《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (5): 1330-1338.DOI: 10.11772/j.issn.1001-9081.2021040654
所属专题: 人工智能
收稿日期:2021-04-25
									
				
											修回日期:2021-07-10
									
				
											接受日期:2021-07-14
									
				
											发布日期:2022-06-11
									
				
											出版日期:2022-05-10
									
				
			通讯作者:
					仇丽青
							作者简介:仇丽青(1978—),女,山东德州人,副教授,博士,主要研究方向:社交网络、数据挖掘 qiuliqing2019@163.com基金资助:Received:2021-04-25
									
				
											Revised:2021-07-10
									
				
											Accepted:2021-07-14
									
				
											Online:2022-06-11
									
				
											Published:2022-05-10
									
			Contact:
					Liqing QIU   
							About author:QIU Liqing, born in 1978, Ph. D., associate professor. Her research interests include social network,data mining.Supported by:摘要:
针对突发事件中负面网络舆情传播的问题,提出了一种基于情感分析和影响力评估的突发事件情感图谱研究方法。提出了一种基于多头自注意力机制和双向长短期记忆网络(Bi-LSTM)的情感分析模型来计算网站用户的情感倾向,并提出了一种融合加权度与K-shell值的节点影响力评估算法来评估用户的影响力,从而综合构建突发事件的情感图谱,有效提高了情感图谱的准确性和科学性。以“7.7安顺公交车坠湖事件”为例,将突发事件的生命周期划分为爆发期、蔓延期、成熟期和衰退期四个阶段,分别生成情感图谱进行可视化分析。实验结果表明,在酒店评论数据集上,所提出的情感分析模型的F1值在积极和消极方面比文本循环神经网络(Text-RNN)模型分别提升了9.92个百分点和2.5个百分点;在Karate网络上,所提影响力评估算法的区分度和准确性比K-shell算法分别提升了46.89个百分点和29.05个百分点。构建基于社交网络的情感图谱有助于相关部门发现意见领袖及其情感倾向,从而把握网络舆情的发展趋势,并降低消极情感对社会造成的影响。
中图分类号:
仇丽青, 曲福帅. 基于情感分析和影响力评估的突发事件情感图谱[J]. 计算机应用, 2022, 42(5): 1330-1338.
Liqing QIU, Fushuai QU. Emotional map of emergency based on sentiment analysis and influence evaluation[J]. Journal of Computer Applications, 2022, 42(5): 1330-1338.
| 参数 | 值 | 
|---|---|
| 词向量维度 | 100 | 
| 卷积核大小 | 3 | 
| 丢弃率 | 0.6 | 
| Bi-LSTM隐藏单元个数 | 128 | 
| 优化器 | Adam Optimizer | 
| 早停法耐心值 | 50 | 
| 学习率 | 0.001 | 
| 批次大小 | 16 | 
表1 实验参数设置
Tab. 1 Experimental parameter setting
| 参数 | 值 | 
|---|---|
| 词向量维度 | 100 | 
| 卷积核大小 | 3 | 
| 丢弃率 | 0.6 | 
| Bi-LSTM隐藏单元个数 | 128 | 
| 优化器 | Adam Optimizer | 
| 早停法耐心值 | 50 | 
| 学习率 | 0.001 | 
| 批次大小 | 16 | 
| 模型 | 情感倾向 | 准确率 | 召回率 | F1值 | 
|---|---|---|---|---|
| Text-RNN | 积极 | 76.08 | 52.70 | 62.27 | 
| 消极 | 79.20 | 91.57 | 84.98 | |
| Text-RCNN | 积极 | 78.75 | 65.55 | 71.54 | 
| 消极 | 83.86 | 91.00 | 87.28 | |
| Text-RNN+Attention | 积极 | 74.32 | 65.74 | 69.76 | 
| 消极 | 83.54 | 88.45 | 85.92 | |
| FastText | 积极 | 80.68 | 52.89 | 63.89 | 
| 消极 | 79.61 | 93.56 | 86.03 | |
| DPCNN | 积极 | 76.50 | 64.25 | 69.84 | 
| 消极 | 83.19 | 89.96 | 86.44 | |
| Transformer | 积极 | 74.69 | 67.60 | 70.97 | 
| 消极 | 84.28 | 88.35 | 86.27 | |
| 本文模型 | 积极 | 78.98 | 66.48 | 72.19 | 
| 消极 | 84.22 | 91.00 | 87.48 | 
