Outlier celestial body classification method based on astronomical spectrum data

Abstract

The invention provides an outlier celestial body classification method based on astronomical spectral data, and the method comprises the steps: obtaining the original astronomical spectral data in batches, taking the astronomical spectral data with known labels as a training set and a test set, and enabling the unknown astronomical spectral data to be classified, carrying out dimension reduction processing on the high-dimensional feature vector of the obtained original astronomical spectrum data, calculating an outlier score according to the obtained low-dimensional feature vector, and carrying out binary classification on the astronomical spectrum data; constructing a triple sample; inputting unknown astronomical spectrum data subjected to dimension reduction processing into the triple random loss neural network model to obtain a new feature vector, obtaining a new feature vector of the unknown spectrum data, calculating an outlier score of the data, and identifying outlier data in the astronomical spectrum data by using a new threshold set after training; the method has the advantages of being short in data processing time, high in accuracy, capable of rapidly recognizing the outlier celestial body and the like.

Description

technical field [0001] The invention relates to a method for classifying outlier celestial bodies based on astronomical spectrum data. Background technique [0002] Astronomy With the development of science and technology, advanced observation equipment enables us to look deeper into the universe, and at the same time brings an explosive growth of astronomical data. As the telescope with the highest spectral acquisition rate in the world, the Guo Shoujing Telescope (LAMOST) can collect more than 10,000 spectra per observation night. The formation and evolution of galaxies, as well as spectral observations combined with various rays provide a large amount of material, which promotes and improves the development of the field of astronomy. Each spectrum in the LAMOST dataset provides a range of radiant intensity values ​​in the wavelength range 3690-9100 angstroms. [0003] Spectral classification is to select and extract the most effective features for classification and rec...

AI Technical Summary

Problems solved by technology

This phenomenon may be due to the contamination of the spectral data by factors such as cosmic background noise, light attenuation, and red shift, making the data unrecognizable; it may also be the spectral data of unknown celestial bodies that have never been observed before

Images