A Method for Constructing Open-Domain Chinese Entity Hypernym Hierarchical Structure

doi:10.3969/j. issn.1673-5188.2017.01.008

ZTE Communications ›› 2017, Vol. 15 ›› Issue (1): 49-54.DOI: 10.3969/j. issn.1673-5188.2017.01.008

收稿日期:2015-11-05 出版日期:2017-02-25 发布日期:2019-12-24

A Method for Constructing Open-Domain Chinese Entity Hypernym Hierarchical Structure

CAI Hongbo¹, CHEN Hong², LIU Shen¹

1.Harbin Institute of Technology, Harbin 150001, China
2.ZTE Corporation, Nanjing 210012, China

Received:2015-11-05 Online:2017-02-25 Published:2019-12-24
About author:CAI Hongbo (hbcai@ir.hit.edu.cn) received his master’s degree from School of Computer Science and Technology, Harbin Institute of Technology, China. He is an engineer at Alibaba Corporation. His main research interest is entity relation graphs.|CHEN Hong (chen.hong3@zte.com.cn) received her B.S. degree from the Department of Information Engineering, Nangjing University of Posts and Telecommunications, China in 2007. She is a senior research engineer at ZTE Corporation. Her research interests incloud social network analysis and intelligent question answering. She holds five patents .|LIU Shen (sliu@ir.hit.edu.cn) received his master’s degree from School of Computer Science and Technology, Harbin Institute of Technology, China. He is an engineer at Miaozhen System. His main research interest is entity relation extraction.
Supported by:
This work was supported by ZTE Industry-Academia-Research Cooperation Funds

摘要/Abstract

Abstract:

Entity relation is an essential component of some famous knowledge bases, such as Freebase, Yago and Knowledge Graph, while the hyponymy plays an important role in entity relations that show the relationship between the more general terms (hypernyms) and the more specific instances of the terms (hyponyms). In this paper, we present a comprehensive scheme of open-domain Chinese entity hypernym hierarchical construction. Some of the most important unsupervised and heuristic approaches for building hierarchical structure are covered in sufficient detail along with reasonable analyses. We experimentally evaluate the proposed methods and compare them with other baselines. The result shows high precision of our method and the proposed scheme will be further improved with larger scale corpora.

Key words: Entity hypernym, hierarchization, Apriori Algorithm, suffix hypernym, open-domain

. [J]. ZTE Communications, 2017, 15(1): 49-54.

CAI Hongbo, CHEN Hong, LIU Shen. A Method for Constructing Open-Domain Chinese Entity Hypernym Hierarchical Structure[J]. ZTE Communications, 2017, 15(1): 49-54.

图/表 10

参考文献 10

[1]	Q. Tang, F. Gao, J. Wang , “Knowledge map concepts analysis and research,” Information Studies: Theory and Application, vol. 34, no. 1, pp. 121-125, 2011.
[2]	L. Qian, G. Zhou, F. Kong, Q. Zhu, P. Qian , “Exploiting constituent dependencies for tree kernel-based semantic relation extraction,” in Proc. 22nd International Conference on Computational Linguistics, Manchester, UK, Aug. 2008, pp. 697-704.
[3]	A. Fader, S. Soderland, O. Etzioni , “Identifying relations for open information extraction,” in Proc. Conference on Empirical Methods in Natural Language Processing, Edinburgh, UK, Jul. 2011, pp. 1535-1545.
[4]	M. Banko, M. J. Cafarella, S. Soderland, M. Broadhead and O. Etzioni , “Open information extraction from the web,” in 20th International Joint Conference on Artificial Intelligence, Hyderabad, India, Jan. 2007, pp. 2670-2676.
[5]	W. Che, T. Liu, S. Li , “Automatic entity relation extraction,” in The First National Conference on Information Retrieval and Content Security (NCIRCS’2004), Shanghai, China, 2004, 1-6.
[6]	C. Chen, Y. Chen, J. Hou, Y. Liang , “CiteSpace II: detecting and visualizing emerging trends and transient patterns in scientific literature,” Journal of the China Society for Scientific and Technical Information vol. 28, no. 3, pp. 401-421, 2009.
[7]	R. Fu, B. Qin, T. Liu , “Exploiting multiple sources for open-domain hypernym discovery,” in Proc. 2013 Conference on Empirical Methods in Natural Language Processing (EMNLP), Seattle, USA, 2013, pp. 1224-1234.
[8]	R. Fu , “Open-domain named entity recognition and hierarchical category acquisition,” Ph.D. dissertation, Harbin Institute of Technology, Harbin, China, 2014.
[9]	W. Che, Z. Li, T. Liu , “LTP: a Chinese language technology platform,” in Proc. 23rd International Conference on Computational Linguistics, Demonstrations Volume, Beijing, China, Aug. 2010, pp. 13-16.
[10]	C. Friedman, G. Hripcsak, W. DuMouchel , et al., “Natural language processing in an operational clinical information system,” Natural Language Engineering, vol. 1, no. 1, pp. 83-108, 1995. DOI URL

