基于OTSRM模型的话题情感演化分析
|
摘要
舆情话题检测与情感演化分析在舆情监控中起着非常重要的作用,但当前方法存在着情感话题含义不明确、情感态势评估不精确等问题。在OLDA (Online Latent Dirichlet Allocation)模型的基础上引入情感强度,并提出一种情感迭代思想,构建在线话题情感识别模型OTSRM(Online Topic and Sentiment Recognition Mode)。该模型通过增加基于β先验的情感遗传度,建立情感演化通道,获取特征词、情感词2个分布矩阵,最后使用相对熵方法计算话题焦点在相邻时间片段上的最大情感值,从而高效地识别不同文本的话题情感。在5个网络事件数据集上对OTSRM模型进行有效性验证,并与主流模型进行了对比,实验表明OTSRM模型在舆情话题识别与话题情感演化分析方面实现了良好效果。
The sentiment evolution of online public topics plays a very important part in the analysis of public opinion,while current methods have problems such as unclear meanings of sentiment topics and inaccurate evaluation of sentiment evolution. This paper introduced sentiment intensity based on the OLDA model and proposed an Online Topic and Senti‐ment Recognition Mode(OTSRM). By adding sentiment heritability with a β prior parameter, this model established a sen‐timent evolution channel and obtained two distribution matrices of feature words and sentiment words. Finally, the relative entropy method was proposed to calculate the maximum value of topic sentiment in adjacent time segments, thereby effi‐ciently identifying the topic sentiment of different texts. The effectiveness of OTSRM was validated using five network da‐tasets and compared with other state-of-the-art models. The experiments showed that our approach achieved good results in the recognition of topic sentiment.
The sentiment evolution of online public topics plays a very important part in the analysis of public opinion,while current methods have problems such as unclear meanings of sentiment topics and inaccurate evaluation of sentiment evolution. This paper introduced sentiment intensity based on the OLDA model and proposed an Online Topic and Senti‐ment Recognition Mode(OTSRM). By adding sentiment heritability with a β prior parameter, this model established a sen‐timent evolution channel and obtained two distribution matrices of feature words and sentiment words. Finally, the relative entropy method was proposed to calculate the maximum value of topic sentiment in adjacent time segments, thereby effi‐ciently identifying the topic sentiment of different texts. The effectiveness of OTSRM was validated using five network da‐tasets and compared with other state-of-the-art models. The experiments showed that our approach achieved good results in the recognition of topic sentiment.
引文
[1]黄晓斌,赵超.文本挖掘在网络舆情信息分析中的应用[J].情报科学, 2009, 27(1):94-99.
[2]聂峰英,张旸.移动社交网络舆情预警指标体系构建[J].情报理论与实践, 2015, 38(12):64-67.
[3] Li G, Jiang S, Zhang W, et al. Online Web video topic detection and tracking with semi-supervised learning[J]. Multimedia Systems, 2016, 22(1):115-125.
[4] Hofmann T. Probabilistic latent semantic indexing[C]//Proceedings of the 22nd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. NewYork:ACM Press, 1999:50-57.
[5] Blei D M, Ng A Y, Jordan M I. Latent Dirichlet allocation[J].Journal of Machine Learning Research, 2003, 3(9):993-1022.
[6] Blei D M, Lafferty J D. Dynamic topic models[C]//Proceedings of the 23rd International Conference on Machine Learning. NewYork:ACM Press, 2006:113-120.
[7] Alsumait L, Domeniconi C. On-line LDA:Adaptive topic models for mining text streams with applications to topic detection and tracking[C]//Proceedings of the Eighth IEEE International Conference on Data Mining. IEEE Computer Society, 2008:3-12.
[8] Lin C, He Y, Everson R, et al. Weakly supervised joint sentimenttopic detection from text[J]. IEEE Transactions on Knowledge and Data Engineering, 2012, 24(6):1134-1145.
[9] Pavitra R, Kalaivaani P C D. Weakly supervised sentiment analysis using joint sentiment topic detection with bigrams[C]//Proceedings of the International Conference on Electronics and Communication Systems. IEEE, 2015:889-893.
[10]周文,张书卿,欧阳纯萍,等.基于情感依存元组的新闻文本主题情感分析[J].山东大学学报(理学版), 2014, 49(12):1-6, 11.
[11] Liu Y, Guo Q, Wu X, et al. Evolution identification approach for news public opinion based on TSSCM[J]. Journal of Intelligence,2017, 36(2):115-121.
[12] Alam M H, Ryu W J, Lee S K. Joint multi-grain topic sentiment:Modeling semantic aspects for online reviews[J]. Information Sciences, 2016, 339:206-223.
