Kun Li is currently a P.h.D student in Tianjin University (TJU). He received his MS, BS degrees of Computer Science and Technology in 2019 and 2016 from Tianjin University, respectively. His main research interests include Parallel computing, Time domain astronomy, Machine learning, and Deep learning.
Storage Schema of Time Series Astronomical Data for Artificial Intelligence Analysis
Time series data is commonly used in time domain astronomy. Traditional manual processing methods are becoming extremely hard and infeasible during the dramatic growth of data volume. Artificial intelligence (AI) techniques provide the possibility for exploring the entire dataset but require that the time series data of all celestial objects in the dataset must be prepared in advance. We have designed a special tool called AstroCatR for reconstructing astronomical time series data from large-scale catalogues. In this paper, we focus on challenges in time series metadata structure, storage schemes, strategies and formats of exposition. A recommendation unified time series data (light curves) release standard needs to be proposed for analysis of time series data based on AI techniques.