基于新浪微博的营销行为及用户偏好研究
发布时间：2019-04-03 12:49
【摘要】：随着微博这一新媒体的发展,微博营销迅速兴起,成为各行各业关注的焦点。在微博这一新兴社交媒体平台上,营销行为不仅引起了广大媒体人的关注,同时也得到了学术界的青睐,微博营销行为所产生的海量数据,正是发挥数据挖掘技术优势的地方。因此,本文通过研究数据挖掘方法,借助相关算法对微博营销行为特征进行挖掘分析。 本文主要研究了新浪微博中的电影微博营销行为及相关用户偏好特征。文章首先提出了一种基于新浪微博文本内容的话题提取算法——基于模糊C均值(FCM)聚类算法的话题提取算法(TEBF)。该算法在FCM算法中隶属度(membership)这一概念的基础上提出了M-membership这一指标,实验证明该指标可以很好地衡量模糊集合中词语的重要性,最终实现以模糊集合的形式描述话题,并通过对比实验说明了TEBF算法话题提取结果的有效性。然后,文章结合微博营销行为和用户偏好分析的具体需求,基于话题提取算法和分类算法,提出了营销行为分析模型和用户偏好分析模型,从微博内容和用户信息两个方面对微博营销行为及用户偏好进行了较为全面系统的分析。 为了对分析结果进行简单直观的展示,本文设计开发了“WMA”微博营销行为分析系统,以期为更多的用户提供一个方便有效的微博营销行为及用户偏好分析平台,同时也证明本文的研究具有一定的实用价值。
[Abstract]:With the development of Weibo, Weibo marketing has become the focus of various industries. On the emerging social media platform, Weibo, the marketing behavior not only attracts the attention of the media, but also gets the favor of the academic circles. The massive data produced by the Weibo marketing behavior is the place where the advantages of data mining technology can be brought into play. Therefore, this paper through the study of data mining methods, with the help of related algorithms to analyze the characteristics of Weibo marketing behavior. This paper mainly studies the marketing behavior of movie Weibo and the characteristics of user preference in Sina Weibo. In this paper, we first propose a topic extraction algorithm based on Sina Weibo text content-topic extraction algorithm (TEBF). Based on fuzzy C-means clustering algorithm. Based on the concept of membership degree (membership) in FCM algorithm, this algorithm puts forward M-membership index. The experiment proves that this index can well measure the importance of words in fuzzy set, and finally describe the topic in the form of fuzzy set. The validity of the topic extraction results of the TEBF algorithm is proved by a comparative experiment. Then, combined with the specific requirements of Weibo marketing behavior and user preference analysis, based on topic extraction algorithm and classification algorithm, this paper proposes a marketing behavior analysis model and a user preference analysis model. This paper makes a comprehensive and systematic analysis of Weibo marketing behavior and user preference from the aspects of Weibo content and user information. In order to display the analysis results simply and intuitively, this paper designs and develops the "WMA" Weibo marketing behavior analysis system, in order to provide more users with a convenient and effective Weibo marketing behavior and user preference analysis platform. At the same time, it is proved that the research in this paper has certain practical value.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.1
本文编号：2453227
[Abstract]:With the development of Weibo, Weibo marketing has become the focus of various industries. On the emerging social media platform, Weibo, the marketing behavior not only attracts the attention of the media, but also gets the favor of the academic circles. The massive data produced by the Weibo marketing behavior is the place where the advantages of data mining technology can be brought into play. Therefore, this paper through the study of data mining methods, with the help of related algorithms to analyze the characteristics of Weibo marketing behavior. This paper mainly studies the marketing behavior of movie Weibo and the characteristics of user preference in Sina Weibo. In this paper, we first propose a topic extraction algorithm based on Sina Weibo text content-topic extraction algorithm (TEBF). Based on fuzzy C-means clustering algorithm. Based on the concept of membership degree (membership) in FCM algorithm, this algorithm puts forward M-membership index. The experiment proves that this index can well measure the importance of words in fuzzy set, and finally describe the topic in the form of fuzzy set. The validity of the topic extraction results of the TEBF algorithm is proved by a comparative experiment. Then, combined with the specific requirements of Weibo marketing behavior and user preference analysis, based on topic extraction algorithm and classification algorithm, this paper proposes a marketing behavior analysis model and a user preference analysis model. This paper makes a comprehensive and systematic analysis of Weibo marketing behavior and user preference from the aspects of Weibo content and user information. In order to display the analysis results simply and intuitively, this paper designs and develops the "WMA" Weibo marketing behavior analysis system, in order to provide more users with a convenient and effective Weibo marketing behavior and user preference analysis platform. At the same time, it is proved that the research in this paper has certain practical value.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.1
【参考文献】
相关期刊论文 前3条
1 葛进平;邹立清;;电影微博立体营销策略探讨[J];当代电影;2012年02期
2 徐戈;王厚峰;;自然语言处理中主题模型的发展[J];计算机学报;2011年08期
3 张保富;施化吉;马素琴;;基于TFIDF文本特征加权方法的改进研究[J];计算机应用与软件;2011年02期
，本文编号：2453227
本文链接：https://www.wllwen.com/guanlilunwen/yingxiaoguanlilunwen/2453227.html
