基于多种词特征的微博突发事件检测方法

doi:10.3969/j.issn.0372-2112.2019.09.015

电子学报 ›› 2019, Vol. 47 ›› Issue (9): 1919-1928.DOI: 10.3969/j.issn.0372-2112.2019.09.015

基于多种词特征的微博突发事件检测方法

张仰森^1,3, 段宇翔¹, 王建¹, 吴云芳²

1. 北京信息科技大学智能信息处理研究所, 北京 100101;
2. 北京大学计算语言学研究所, 北京 100871;
3. 国家经济安全预警工程北京实验室, 北京 100044

收稿日期:2018-08-13 修回日期:2018-11-20 出版日期:2019-09-25
- 作者简介:
- 张仰森 男,1962年6月出生于山西临猗,博士后,教授,研究方向为中文信息处理、人工智能.E-mail:zhangyangsen@163.com;段宇翔 男,1992年3月出生于山西太原,硕士研究生,研究方向为中文信息处理、突发事件检测.E-mail:duanyx5173@163.com;王建男,1993年5月生于浙江温州,硕士,研究方向为中文信息处理、信息安全.E-mail:455858538@qq.com;吴云芳 女,1973年3月生于山西,博士,副教授,研究方向为语义计算,智能问答.E-mail:wuyf@pku.edu.cn
- 基金资助:
- 国家自然科学基金 (No.61772081）; 科技创新服务能力建设-科研基地建设-北京实验室-国家经济安全预警工程北京实验室项目 (No.PXM2018_014224_000010）

Microblog Bursty Events Detection Method Based on Multiple Word Features

ZHANG Yang-sen^1,3, DUAN Yu-xiang¹, WANG Jian¹, WU Yun-fang²

1. Institute of Intelligent Information Processing, Beijing Information Science and Technology University, Beijing 100101, China;
2. Institute of Computational Linguistics, Peking Universit,, Beijing, 100871, China;
3. Beijing Laboratory of National Economic Security Early-warning Engineering, Beijing 100044, China

Received:2018-08-13 Revised:2018-11-20 Online:2019-09-25 Published:2019-09-25
- Supported by:
- National Natural Science Foundation of China (No.61772081); Capacity Building for Sci-Tech Innovation - Construction of Research Base- Beijing Laboratory - Beijing Laboratory of National Economic Security Early-warning Engineering, Beijing Jiaotong University (No.PXM2018_014224_000010)

摘要/Abstract

摘要： 近年来，各领域内频频发生各类突发事件，对社会稳定发展产生了一定程度的影响.本文提出了一种基于多种词特征的微博突发事件检测模型，可以在海量微博数据中对突发事件进行检测，便于相关决策者进行微博监控和舆论引导，尽可能减少突发事件给社会带来的危害.首先根据时间信息对微博数据进行时间切片，对每一个时间窗口内的数据分别计算各个词语的词频特征、话题标签特征和词频增长率特征；然后基于D-S证据理论和层次分析法，确定词的各个特征权重，并进行加权融合得到词的突发特征值，将突发特征值大的词挑选出来构成突发特征词集，构建基于共现度和结合紧密度的突发事件特征词集的耦合度矩阵；最后将该耦合度矩阵作为凝聚式层次聚类算法的输入，生成一棵由突发词为叶子节点的二叉树，并采用内部相似度的二叉树剪枝算法对聚类结果进行划分，即可实现对相应时间窗口突发事件的检测.实验结果表明，基于突发词的事件检测模型在簇内部相似度阈值等于1.1时效果最好，正确率达到0.8462、召回率达到0.8684、F值为0.8571，表明了本文所提方法的有效性.