[13] Lim K W, Buntine W. Twitter opinion topic model:Extracting product opinions from Tweets by leveraging hashtags and sentiment lexicon[C]//Proceedings of the 23rd ACM International Conference on Information and Knowledge Management. New York:ACM Press, 2014:1319-1328.
[14]黄卫东,陈凌云,吴美蓉.网络舆情话题情感演化研究[J].情报杂志, 2014(1):102-107.
[15] Rao Y. Contextual sentiment topic model for adaptive social emotion classification[J]. IEEE Intelligent Systems, 2016, 31(1):41-47.
[16] Steuer R, Kurths J, Daub C O, et al. The mutual information:Detecting and evaluating dependencies between variables[J]. Bioinformatics, 2002, 18(Suppl 2):S231-S240.
[17] Haselmayer M, Jenny M. Sentiment analysis of political communication:Combining a dictionary approach with crowdcoding[J].Quality&Quantity, 2017, 51(6):2623-2646.
[18] Moreira C, Wichert A. Finding academic experts on a multisensor approach using Shannon’s entropy[J]. Expert Systems with Applications, 2013, 40(14):5740-5754.
[19] GooSeeker. MetaSeeker[EB/OL].[2016-08-16]. http://www.gooseeker.com/product.
[2]聂峰英,张旸.移动社交网络舆情预警指标体系构建[J].情报理论与实践, 2015, 38(12):64-67.
[3] Li G, Jiang S, Zhang W, et al. Online Web video topic detection and tracking with semi-supervised learning[J]. Multimedia Systems, 2016, 22(1):115-125.
[4] Hofmann T. Probabilistic latent semantic indexing[C]//Proceedings of the 22nd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. NewYork:ACM Press, 1999:50-57.
[5] Blei D M, Ng A Y, Jordan M I. Latent Dirichlet allocation[J].Journal of Machine Learning Research, 2003, 3(9):993-1022.
[6] Blei D M, Lafferty J D. Dynamic topic models[C]//Proceedings of the 23rd International Conference on Machine Learning. NewYork:ACM Press, 2006:113-120.
[7] Alsumait L, Domeniconi C. On-line LDA:Adaptive topic models for mining text streams with applications to topic detection and tracking[C]//Proceedings of the Eighth IEEE International Conference on Data Mining. IEEE Computer Society, 2008:3-12.
[8] Lin C, He Y, Everson R, et al. Weakly supervised joint sentimenttopic detection from text[J]. IEEE Transactions on Knowledge and Data Engineering, 2012, 24(6):1134-1145.
[9] Pavitra R, Kalaivaani P C D. Weakly supervised sentiment analysis using joint sentiment topic detection with bigrams[C]//Proceedings of the International Conference on Electronics and Communication Systems. IEEE, 2015:889-893.
[10]周文,张书卿,欧阳纯萍,等.基于情感依存元组的新闻文本主题情感分析[J].山东大学学报(理学版), 2014, 49(12):1-6, 11.
[11] Liu Y, Guo Q, Wu X, et al. Evolution identification approach for news public opinion based on TSSCM[J]. Journal of Intelligence,2017, 36(2):115-121.
[12] Alam M H, Ryu W J, Lee S K. Joint multi-grain topic sentiment:Modeling semantic aspects for online reviews[J]. Information Sciences, 2016, 339:206-223.
[13] Lim K W, Buntine W. Twitter opinion topic model:Extracting product opinions from Tweets by leveraging hashtags and sentiment lexicon[C]//Proceedings of the 23rd ACM International Conference on Information and Knowledge Management. New York:ACM Press, 2014:1319-1328.
[14]黄卫东,陈凌云,吴美蓉.网络舆情话题情感演化研究[J].情报杂志, 2014(1):102-107.
[15] Rao Y. Contextual sentiment topic model for adaptive social emotion classification[J]. IEEE Intelligent Systems, 2016, 31(1):41-47.
[16] Steuer R, Kurths J, Daub C O, et al. The mutual information:Detecting and evaluating dependencies between variables[J]. Bioinformatics, 2002, 18(Suppl 2):S231-S240.
[17] Haselmayer M, Jenny M. Sentiment analysis of political communication:Combining a dictionary approach with crowdcoding[J].Quality&Quantity, 2017, 51(6):2623-2646.
[18] Moreira C, Wichert A. Finding academic experts on a multisensor approach using Shannon’s entropy[J]. Expert Systems with Applications, 2013, 40(14):5740-5754.
[19] GooSeeker. MetaSeeker[EB/OL].[2016-08-16]. http://www.gooseeker.com/product.