社会化媒体中提升用户参与度的关键因素研究
发布时间：2018-10-14 13:33
【摘要】：短短几年间，社会化媒体得到了迅猛的发展，用户数量和覆盖率不断刷新记录，在社会生活中的地位和作用日渐重要。社会化媒体的核心是“社会化”，即用户的参与和互动。可以说，社会化媒体的根本价值来自用户参与的广泛性与互动性，参与度的低迷将直接导致用户的流失和平台本身的没落。而只有从理论上和本质上深刻影响用户参与的因素，才能为社会化媒体的实际应用如推荐和搜索提供有意义的指导。 本文从多角度展开了对社会化媒体用户参与度的研究。首先，需要避免千遍一律的枯燥和雷同，单一的内容会让用户乏味而离开，即需要保证多样性；其次，仅有多样性是不够的，必须同时保证内容的相关性和有用性，让用户收获意料之外的发现，即在多样性之上为用户带来眼前一亮的意外惊喜；最后，考虑到多样性和意外惊喜仅提升了用户个体层面的体验，应该继续挖掘用户关系，在网络层面上激发更多互动和共鸣，实现广泛的信息传播，由此，本文进一步对用户之间的影响关系进行深入挖掘以最大化整体参与度。对于以上激励用户参与的三个重要因素—多样性、意外惊喜和影响关系，本文分别展开了以下深入研究。 在多样性与参与度的研究上，以微博为例，本文对社会化媒体用户的个体网络和所读内容的多样性进行了实证研究。首先，使用四种不同的度量方法量化了多样性；之后，对多样性进行了时序分析，发现了微博用户的多样性随着时间增长；最后，考察了多样性与用户参与度的关系，实验发现：结构层面的多样性与原创数量显著正相关，而内容层面的多样性则对原创数量没有太大影响，这说明平台应该有意识地引导用户加入多个不同的圈子；不同度量方式下，转发数都随着多样性的增长而增长，这说明在平台设计中加入多样性元素能有效提升用户的参与度。 在意外惊喜与参与度的研究上，本文首次对意外惊喜现象进行了基于大规模用户行为数据的量化研究，提出了一种识别意外惊喜的高效算法，并计算了意外惊喜在社会化媒体中的存在比例，揭示了其对用户参与度的正面作用。意外惊喜指的是一种非预期的收获或无意中的发现，其在信息系统中对用户体验和用户参与的积极作用已得到了学术界和工业界的普遍认同，但这种作用仍缺乏由大规模数据下的理论研究支持。本文定义社会化媒体中的意外惊喜为“意外的相关性”。在该定义下，基于统计假设检验，本文提出了一种全新的方法来自动、快速、准确识别信息传播中的意外性、相关性和意外惊喜，该方法适用于多种信息系统，如推荐系统、检索系统和广告平台。使用该识别方法，本文计算了意外惊喜在微博信息传播中的存在比例，在Twitter的转发中约占27%，在新浪微博的转发中约占30%。最后，通过相关关系分析和因果关系分析，本文揭示了意外惊喜对社会化媒体中用户参与度（活跃度和社交度）的正面作用。 在影响关系与参与度的研究上，本文利用影响关系提升社会化媒体的整体参与度，抽象并公式化了参与度最大化问题。为了解决此问题，首先，通过随机测试验证了影响关系对用户参与行为的驱动作用；其次，提出了一种迭代算法，根据用户历史交互数据计算用户之间的影响关系；最后，针对参与度最大化问题，提出了一种高效的启发式算法TABI，实验显示该算法在整体参与度的提升上，，性能显著优于推荐算法和社会财富最大化问题的近似算法。基于影响关系的参与度最大化是推荐系统新思路的一种探索，即出于提升整体参与度的考虑，在推荐中不仅需要匹配当前用户的兴趣，还需要考虑当前用户影响力带来的未来参与度。 综上所述，本文深入研究了提高社会化媒体用户参与度的三个关键因素：多样性、意外惊喜和影响关系。实验结果表明，以上三个因素均对用户参与度均产生积极作用。因此，在实际应用和系统设计中，可以借鉴本文提出的算法、技术和框架，在信息内容和用户关系两个层面为用户带来更好的用户体验，从而有效提升社会化媒体的互动程度和参与程度。
[Abstract]:In just a few years, the social media has developed rapidly, the number of users and the coverage rate are constantly being recorded, and the status and role of social life become more and more important. The core of social media is" Socialization "i.e. the user's participation and interaction. It can be said that the basic value of social media comes from the extensive and interactive participation of users, and the downturn of participation will directly affect the loss of users and the loss of the platform itself. Only the factors that profoundly affect the user's participation in theory and nature can provide meaningful guidance for the practical application of social media, such as recommendation and search. In this paper, we expand the participation of social media users from various angles Research. First, there is a need to avoid the boring and thunder of thousands of times, the single content will make the user dull and leave, that is, need to ensure diversity; secondly, only the diversity is not enough, must ensure the relevance and usefulness of the content at the same time, let the user harvest unexpected It has been found that, on the basis of diversity, the user brings an unexpected surprise to the user; finally, taking into account the diversity and unexpected surprise only improves the experience of the individual level of the user, the user relationship should continue to be mined, more interaction and resonance can be stimulated at the network level, and wide information transmission can be realized, Therefore, the relationship between users is further explored in this paper to maximize the whole parameter. With respect to the three important factors, such as diversity, surprise and influence of the above incentive users, this paper respectively expands the following in depth: In this paper, the diversity and the diversity of the individual network and the content of the social media users are studied in the research of diversity and