Entity	Hypernym
百金花 (centaury)	植物 (plant)
百金花 (centaury)	花 (flower)
百金花 (centaury)	中药 (traditional Chinese medicine)
日本角鲨 (Squalus japonicus)	生物 (living thing)
日本角鲨 (Squalus japonicus)	动物 (animal)
黄色白茧蜂 (Phanerotoma flava Ashmead)	昆虫 (insect)

Entity	Hypernym
百金花 (centaury)	植物 (plant)
百金花 (centaury)	花 (flower)
百金花 (centaury)	中药 (traditional Chinese medicine)
日本角鲨 (Squalus japonicus)	生物 (living thing)
日本角鲨 (Squalus japonicus)	动物 (animal)
黄色白茧蜂 (Phanerotoma flava Ashmead)	昆虫 (insect)

Hypernym A	The hypernym of A	Confidence	Support
川菜 (Sichuan Cuisine)	饮食 (diet)	1	38
石竹亚纲 (Caryophyllidae)	植物 (plant)	0.968994	64
种子植物门 (Spermatophyta)	植物 (plant)	0.977477	1501
石竹目 (Caryophyllales)	种子植物门 (Spermatophyta)	1	63
鸟类 (bird)	动物 (animal)	1	58
冬青属 (ilex L.)	双子叶植物纲 (Dicotyledoneae)	0.907029	21

Hypernym A	The hypernym of A	Confidence	Support
川菜 (Sichuan Cuisine)	饮食 (diet)	1	38
石竹亚纲 (Caryophyllidae)	植物 (plant)	0.968994	64
种子植物门 (Spermatophyta)	植物 (plant)	0.977477	1501
石竹目 (Caryophyllales)	种子植物门 (Spermatophyta)	1	63
鸟类 (bird)	动物 (animal)	1	58
冬青属 (ilex L.)	双子叶植物纲 (Dicotyledoneae)	0.907029	21

Variables and functions	Definition
Confidence (A, B)	The probability that B is one hypernym of A for pre-estimation accuracy.
Support (A, B)	The statistical support that B is one hypernym of A.
Count (A, B)	The number of times A and B co-occur in the same entity hypernym set.
Count (A)	The number of times A occurs in all the entity hypernym set.

A Method for Constructing Open-Domain Chinese Entity Hypernym Hierarchical Structure

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 10

相关文章 0

编辑推荐

Metrics

Parameter	Definition	Value
C₁	Forward direction confidence	0.9
C₂	Backward direction confidence	0.8
S₁	Minimum support	5
S₂	Basic support	10

	Precision	The relation number (no indirect edge)	Percentage of number increase
No optimization	97.0% (±1%)	3396	-
Linear optimization	97.0% (±1%)	5422	60.00%
Logarithmic optimization	84.5% (±1%)	4073	19.90%