表2 不同模型在酒店评论数据集上的实验结果 (%)
Tab. 2 Experimental results of different models on hotel review dataset
| 模型 | 情感倾向 | 准确率 | 召回率 | F1值 | 
|---|---|---|---|---|
| Text-RNN | 积极 | 76.08 | 52.70 | 62.27 | 
| 消极 | 79.20 | 91.57 | 84.98 | |
| Text-RCNN | 积极 | 78.75 | 65.55 | 71.54 | 
| 消极 | 83.86 | 91.00 | 87.28 | |
| Text-RNN+Attention | 积极 | 74.32 | 65.74 | 69.76 | 
| 消极 | 83.54 | 88.45 | 85.92 | |
| FastText | 积极 | 80.68 | 52.89 | 63.89 | 
| 消极 | 79.61 | 93.56 | 86.03 | |
| DPCNN | 积极 | 76.50 | 64.25 | 69.84 | 
| 消极 | 83.19 | 89.96 | 86.44 | |
| Transformer | 积极 | 74.69 | 67.60 | 70.97 | 
| 消极 | 84.28 | 88.35 | 86.27 | |
| 本文模型 | 积极 | 78.98 | 66.48 | 72.19 | 
| 消极 | 84.22 | 91.00 | 87.48 | 
| 模型 | 情感倾向 | 准确率 | 召回率 | F1值 | 
|---|---|---|---|---|
| Text-RNN | 积极 | 78.75 | 77.63 | 78.19 | 
| 消极 | 78.38 | 79.48 | 78.93 | |
| Text-RCNN | 积极 | 78.03 | 79.61 | 79.86 | 
| 消极 | 81.26 | 77.45 | 79.31 | |
| Text-RNN+Attention | 积极 | 77.10 | 80.33 | 78.68 | 
| 消极 | 79.90 | 76.63 | 78.23 | |
| FastText | 积极 | 48.61 | 18.09 | 26.37 | 
| 消极 | 50.31 | 81.27 | 62.15 | |
| DPCNN | 积极 | 86.03 | 69.83 | 77.09 | 
| 消极 | 75.04 | 88.89 | 81.38 | |
| Transformer | 积极 | 78.24 | 78.46 | 78.35 | 
| 消极 | 78.84 | 78.61 | 78.73 | |
| 本文模型 | 积极 | 79.31 | 83.97 | 81.57 | 
| 消极 | 81.46 | 77.60 | 79.49 | 
表3 不同模型在电商评论数据集上的实验结果 (%)
Tab. 3 Experimental results of different models on e-commerce review dataset
| 模型 | 情感倾向 | 准确率 | 召回率 | F1值 | 
|---|---|---|---|---|
| Text-RNN | 积极 | 78.75 | 77.63 | 78.19 | 
| 消极 | 78.38 | 79.48 | 78.93 | |
| Text-RCNN | 积极 | 78.03 | 79.61 | 79.86 | 
| 消极 | 81.26 | 77.45 | 79.31 | |
| Text-RNN+Attention | 积极 | 77.10 | 80.33 | 78.68 | 
| 消极 | 79.90 | 76.63 | 78.23 | |
| FastText | 积极 | 48.61 | 18.09 | 26.37 | 
| 消极 | 50.31 | 81.27 | 62.15 | |
| DPCNN | 积极 | 86.03 | 69.83 | 77.09 | 
| 消极 | 75.04 | 88.89 | 81.38 | |
| Transformer | 积极 | 78.24 | 78.46 | 78.35 | 
| 消极 | 78.84 | 78.61 | 78.73 | |
| 本文模型 | 积极 | 79.31 | 83.97 | 81.57 | 
| 消极 | 81.46 | 77.60 | 79.49 | 
| 网络 | 节点数 | 边数 | 最大度 | 平均度 | 同配性 | 
|---|---|---|---|---|---|
| Karate | 34 | 78 | 17 | 4.558 2 | |
| Copperfield | 112 | 425 | 49 | 7.589 3 | |
| USAir97 | 332 | 2 126 | 139 | 12.000 0 | |
| Euroroad | 1 174 | 1 417 | 10 | 2.414 0 | 0.126 7 | 
| PowerGrid | 4 941 | 6 594 | 19 | 2.669 1 | 0.003 5 | 
表4 社交网络数据集统计
Tab. 4 Statistics of social network datasets
| 网络 | 节点数 | 边数 | 最大度 | 平均度 | 同配性 | 
|---|---|---|---|---|---|
| Karate | 34 | 78 | 17 | 4.558 2 | |
| Copperfield | 112 | 425 | 49 | 7.589 3 | |
