilistic models
Manabu KIMURA, Kazumi SAITO, Naonori UEDA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we address the task of extracting hot topics from document streams such as a series of news papers. To this end, we present a method based on a bursty appearance of words, as a variant of the Kleinberg's method based on a bursty appearance of documents. In our experiments using news papers during one year, we report that the proposed method showed a better extraction performance on a set of benchmark topics extracted by humans, in comparson to the Kleinberg's original method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) hot topic extraction / document stream / burst / probabilistic model
Paper # AI2006-10
Date of Issue

Conference Information
Committee AI
Conference Date 2006/5/11(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Artificial Intelligence and Knowledge-Based Processing (AI)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A study of hot topic extraction from document stream based on probabilistic models
Sub Title (in English)
Keyword(1) hot topic extraction
Keyword(2) document stream
Keyword(3) burst
Keyword(4) probabilistic model
1st Author's Name Manabu KIMURA
1st Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology()
2nd Author's Name Kazumi SAITO
2nd Author's Affiliation NIPPON TELEGRAPH AND TELEPHONE CORPORATION, NTT Communication Science Laboratories
3rd Author's Name Naonori UEDA
3rd Author's Affiliation NIPPON TELEGRAPH AND TELEPHONE CORPORATION, NTT Communication Science Laboratories
Date 2006-05-18
Paper # AI2006-10
Volume (vol) vol.106
Number (no) 38
Page pp.pp.-
#Pages 6
Date of Issue