关键词: 微博, 突发事件, 突发特征词, D-S证据理论, 凝聚式层次聚类

Abstract: In recent years, a wide variety of bursty events have been occurring frequently in many fields,impacting both the stability and the development of our society. This paper proposes an event detection model based on multiple word features, which is intended to detect bursty events in the massive microblog data. The model will assist decision-makers to monitor microblogs and guide public opinions and will minimize the negative effect of bursty events to society. Firstly, the model slices the microblog data according to the time information. In each time window, the word frequency feature, the topic tag feature and the word frequency growth rate feature of each word are calculated separately. Then, the D-S evidence theory and the analytic hierarchy process are utilized to determine each word's feature weights, which are then merged to obtain the bursty feature value of the word. Words with large bursty feature value are selected to form the bursty feature word set and to construct a coupling degree matrix of bursty feature word set based on co-occurrence degree and tightness. Finally, the coupling degree matrix is used as the input of the hierarchical agglomerative clustering algorithm to generate a binary tree with bursty words being leaf nodes, and the internal similarity binary tree pruning algorithm is used to divide the clustering results.In this way, the detection of the corresponding time window's bursty events can be realized. The experimental results show that the event detection model based on bursty words has the best effect when the intra-cluster similarity threshold is 1.1, the correct rate is as high as 0.8462, the recall rate reaches 0.8684, and the F value is 0.8571, indicating the effectiveness of the proposed method.

Key words: microblog, bursty events, bursty feature words, D-S evidence theory, hierarchical agglomerative clustering

中图分类号:

TP393.2

张仰森, 段宇翔, 王建, 等. 基于多种词特征的微博突发事件检测方法[J]. 电子学报, 2019, 47(9): 1919-1928.

ZHANG Yang-sen, DUAN Yu-xiang, WANG Jian, et al. Microblog Bursty Events Detection Method Based on Multiple Word Features[J]. Acta Electronica Sinica, 2019, 47(9): 1919-1928.

参考文献

[1] Goto J,Miyazaki T,Takei Y,et al.Automatictweet detection based on data specified through news production[A].Proceedings of the 23rd International Conference on Intelligent User Interfaces Companion[C].Tokyo:ACM,2018.1.
[2] 周刚,邹鸿程,熊小兵,等.MB-SinglePass:基于组合相似度的微博话题检测[J].计算机科学,2012,39(10):198-202. ZHOU Gang,ZOU Hongcheng,XIONG Xiaobing,et al.MB-SinglePass:Microblog topic detection based on combined similarity[J].Computer Science,2012,39(10):198-202.(in Chinese)
[3] Qiu YF,Cheng L B.Research onsudden topic detection method for microblog[J].Computer Engineering,2012,38(9):288-290.
[4] Du Y,Wu W,He Y,et al.Microblog bursty feature detection based on dynamics model[A].International Conference on Systems and Informatics[C].Bandung:IEEE,2012.2304-2308.
[5] Salas A,Georgakis P,Petalas Y.Incident detection using data from social media[A] Proceedings of the 20th International Conference on Intelligent Transportation Systems[C].Yokohama:IEEE,2017.751-755.
[6] Schmidt A,Wiegand M.A survey on hate speech detection using natural language processing[A]Proceedings of the Fifth International Workshop on Natural Language Processing for Social Media[C].Boston:Association for Computational Linguistics,2017.1-10.
[7] Kalden J P H.Dataanalysis within the netherlands coastguard:risk mapping,social network analysis and anomaly detection[A].NL ARMS Netherlands Annual Review of Military Studies 2018[C].The Hague:TMC Asser Press,2018.193-200.
[8] 郭跇秀,吕学强,李卓.基于突发词聚类的微博突发事件检测方法[J].计算机应用,2014,34(02):486-490,505. GUO Yixiu,LYU Xueqiang,LI Zhuo.Bursty topics detection approach on Chinese microblog based on burst words clustering[J].Journal of Computer Applications,2014,34(02):486-490,505.(in Chinese)
[9] Unankard S,Li X,Sharaf M A.Emerging event detection in social networks with location sensitivity[J].World Wide Web,2015,18(5):1393-1417.
[10] Quezada M,Pe†a-Araya V,Poblete B.Location-aware model for news events in social media[A]Proceedingsof the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval[C].New York:ACM,2015.935-938.
[11] Krumm J,Horvitz E.Eyewitness:Identifying local events via space-time signals in twitter feeds[A]Proceedings of the 23rd SIGSPATIAL International Conference on Advances in Geographic Information Systems[C].Washington:ACM,2015.20.
[12] Cecaj A,Mamei M.Data fusion for city life event detection[J].Journal of Ambient Intelligence and Humanized Computing,2017,8(1):117-131.
[13] Jinhua L,Zhongjie A.Analyzing geographical coordinates data for micro-blog trending events[J].Data Analysis and Knowledge Discovery,2016,32(2):90-101.
[14] Hua T,Chen F,Zhao L,et al.Automatic targeted-domain spatiotemporal event detection in twitter[J].GeoInformatica,2016,20(4):765-795.
[15] Wakamiya S,Jatowt A,Kawai Y,et al.Analyzingglobal and pairwise collective spatial attention for geo-social event detection in microblogs[A]Proceedings of the 25th International Conference Companion on World Wide Web[C].Montreal:WWW,2016.263-266.
[16] Zheng F R,Miao D Q,Zhang Z F,et al.News topic detection approach on Chinese microblog[J].Computer science,2012,39(1):138-141.
[17] 张仰森,郑佳,唐安杰.基于多特征融合的微博用户权威度定量评价方法[J].电子学报,2017,45(11):2800-2809. Zhang Yangsen,Zheng Jia,Tang Anjie.A quantitative evaluation method of micro-blog user authority based on multi-feature fusion[J].Acta Electrinica Sinica,2017,45(11):2800-2809.(in Chinese)