| USAir97 | 332 | 2 126 | 139 | 12.000 0 | |
| Euroroad | 1 174 | 1 417 | 10 | 2.414 0 | 0.126 7 | 
| PowerGrid | 4 941 | 6 594 | 19 | 2.669 1 | 0.003 5 | 
| 网络 | 算法 | |||||
|---|---|---|---|---|---|---|
| BC | CC | DC | HX | KS | WDK | |
| Karate | 0.775 4 | 0.899 3 | 0.707 9 | 0.576 6 | 0.495 8 | 0.964 7 | 
| Copperfiled | 0.978 9 | 0.983 7 | 0.866 1 | 0.811 0 | 0.599 0 | 0.991 2 | 
| USAir97 | 0.697 0 | 0.989 2 | 0.858 6 | 0.835 5 | 0.811 4 | 0.995 9 | 
| Euroroad | 0.937 4 | 0.998 8 | 0.444 2 | 0.253 4 | 0.212 9 | 0.999 4 | 
| PowerGrid | 0.831 9 | 0.999 8 | 0.592 7 | 0.393 0 | 0.246 0 | 0.999 9 | 
表5 节点影响力算法的M函数值
Tab. 5 M-function value of node influence algorithm
| 网络 | 算法 | |||||
|---|---|---|---|---|---|---|
| BC | CC | DC | HX | KS | WDK | |
| Karate | 0.775 4 | 0.899 3 | 0.707 9 | 0.576 6 | 0.495 8 | 0.964 7 | 
| Copperfiled | 0.978 9 | 0.983 7 | 0.866 1 | 0.811 0 | 0.599 0 | 0.991 2 | 
| USAir97 | 0.697 0 | 0.989 2 | 0.858 6 | 0.835 5 | 0.811 4 | 0.995 9 | 
| Euroroad | 0.937 4 | 0.998 8 | 0.444 2 | 0.253 4 | 0.212 9 | 0.999 4 | 
| PowerGrid | 0.831 9 | 0.999 8 | 0.592 7 | 0.393 0 | 0.246 0 | 0.999 9 | 
| 网络 | τ(σ,WDK) | |||||
|---|---|---|---|---|---|---|
| Karate | 0.556 1 | 0.702 3 | 0.830 7 | 0.616 8 | 0.572 2 | 0.862 7 | 
| Copperfield | 0.638 0 | 0.851 2 | 0.903 0 | 0.831 4 | 0.725 7 | 0.916 5 | 
| USAir97 | 0.517 8 | 0.804 8 | 0.916 8 | 0.763 0 | 0.759 3 | 0.943 1 | 
| Euroroad | 0.393 4 | 0.608 0 | 0.489 5 | 0.429 1 | 0.408 8 | 0.724 1 | 
| PowerGrid | 0.377 4 | 0.349 2 | 0.488 5 | 0.444 2 | 0.346 8 | 0.682 9 | 
表6 节点影响力算法的肯德尔相关系数
Tab. 6 Kendall coefficients of node influence algorithms
| 网络 | τ(σ,WDK) | |||||
|---|---|---|---|---|---|---|
| Karate | 0.556 1 | 0.702 3 | 0.830 7 | 0.616 8 | 0.572 2 | 0.862 7 | 
| Copperfield | 0.638 0 | 0.851 2 | 0.903 0 | 0.831 4 | 0.725 7 | 0.916 5 | 
| USAir97 | 0.517 8 | 0.804 8 | 0.916 8 | 0.763 0 | 0.759 3 | 0.943 1 | 
| Euroroad | 0.393 4 | 0.608 0 | 0.489 5 | 0.429 1 | 0.408 8 | 0.724 1 | 
| PowerGrid | 0.377 4 | 0.349 2 | 0.488 5 | 0.444 2 | 0.346 8 | 0.682 9 | 
| 1 | 中国互联网络信息中心.第47 次中国互联网络发展状况统计报告[EB/OL].[2021-03-12].. 10.1007/978-981-33-6930-6_2 | 
| China Internet Network Information Center. The 47th China statistical report on Internet development [EB/OL]. [2021-03-12]. . 10.1007/978-981-33-6930-6_2 | |
| 2 | 戴杏云,张柳,戴伟辉,等.社交网络的情感图谱研究[J].管理评论,2016,28(8):79-86. 10.1111/jce.13217 | 
| DAI X Y, ZHANG L, DAI W H, et al. Research on emotional mapping of social networks [J]. Management Review, 2016, 28(8): 79-86. 10.1111/jce.13217 | |
