arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2004.00160stat.MEstat.CO
收藏

动物运动建模中存在测量误差的运动-静止过程

Moving-Resting Process with Measurement Error in Animal Movement Modeling

Chaoran Hu, Mark Elbroch, Thomas Meyer, Vladimir Pozdnyakov, Jun Yan

摘要:

动物运动的统计建模是至关重要的。动物运动的连续轨迹只能在离散的、通常间隔不规则的时间点观察到。大多数现有的模型不能自然地处理不等的采样间隔和/或不允许诸如休息或睡眠的不活动时段。最近提出的运动-静止(MR)模型是一种由电报过程控制的布朗运动,它允许电报过程的一个状态中有一段不活动的时间。MR模型在模拟具有长不活动期的捕食者(如许多猫科动物)的运动方面显示出了希望,但缺乏对测量误差的适应性严重阻碍了其在实际中的应用。在这里,我们将测量误差的MR模型,并得出模型的基本属性。推理是基于一个复合似然使用的马尔可夫性的链组成的每隔一个观察到的增量。在有限样本仿真研究中验证了该方法的性能。应用于怀俄明州的山狮的运动数据说明了该方法的实用性。

英文摘要:

Statistical modeling of animal movement is of critical importance. The continuous trajectory of an animal's movements is only observed at discrete, often irregularly spaced time points. Most existing models cannot handle the unequal sampling interval naturally and/or do not allow inactivity periods such as resting or sleeping. The recently proposed moving-resting (MR) model is a Brownian motion governed by a telegraph process, which allows periods of inactivity in one state of the telegraph process. The MR model shows promise in modeling the movements of predators with long inactive periods such as many felids, but the lack of accommodation of measurement errors seriously prohibits its application in practice. Here we incorporate measurement errors in the MR model and derive basic properties of the model. Inferences are based on a composite likelihood using the Markov property of the chain composed by every other observed increments. The performance of the method is validated in finite sample simulation studies. Application to the movement data of a mountain lion in Wyoming illustrates the utility of the method.

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。

↑