Multi-label Classification for Past Events

被引：0

作者：

Sumikawa, Yasunobu ^{[1
]}

Ikejiri, Ryohei ^{[2
]}

机构：

[1] Tokyo Metropolitan Univ, Univ Educ Ctr, Tokyo, Japan

[2] Univ Tokyo, Interfac Initiat Informat Studies, Tokyo, Japan

来源：

2018 IEEE/WIC/ACM INTERNATIONAL CONFERENCE ON WEB INTELLIGENCE (WI 2018) | 2018年

关键词：

Multi-label classification; document classification; history; event;

D O I：

10.1109/WI.2018.00-37

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Study and analysis of past events can provide numerous benefits. While event categorization has been previously studied, it was usually assigned only one event category to an event. In this work we focus on multi-label classification for past events that is a more general and challenging problem than the previous studies. We categorize them into 13 event categories using a range of diverse features and report micro-average F-1 score is improved approximately by 10% compared with the state-of-the-art algorithm.

引用

页码：562 / 567

页数：6

共 24 条

[1]

[Anonymous], 2013, NAACL

[2]

[Anonymous], 2006, TEACHING HIST

[3]

[Anonymous], 2014, P INT C INT C MACH L

[4]

AuYeung C.-M., 2011, CIKM 11, P1231, DOI [10.1145/2063576.2063755, DOI 10.1145/2063576.2063755]

[5] Latent Dirichlet allocation [J].

Blei, DM ;

Ng, AY ;

Jordan, MI .

JOURNAL OF MACHINE LEARNING RESEARCH, 2003, 3 (4-5) :993-1022

[6]

DEERWESTER S, 1990, J AM SOC INFORM SCI, V41, P391, DOI 10.1002/(SICI)1097-4571(199009)41:6<391::AID-ASI1>3.0.CO

[7]

2-9

[8]

Ikejiri R., 2016, J ED RES SOCIAL STUD, V84, P37, DOI DOI 10.20799/JERASSKENKYU.84.0_37

[9] Towards a Complete Event Type Taxonomy [J].

Kosmerlj, Aljaz ;

Belyaeva, Evgenia ;

Leban, Gregor ;

Fortuna, Blaz ;

Grobelnik, Marko .

WWW'15 COMPANION: PROCEEDINGS OF THE 24TH INTERNATIONAL CONFERENCE ON WORLD WIDE WEB, 2015, :899-902

[10]

Kudo T., EMNLP 04, P230

← 1 2 3 →