# Performant feature extraction for photometric time series

**Authors:** Anastasia Lavrukhina, Konstantin Malanchev  
**arXiv:** [2302.10837](https://arxiv.org/abs/2302.10837)  
**Date:** 2023-02-21  
**Summary page:** https://www.summarizepaper.com/en/paper/2302.10837v1/

## Key points

- Astronomy is entering a new era of large surveys of the variable sky
- Zwicky Transient Facility (ZTF) and Legacy Survey of Space and Time (LSST) are expected to produce up to a million alerts per night
- Efficient light-curve pre-processing algorithms are required for subsequent data quality cuts, classification, and characterization analysis
- Lavrukhina and Malanchev's "light-curve" library provides an efficient solution for feature extraction from light curves in large astronomical surveys like ZTF and LSST
- The library outperforms other publicly available codes while providing dozens of features describing shape, magnitude distribution, and periodic properties of light curves
- It includes not only features that have been shown to provide high performance in classification tasks but also new features developed to improve the classification quality of selected types of objects
- The library performs well on real-world data as demonstrated by its usage in three broker systems: ANTARES, AMPEL, and Fink which analyze the ZTF alert stream
- The "light-curve" library has been selected for use with LSST
- Its successful application in real-world scenarios makes it a valuable tool for future astronomical research projects.

## Summary

Astronomy is entering a new era of large surveys of the variable sky, such as the Zwicky Transient Facility (ZTF) and the forthcoming Legacy Survey of Space and Time (LSST), which are expected to produce up to a million alerts per night. The sheer amount of photometric data generated by these surveys requires efficient light-curve pre-processing algorithms for subsequent data quality cuts, classification, and characterization analysis. Lavrukhina and Malanchev's "light-curve" library provides an efficient solution for feature extraction from light curves in large astronomical surveys like ZTF and LSST. Its fast implementation outperforms other publicly available codes while providing dozens of features describing shape, magnitude distribution, and periodic properties of light curves. It includes not only features that have been shown to provide high performance in classification tasks but also new features developed to improve the classification quality of selected types of objects. The authors demonstrate that their library performs well on real-world data by showing its usage in three broker systems: ANTARES, AMPEL, and Fink, which analyze the ZTF alert stream. Moreover, the "light-curve" library has been selected for use with LSST. Its successful application in real-world scenarios makes it a valuable tool for future astronomical research projects.