Tests for time series data using recurrence plots

本研究的目的是提出使用递归图的时间序列数据的测试。特别是，我们提出的方法，通过使用概率和/或拓扑性质的基础动力学特征。所提出的测试包括平稳性，串行依赖，周期性，平稳线性，确定性混沌和同步。我们使用其特征众所周知的玩具模型验证了这些测试，并将测试应用于各种真实的数据集。

This research aimed at proposing tests for time series data using recurrence plots. Especially, we proposed methods that characterize the underlying dynamics by using the probabilistic and/or topological properties. The proposed tests include those for stationarity, serial dependence, periodicity, stationary linearity, deterministic chaos, and synchronization. We validated these tests using toy models whose characteristics are well known, and applied the tests to various real datasets.