运动数据的复杂领导-追随动态挖掘与建模
Mining and modeling complex leadership-followership dynamics of movement data
摘要：
领导和追随是社会性动物（包括人类）集体决策和组织的重要组成部分。在本质上，领导者和追随者的关系是动态的，并随着上下文或时间因素而变化。了解领导和追随者的动态，例如领导者和追随者如何改变，出现或收敛，使科学家能够更深入地了解群体决策和集体行为。然而，仅给出个人活动的数据，推断领导者和追随者的动态是具有挑战性的。在本文中，我们专注于挖掘和建模频繁模式的领先和跟随。我们正式提出了新的计算问题，并提出了一个框架，可用于解决有关群体运动的几个问题。我们使用的领导推理框架，mFLICA，从运动数据集推断的时间序列的领导者和他们的派别，然后提出了一种方法来挖掘和模型的领导和追随者的动态频繁模式。我们使用几个模拟数据集，以及狒狒运动的真实世界的数据集，以证明我们的框架的应用程序来评估我们的框架的性能。这些都是新的计算问题，据我们所知，没有现有的可比方法来解决这些问题。因此，我们修改和扩展现有的领导推理框架，以提供一个非平凡的基线比较。我们的框架在所有数据集中的表现都比这个基线更好。我们的框架为科学家们提供了机会，可以生成关于运动数据中领导力动态的可验证的科学假设。
英文摘要：
Leadership and followership are essential parts of collective decision and organization in social animals, including humans. In nature, relationships of leaders and followers are dynamic and vary with context or temporal factors. Understanding dynamics of leadership and followership, such as how leaders and followers change, emerge, or converge, allows scientists to gain more insight into group decision-making and collective behavior in general. However, given only data of individual activities, it is challenging to infer the dynamics of leaders and followers. In this paper, we focus on mining and modeling frequent patterns of leading and following. We formalize new computational problems and propose a framework that can be used to address several questions regarding group movement. We use the leadership inference framework, mFLICA, to infer the time series of leaders and their factions from movement datasets and then propose an approach to mine and model frequent patterns of both leadership and followership dynamics. We evaluate our framework performance by using several simulated datasets, as well as the real-world dataset of baboon movement to demonstrate the applications of our framework. These are novel computational problems and, to the best of our knowledge, there are no existing comparable methods to address them. Thus, we modify and extend an existing leadership inference framework to provide a non-trivial baseline for comparison. Our framework performs better than this baseline in all datasets. Our framework opens the opportunities for scientists to generate testable scientific hypotheses about the dynamics of leadership in movement data.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