participation. Firstly, the diversity is quantified by four different measurement methods, then the diversity is analyzed in time series, and the diversity of micro-Bo users is found to grow with time. Finally, the relationship between diversity and user participation is investigated. The experiment shows that the diversity of the structure level is positively correlated with the original quantity, while the diversity of the content level has little influence on the original quantity, which means that the platform should consciously guide the user into a plurality of different circles; in different measures, the number of forwarding is varied with diversity. Growth, which suggests that the addition of diversity elements in the platform design can be effectively improved In the research of surprise and participation, this paper makes a quantitative study on unexpected surprises based on large-scale user behavior data for the first time, and presents an efficient algorithm for identifying unexpected surprises, and calculates unexpected surprises in socialization. the presence scale in the media reveals its access to the user, Positive role of engagement. Unexpected surprise refers to a non-expected harvest or unintended discovery that has been universally recognized by academia and industry in the information system, but this role is still lacking in large-scale data The next theoretical research support. This paper defines socialization. Unexpected surprises in the media To Under this definition, based on statistical hypothesis testing, this paper proposes a new method to automatically, quickly and accurately identify the accident, correlation and unexpected surprises in information propagation, which is suitable for various information systems, such as recommendation system and inspection. Cable system and advertising platform. Using this recognition method, this paper calculates the proportion of unexpected surprises in micro-bo information propagation, accounting for about 27% of Twitter's forwarding, Finally, through correlation analysis and causality analysis, this paper reveals the users' participation (activity and society) in social media by surprise surprise. On the research of influence relation and participation, this paper uses influence relation to promote the whole engagement of social media, abstract and public. In order to solve this problem, in order to solve this problem, firstly, the driving effect of the influence relation on the user's participation behavior is verified through the random test; secondly, an iterative algorithm is proposed to calculate the influence relation among users according to the user history interactive data; and finally, an iterative algorithm is proposed. In view of the maximization of engagement, an efficient heuristic algorithm is presented in this paper. The experiment shows that the algorithm is superior to the recommendation algorithm and social wealth in the improvement of the overall engagement. In the recommendation, not only needs to match the interest of the current user, but also needs to take into account the current user. In conclusion, this paper deeply studies the three key factors to improve the participation of social media users. Diversity, surprise and impact. The experimental results show that the above three factors Therefore, in the practical application and the system design, the algorithm, technology and framework proposed in this paper can be used for the user to bring a better user experience at both the information content and the user relation, thus effectively improving the society.
