近年来，互联网的普及和发展，深刻地改变了众多传统行业。泛娱乐产业链作为“互联网+文化娱乐”的新业态，包括游戏、影视、动漫、音乐等多重媒介广泛互联并深度融合，逐渐成为数字经济发展的重要支柱。而且，其中孕育出的精品IP通过不同的内容表现形式，满足了粉丝的多元化需求，进而为整个文化娱乐产业链注入了新的动力。在线社会化媒体平台作为用户分享热点问题和领域话题的主要阵地，为公众讨论文娱领域相关知识搭建了一个重要平台。

      本研究基于知识协同建构理论，以2011至2018年从“知乎”平台上抓取的所有文娱标签话题数据为对象，运用大数据分析技术，文本挖掘、社会网络分析等方法将文娱领域内的知识标签量化为标签共现网络开展实证研究，旨在描述以“知乎”为代表的在线问答社区文娱标签的内容结构和演化特征。并以“漫威”IP文娱作品为个案，从标签结构和网络演化两大维度来研究具有IP属性的文娱热点标签其知识资源的建构和发展规律。

      首先，在文娱标签的内容与结构层面，将八年内所有标签节点经过去噪后进行社团划分，发现文娱标签网络具有良好的社区聚类特征，每年都含有的主题词包括游戏、摄影、设计、影视、文学、音乐、动漫、新闻媒体八大类别。其次，在文娱标签的演化规律层面，通过对文娱知识标签的生成与消亡及其在社区间的流动过程进行研究，揭示出文娱知识协同生产的基本过程。最后，在漫威文娱作品个案层面，通过可视化呈现出知乎平台中漫威文娱标签的节点生成过程，以探究其知识分享的规律和特征。

      本研究具有一定的理论、方法和实践意义。在理论贡献层面，厘清了“知识协同建构理论”视角下文娱标签的内容、结构特征及其标签的演化规律。在方法论层面，基于大数据分析技术，运用文本挖掘、Louvain社区发现算法和Pathfinder关键路径算法，对大量有关文娱领域的知识标签文本进行分析和可视化呈现，有效规避了传统研究方法所产生的主观偏差和经验论断。在实践意义层面，综合分析了大众对文娱产品讨论的内在规律，实现了对问答社区内由用户协同生产的文娱知识图谱的构建。

    In recent years, the popularization and development of the Internet has profoundly changed many traditional industries. As a new format of "Internet + cultural entertainment", the pan-entertainment industry chain, including games, film and television, animation, music and other media, is widely interconnected and deeply integrated, and has gradually become an important pillar of the development of the digital economy. Moreover, the high-quality IP nurtured in them meet the diversified needs of fans through different content expressions, thereby injecting new impetus into the entire cultural and entertainment industry chain. Online social media platforms serve as the main platform for users to share hot issues and field topics, and build an important platform for the public to discuss related knowledge in the field of entertainment.

    Based on the theory of "collaborative knowledge building", this research takes all the topic data of entertainment labels captured from the "Zhihu" platform from 2011 to 2018, and uses big data analysis technology, text mining, social network analysis and other methods to quantify knowledge labels in the field of entertainment into label co-occurrence networks to carry out empirical research, aiming to describe the content structure and evolution characteristics of entertainment labels in online Q&A communities represented by "Zhihu". And take "Marvel" IP entertainment works as a case, from the two dimensions of label structure and network evolution, to study the construction and development law of knowledge resources of entertainment hotspot labels with IP attributes.

    First of all, in terms of the content and structure of entertainment labels, all label nodes within eight years are divided into communities after denoising. It is found that the entertainment label network has good community clustering characteristics. The themes of entertainment labels every year include eight categories: games, photography, design, film and television, literature, music, animation, and news media. Secondly, at the level of the evolution characteristics of entertainment labels, the basic process of collaborative production of entertainment knowledge is revealed by studying the generation and extinction of cultural and entertainment knowledge labels and their flow process among communities. Finally, at the case level of Marvel Entertainment's works, the node generation process of the Marvel Entertainment label in the Zhihu platform is visualized to explore the rules and characteristics of its knowledge sharing.

    This research has certain theoretical, method and practical significance. At the level of theoretical contribution, it clarifies the content and structural characteristics of entertainment labels and the evolution law of labels from the perspective of "collaborative knowledge building"theory. At the methodological level, based on big data analysis technology, text mining, Louvain community discovery algorithm and Pathfinder critical path algorithm are used to analyze and visualize a large number of knowledge label texts in the field of entertainment, effectively avoiding the subjective bias and empirical conclusions generated by traditional research methods. At the level of practical significance, this paper comprehensively analyzes the internal laws of the public's discussion of entertainment products, and realizes the construction of the entertainment knowledge graph produced by users in the Q&A community.