An approach to represent time series by patterns and process dependent statistics to present properties of signals
Sunjung Kim, Chulhyun Park, Soohyun Chae, Min-kyu Sohn, Chang-Ho Lee, Young-Hoon Kim, Jae-Yong Park · 2023
This paper proposes a method to manage time series which is sensor raw traces based on process and various statistics in order to express information that cannot be explained by basic statistics. We present seven algorithms that represent features of signals properly and explain what kind of characteristics those algorithms reflect. This paper also proposes a procedure from determining a pattern based on process to choosing an algorithm. This method will be the basis for managing the status of equipment.