社交网络热点内容的时间序列研究

发布时间：2018-04-16 11:05

本文选题：社交网络 + 时间序列分析　；参考：《北京邮电大学》2017年硕士论文

【摘要】：随着信息社会的高速发展,信息传播越来越趋向于多维化,互联网作为飞速发展的新媒体,在信息传播方面功不可没。社交网络的出现改变了人们获取信息的方式,使得消息交换和共享变得快捷方便。社交网络不断涌现的热门词语,不仅影响着人们现实生活的观点和立场,而且很大程度上反映了社会舆论。因此,分析热点内容的时间序列传播趋势,有效抑制负面消息传播,促进正面积极消息传播,揭示其传播动力学特征,从而能进一步分析舆论导向,具有非常重要的现实意义。基于社交网络热点内容的多元性和随机性,本文从时间序列的角度来分析热点内容的传播规律,同时引入模糊数学理论,提出模糊集分区算法进行时间序列集合划分,提出模糊趋势分析算法来求解传播趋势,反映了热点内容的传播过程和突变规律。针对社交网络热点内容传播的场景,本文提出了一种基于模糊数学的热点内容时间序列预测模型。针对该模型,提出了模糊集分区算法,通过模糊隶属函数对热点内容时间序列进行模糊区间划分,采用软计算方法对热词传播数值进行灵活分区,同时将异常值加以考虑。然后提出了模糊趋势分析算法,分析单位时间数据对应的模糊集合和传播趋势,得出合理的预测结果。与传统时间序列预测模型相比,本文提出的模糊时间序列分析预测模型不仅拓宽了时间序列分析在社交网络的应用范围,同时更加准确地分析社交网络热点内容的传播规律。在实验仿真验证方面,本文提出了基于模糊数学的时间序列分析预测模型,通过拟合程度和预测准确度证明了算法的有效性。本文选取新浪微博热点词语作为实验数据,采用海量热词数据对模糊时间序列预测模型进行仿真。与传统时间序列预测模型相比,模糊时间序列预测模型能够更好地拟合社交网络热词的传播过程,具有更高的预测准确度。
[Abstract]:With the rapid development of the information society, the information communication tends to be multidimensional. The Internet, as a new media with rapid development, has contributed to the information dissemination.The emergence of social networks has changed the way people access information, making message exchange and sharing become fast and convenient.The popular words of social network not only affect people's viewpoint and stand in real life, but also reflect public opinion to a great extent.Therefore, it is of great practical significance to analyze the trend of time series communication of hot topics, effectively restrain the spread of negative news, promote the spread of positive news, reveal the dynamic characteristics of communication, and further analyze the orientation of public opinion.Based on the diversity and randomness of hot spots in social networks, this paper analyzes the propagation rules of hot spots from the point of view of time series, and introduces the theory of fuzzy mathematics, and proposes a fuzzy set partition algorithm for time series partitioning.A fuzzy trend analysis algorithm is proposed to solve the propagation trend, which reflects the propagation process and mutation law of hot topics.Aiming at the scene of hot content propagation in social networks, this paper presents a prediction model of hot content time series based on fuzzy mathematics.According to this model, a fuzzy set partition algorithm is proposed. The fuzzy interval partition of the time series of hot spots is carried out by fuzzy membership function, and the numerical value of hot word propagation is flexibly partitioned by soft computing method, and the outliers are considered at the same time.Then a fuzzy trend analysis algorithm is proposed to analyze the fuzzy set and propagation trend corresponding to the unit time data, and the reasonable prediction results are obtained.Compared with the traditional time series prediction model, the proposed fuzzy time series analysis and prediction model not only widens the scope of application of time series analysis in social networks, but also more accurately analyzes the propagation rules of hot spots in social networks.In the aspect of experimental simulation, a time series analysis and prediction model based on fuzzy mathematics is proposed. The validity of the algorithm is proved by the degree of fitting and the accuracy of prediction.In this paper, the hot words of Sina Weibo are selected as experimental data, and the fuzzy time series prediction model is simulated by massive hot word data.Compared with the traditional time series prediction model, the fuzzy time series prediction model can better fit the propagation process of social network hot words, and has higher prediction accuracy.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O159;O211.61

【相似文献】