| 3 | 黄星,刘樑.突发事件网络舆情风险评价方法及应用[J].情报科学,2018,36(4):3-9. | 
| HUANG X, LIU L. The evaluation method and application of unexpected events network public opinion [J]. Information Science, 2018, 36(4): 3-9. | |
| 4 | 安璐,欧孟花.突发公共卫生事件利益相关者的社会网络情感图谱研究[J].图书情报工作,2017,61(20):120-130. 10.13266/j.issn.0252-3116.2017.20.013 | 
| AN L, OU M H. Social network sentiment map of the stakeholders in public health emergencies [J]. Library and Information Service, 2017, 61(20): 120-130. 10.13266/j.issn.0252-3116.2017.20.013 | |
| 5 | 赵蓉英,王旭.突发事件网络舆情关键节点识别及导控对策研究——以“大贤村遭洪灾事件”为例[J].现代情报,2018,38(1):19-24,30. 10.3969/j.issn.1008-0821.2018.01.002 | 
| ZHAO R Y, WANG X. Research on identifying key nodes and guiding and controlling strategies of network public opinion in emergency — a case study of being suffered by flooding in Da Xian Village [J]. Journal of Modern Information, 2018, 38(1): 19-24, 30. 10.3969/j.issn.1008-0821.2018.01.002 | |
| 6 | 张柳,王晰巍,王铎,等.微博环境下高校舆情情感演化图谱研究——以新浪微博“高校学术不端”话题为例[J].现代情报,2019,39(10):119-126,135. 10.3969/j.issn.1008-0821.2019.10.014 | 
| ZHANG L, WANG X W, WANG D, et al. The study of emotional evolution map of public opinions in university under the microblog environment — a case of “academic misconduct in universities” in Sina Weibo [J]. Journal of Modern Information, 2019, 39(10): 119-126, 135. 10.3969/j.issn.1008-0821.2019.10.014 | |
| 7 | HEMMATIAN F, SOHRABI M K. A survey on classification techniques for opinion mining and sentiment analysis [J]. Artificial Intelligence Review, 2019, 52(3): 1495-1545. 10.1007/s10462-017-9599-6 | 
| 8 | VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need [C]// Proceedings of the 2017 31st International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2017: 6000-6010. 10.1016/s0262-4079(17)32358-8 | 
| 9 | WANG Y Q, HUANG M L, ZHAO L, et al. Attention-based LSTM for aspect-level sentiment classification [C]// Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2016: 606-615. 10.18653/v1/d16-1058 | 
| 10 | LIU Z, ZHOU W, LI H. AB-LSTM: attention-based bidirectional LSTM model for scene text detection [J]. ACM Transactions on Multimedia Computing, Communications, and Applications, 2019, 15(4): Article No.107. 10.1145/3356728 | 
| 11 | 关鹏飞,李宝安,吕学强,等.注意力增强的双向LSTM情感分析[J].中文信息学报,2019,33(2):105-111. 10.3969/j.issn.1003-0077.2019.02.014 | 
| GUAN P F, LI B A, LYU X Q, et al. Attention enhanced Bi-directional LSTM for sentiment analysis [J]. Journal of Chinese Information Processing, 2019, 33(2): 105-111. 10.3969/j.issn.1003-0077.2019.02.014 | |
