篇名 | Visual Early-Warning Signal Detection for Critical Transitions |
---|---|
卷期 | 28:2 |
作者 | Hu, Min 、 Jiang, Peng 、 Zhou, Sheng-Chen 、 Sun, Yu-Feng |
頁次 | 001-013 |
關鍵字 | early warnings 、 time series 、 visualization technology 、 EI 、 MEDLINE 、 Scopus |
出刊日期 | 201704 |
DOI | 10.3966/199115592017042802001 |
Critical transitions occur in a wide variety of real systems, and it is very desirable to find early warning signals that a particular system may be approaching an undesired transition. This paper proposes an innovative visual method for critical transition detection, called TDTD (Time-series Data Trajectory Diagram). TDTD is a graphic tool used to time series data analysis, which can visualize the behavior of a trajectory of dynamic system. In order to go beyond the visual impression, Entropy Change Rate per Area (ECRA) is introduced as a universal indicator to analyze TDTD quantitatively, regardless difference in the details of each dynamic system. This method has been tested and compared with other related techniques in some fields, such as climate change and engineering disaster. The results of experiments indicate that TDTD is a valuable approach and reveals a new perspective for early detection of critical transitions.