题目：Causation analysis in dynamical systems

报告人：林伟教授  复旦大学

摘要：Existing methods on causation detection often misidentify indirect causal influences as direct ones, due to the effect of causation transitivity. We solve this problem by developing a method based on an articulated integration of three basic tools from nonlinear dynamics and statistics. We demonstrate our method by using data from a number of real-world systems. As direct causal links are key to the fundamental dynamical underpinnings of a variety of complex systems, our PCM method will become a useful tool to unlock from data the inner mechanisms of real systems in diverse disciplines. In addition to simultaneous causation between variables, the time delayed causation is also taken into account in our investigations.

时间：2018年5月23日 10:00-11:00

地点：知新楼B1032

邀请人：陈章