| 12 | ZAREIE A, SHEIKHAHMADI A, FATEMI A. Influential nodes ranking in complex networks: an entropy-based approach [J]. Chaos, Solitons and Fractals, 2017, 104: 485-494. 10.1016/j.chaos.2017.09.010 | 
| 13 | FREEMAN L C. Centrality in social networks conceptual clarification [J]. Social Networks, 1978, 1(3): 215-239. 10.1016/0378-8733(78)90021-7 | 
| 14 | SABIDUSSI G. The centrality index of a graph [J]. Psychometrika, 1966, 31(4): 581-603. 10.1007/bf02289527 | 
| 15 | NEWMAN M E J. A measure of betweenness centrality based on random walks [J]. Social Networks, 2005, 27(1): 39-54. 10.1016/j.socnet.2004.11.009 | 
| 16 | BRIN S, PAGE L. The anatomy of a large-scale hypertextual Web search engine [J]. Computer Networks and ISDN Systems, 1998, 30 (1/2/3/4/5/6/7) :107-117. | 
| 17 | GARAS A, SCHWEITZER F, HAVLIN S. A k-shell de-composition method for weighted networks [J]. New Journal of Physics, 2012, 14(8): 2017 No. 083030. 10.1088/1367-2630/14/8/083030 | 
| 18 | 邓凯旋,陈鸿昶,黄瑞阳.一种基于改进K-shell的节点重要性排序方法[J].计算机应用研究,2017,34(10):3017-3019, 3084. 10.3969/j.issn.1001-3695.2017.10.031 | 
| DENG K X, CHEN H C, HUANG R Y. Method of node importance ranking based on improved K-shell [J]. Application Research of Computers, 2017, 34(10):3017-3019, 3084. 10.3969/j.issn.1001-3695.2017.10.031 | |
| 19 | IBNOULOUAFI A, HAZITI M EL. Density centrality: identifying influential nodes based on area density formula [J]. Chaos, Solitons and Fractals, 2018, 114: 69-80. 10.1016/j.chaos.2018.06.022 | 
| 20 | 王安,顾益军.基于社区划分的节点重要性评估方法[J].计算机工程与应用,2020,56(8):42-48. | 
| WANG A, GU Y J. Nodes importance ranking method based on community detection [J]. Computer Engineering and Applications, 2020, 56(8): 42-48. | |
| 21 | HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. 10.1162/neco.1997.9.8.1735 | 
| 22 | LAI S W, XU L H, LIU K, et al. Recurrent convolutional neural networks for text classification [C]// Proceedings of the 2015 29th AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2015: 2267-2273. 10.1609/aaai.v33i01.33017370 | 
| 23 | LIU P, QIU X, HUANG X. Recurrent neural network for text classification with multi-task learning [C]// Proceedings of the 2016 Twenty-Fifth International Joint Conference on Artificial Intelligence. New York: AAAI Press, 2016: 2873-2879. | 
