为对用户新闻阅读行为进行剖析，根据2019年某典型新闻APP中用户所产生二百余万条非算法推荐及特色栏目行为日志记录，基于流网络对用户集体注意力规律及其时间分布进行了研究。研究发现：硬新闻在用户主动阅读中占较大比例，集体层面用户在时间维度上未呈现明显的新闻阅读行为习惯，但不同类别新闻集体注意力特征相差极大，用户开始阅读时会在“时事”类和“热点”类新闻上分配最多注意力，且更倾向于持续阅读以“文化”类新闻为代表的某类新闻；同时用户很大概率会主动选择多种类别的新闻资源，并在晚上相对频繁地在不同新闻类别之间跳转。本研究从“能量流动”的视角采用复杂网络分析方法更深层次地挖掘了新闻APP中的用户新闻阅读行为规律，从集体的角度对于用户的注意力流转规律加以剖析，给予大众传播新闻分发过程建议，为互联网时代传统媒体转型提供了新视角的依据。

In order to analyze the user's news reading behavior, this study mined the rules and time distribution of users' collective attention based on flow network, by more than 2 million non-algorithmic users' recommendation and characteristic column behavior logs generated in a typical news APP in 2019. The results show that: Hard news accounts for a large proportion of users' active reading. At the collective level, users do not show obvious news browsing habits in the time dimension, but the features of collective attention vary greatly among different news categories. When users start using app, they will allocate the most collective attention to current affairs and hot news, and they are more inclined to continue to pay attention to a certain type of news represented by culture news. Also, users are likely to choose a variety of news and jump between different categories more frequently in the evening. From the perspective of "energy flows", this study uses the complex network analysis method to dig out the rules of user news reading behavior, and analyzes the user attention flow rules from the collective perspective, which gives advice on the mass media news distribution process and provides a new perspective for the transformation of traditional media in the Internet era.