The collective behavior describing spontaneously emerging social processes and events is ubiquitous in both physical society and online social media. The knowledge of collective behavior is critical in understanding and predicting social movements, fads, riots, and so on. However, detecting, quantifying, and modeling the collective behavior in cascading systems at large scale are seldom explored. In this paper, we examine a real-world online social media with more than 1.7 million information spreading records. We observe evident collective behavior in information cascading systems and then propose metrics to quantify the collectivity. We find that previous information cascading models cannot capture the collective behavior in the real-world data and thus never utilize it. Furthermore, we propose a comprehensive generative framework with a latent user interest layer to capture the collective behavior. Our framework accurately models the information cascades with respect to dynamics, popularity, structure, and collectivity. By leveraging the knowledge behind collective behavior, our model successfully predicts the popularity and participants of information cascades without temporal features or early stage information. Our framework may serve as a more generalized one in modeling cascading systems, and, together with empirical discovery and applications, advance our understanding of human behavior.

在自然社会和在线社交媒体中，描述自发出现的社会过程和事件的集体行为无处不在。集体行为的知识对于理解和预测社会运动，时尚，暴动等至关重要。但是，很少探索在级联系统中对集体行为进行检测，量化和建模。在本文中，我们研究了具有超过170万条信息传播记录的现实世界在线社交媒体。我们观察到信息级联系统中明显的集体行为，然后提出度量标准以量化集体性。我们发现以前的信息级联模型无法捕获现实世界数据中的集体行为，因此从不利用它。此外，我们提出了一个具有潜在用户兴趣层的综合生成框架，以捕获集体行为。我们的框架可以准确地建模有关动态，受欢迎程度，结构和整体性的信息级联。通过利用集体行为背后的知识，我们的模型成功地预测了信息的流行程度和参与程度，而没有时间特征或早期信息。我们的框架可以用作级联系统建模的更通用的框架，并且与经验发现和应用一起，可以增进我们对人类行为的理解。我们的模型成功地预测了没有时间特征或早期信息的信息级联的受欢迎程度和参与者。我们的框架可以用作级联系统建模的更通用的框架，并且与经验发现和应用一起，可以增进我们对人类行为的理解。我们的模型成功地预测了没有时间特征或早期信息的信息级联的受欢迎程度和参与者。我们的框架可以用作级联系统建模的更通用的框架，并且与经验发现和应用一起，可以增进我们对人类行为的理解。