基于多种词特征的微博突发事件检测方法

Microblog Bursty Events Detection Method Based on Multiple Word Features

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	程晓涛, 吉立新, 尹赢, 黄瑞阳. 基于D-S证据理论的网络表示融合方法[J]. 电子学报, 2020, 48(5): 854-860.
[2]	陈维高, 贾鑫, 朱卫纲, 唐晓婧. 基于机动特征辅助的MFR状态预测方法[J]. 电子学报, 2018, 46(6): 1404-1409.
[3]	柏文言, 张闯, 徐克付, 张志明. 一种融合用户关系的自适应微博话题跟踪方法[J]. 电子学报, 2017, 45(6): 1375-1381.
[4]	陈杰, 刘学军, 李斌, 章玮. 一种基于用户动态兴趣和社交网络的微博推荐方法[J]. 电子学报, 2017, 45(4): 898-905.
[5]	马慧芳, 谢蒙, 何廷年, 蔺想红. 基于核心标签的可重叠微博网络社区划分方法[J]. 电子学报, 2017, 45(4): 769-776.
[6]	王伟, 张效尉, 任国恒, 秦东霞, 刘琳琳. 基于能量优化的微博用户转发行为预测[J]. 电子学报, 2017, 45(12): 2987-2996.
[7]	张仰森, 郑佳, 唐安杰. 基于多特征融合的微博用户权威度定量评价方法[J]. 电子学报, 2017, 45(11): 2800-2809.
[8]	马慧芳, 贾美惠子, 张迪, 蔺想红. 融合标签关联关系与用户社交关系的微博推荐方法[J]. 电子学报, 2017, 45(1): 112-118.
[9]	刘全超, 黄河燕, 冯冲. 面向中文微博的评价对象与评价词语联合抽取[J]. 电子学报, 2016, 44(7): 1662-1670.
[10]	于岩, 陈鸿昶, 于洪涛. 基于霍克斯过程的社交网络用户关系强度模型[J]. 电子学报, 2016, 44(6): 1362-1368.
[11]	朱江, 王柏, 吴斌, 李小明. 一种微博用户情感影响者发现模型[J]. 电子学报, 2015, 43(12): 2497-2504.
[12]	王佰玲, 曲芸, 张永铮, 田志宏. 基于数据流的网页内容分析技术研究[J]. 电子学报, 2013, 41(4): 751-756.
[13]	郎风华;谷利泽;杨义先;钮心忻. 一种基于均衡交补分担准则的证据组合新方法[J]. 电子学报, 2009, 37(1): 95-100.
[14]	朱大奇;于盛林. 基于D-S证据理论的数据融合算法及其在电路故障诊断中的应用[J]. 电子学报, 2002, 30(2): 221-223.
[15]	孙全;叶秀清;顾伟康. 一种新的基于证据理论的合成公式[J]. 电子学报, 2000, 28(8): 117-119.