【学位授予单位】：北京大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TP391.3
本文编号：2270619
[Abstract]:In just a few years, the social media has developed rapidly, the number of users and the coverage rate are constantly being recorded, and the status and role of social life become more and more important. The core of social media is" Socialization "i.e. the user's participation and interaction. It can be said that the basic value of social media comes from the extensive and interactive participation of users, and the downturn of participation will directly affect the loss of users and the loss of the platform itself. Only the factors that profoundly affect the user's participation in theory and nature can provide meaningful guidance for the practical application of social media, such as recommendation and search. In this paper, we expand the participation of social media users from various angles Research. First, there is a need to avoid the boring and thunder of thousands of times, the single content will make the user dull and leave, that is, need to ensure diversity; secondly, only the diversity is not enough, must ensure the relevance and usefulness of the content at the same time, let the user harvest unexpected It has been found that, on the basis of diversity, the user brings an unexpected surprise to the user; finally, taking into account the diversity and unexpected surprise only improves the experience of the individual level of the user, the user relationship should continue to be mined, more interaction and resonance can be stimulated at the network level, and wide information transmission can be realized, Therefore, the relationship between users is further explored in this paper to maximize the whole parameter. With respect to the three important factors, such as diversity, surprise and influence of the above incentive users, this paper respectively expands the following in depth: In this paper, the diversity and the diversity of the individual network and the content of the social media users are studied in the research of diversity and participation. Firstly, the diversity is quantified by four different measurement methods, then the diversity is analyzed in time series, and the diversity of micro-Bo users is found to grow with time. Finally, the relationship between diversity and user participation is investigated. The experiment shows that the diversity of the structure level is positively correlated with the original quantity, while the diversity of the content level has little influence on the original quantity, which means that the platform should consciously guide the user into a plurality of different circles; in different measures, the number of forwarding is varied with diversity. Growth, which suggests that the addition of diversity elements in the platform design can be effectively improved In the research of surprise and participation, this paper makes a quantitative study on unexpected surprises based on large-scale user behavior data for the first time, and presents an efficient algorithm for identifying unexpected surprises, and calculates unexpected surprises in socialization. the presence scale in the media reveals its access to the user, Positive role of engagement. Unexpected surprise refers to a non-expected harvest or unintended discovery that has been universally recognized by academia and industry in the information system, but this role is still lacking in large-scale data The next theoretical research support. This paper defines socialization. Unexpected surprises in the media To Under this definition, based on statistical hypothesis testing, this paper proposes a new method to automatically, quickly and accurately identify the accident, correlation and unexpected surprises in information propagation, which is suitable for various information systems, such as recommendation system and inspection. Cable system and advertising platform. Using this recognition method, this paper calculates the proportion of unexpected surprises in micro-bo information propagation, accounting for about 27% of Twitter's forwarding, Finally, through correlation analysis and causality analysis, this paper reveals the users' participation (activity and society) in social media by surprise surprise. On the research of influence relation and participation, this paper uses influence relation to promote the whole engagement of social media, abstract and public. In order to solve this problem, in order to solve this problem, firstly, the driving effect of the influence relation on the user's participation behavior is verified through the random test; secondly, an iterative algorithm is proposed to calculate the influence relation among users according to the user history interactive data; and finally, an iterative algorithm is proposed. In view of the maximization of engagement, an efficient heuristic algorithm is presented in this paper. The experiment shows that the algorithm is superior to the recommendation algorithm and social wealth in the improvement of the overall engagement. In the recommendation, not only needs to match the interest of the current user, but also needs to take into account the current user. In conclusion, this paper deeply studies the three key factors to improve the participation of social media users. Diversity, surprise and impact. The experimental results show that the above three factors Therefore, in the practical application and the system design, the algorithm, technology and framework proposed in this paper can be used for the user to bring a better user experience at both the information content and the user relation, thus effectively improving the society.
【学位授予单位】：北京大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TP391.3
【引证文献】
相关期刊论文 前1条
1 朱长春;李娜;;社会化媒体技术在大学教学中的实践应用研究[J];经济视角(下);2013年12期
本文编号：2270619
本文链接：https://www.wllwen.com/wenyilunwen/guanggaoshejilunwen/2270619.html