| 24 | YANG Z C, YANG D Y, DYER C, et al. Hierarchical attention networks for document classification [C]// Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Stroudsburg: ACL, 2016: 1480-1489. 10.18653/v1/n16-1174 | 
| 25 | JOULIN A, GRAVE E, BOJANOWSKI P, et al. Bag of tricks for efficient text classification [C]// Proceedings of the 2017 15th Conference of the European Chapter of the Association for Computational Linguistics (Volume 2, Short Papers). Stroudsburg: ACL, 2017: 427-431. 10.18653/v1/e17-2068 | 
| 26 | JOHNSON R, ZHANG T. Deep pyramid convolutional neural networks for text categorization [C]// Proceedings of the 2017 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Stroudsburg: ACL, 2017:562-570. 10.18653/v1/p17-1052 | 
| 27 | LÜ L Y, ZHOU T, ZHANG Q M, et al. The H-index of a network node and its relation to degree and coreness [J]. Nature Communications, 2016, 7: Article No.10168. 10.1038/ncomms10168 | 
| 28 | LI H J, ZHANG X S. Analysis of stability of community structure across multiple hierarchical levels[J]. Europhysics Letters, 2013, 103(5): Article No.58002. 10.1209/0295-5075/103/58002 | 
| [1] | 薛凯鹏, 徐涛, 廖春节. 融合自监督和多层交叉注意力的多模态情感分析网络[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2387-2392. | 
| [2] | 林欣蕊, 王晓菲, 朱焱. 基于局部扩展社区发现的学术异常引用群体检测[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1855-1861. | 
| [3] | 柯添赐, 刘建华, 孙水华, 郑智雄, 蔡子杰. 融合强关联依赖和简洁语法的方面级情感分析模型[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1786-1795. | 
| [4] | 高龙涛, 李娜娜. 基于方面感知注意力增强的方面情感三元组抽取[J]. 《计算机应用》唯一官方网站, 2024, 44(4): 1049-1057. | 
| [5] | 杨先凤, 汤依磊, 李自强. 基于交替注意力机制和图卷积网络的方面级情感分析模型[J]. 《计算机应用》唯一官方网站, 2024, 44(4): 1058-1064. | 
| [6] | 郭磊, 贾真, 李天瑞. 面向方面级情感分析的交互式关系图注意力网络[J]. 《计算机应用》唯一官方网站, 2024, 44(3): 696-701. | 
| [7] | 高瑞, 陈学斌, 张祖篡. 面向部分图更新的动态社交网络隐私发布方法[J]. 《计算机应用》唯一官方网站, 2024, 44(12): 3831-3838. | 
| [8] | 刘世梁, 王义, 马应龙. 考虑社区规模不平衡的非重叠社区检测[J]. 《计算机应用》唯一官方网站, 2024, 44(11): 3396-3402. | 
| [9] | 李言博, 何庆, 陆顺意. 融合语义和句法信息的方面情感三元组抽取[J]. 《计算机应用》唯一官方网站, 2024, 44(10): 3275-3280. | 
| [10] | 罗俊豪, 朱焱. 用于未对齐多模态语言序列情感分析的多交互感知网络[J]. 《计算机应用》唯一官方网站, 2024, 44(1): 79-85. | 
| [11] | 李莉, 杨春艳, 朱江文, 胡荣磊. 区块链下社交网络用户抄袭识别方案[J]. 《计算机应用》唯一官方网站, 2024, 44(1): 242-251. | 
| [12] | 陈丽安, 过弋. 融合个体偏差信息的文本情感分析模型[J]. 《计算机应用》唯一官方网站, 2024, 44(1): 145-151. | 
| [13] | 张心月, 刘蓉, 魏驰宇, 方可. 融合提示知识的方面级情感分析方法[J]. 《计算机应用》唯一官方网站, 2023, 43(9): 2753-2759. | 
| [14] | 彭诗杰, 陈红梅, 王丽珍, 肖清. 基于地理偏好排序的兴趣点混合推荐模型[J]. 《计算机应用》唯一官方网站, 2023, 43(8): 2448-2455. | 
| [15] | 衡红军, 杨鼎诚. 知识增强的方面词交互图神经网络[J]. 《计算机应用》唯一官方网站, 2023, 43(8): 2412-2419. | 
| 阅读次数 | ||||||
| 全文 |  | |||||
| 摘要 |  | |||||
