Back to Volume
Paper: Clustering Analysis on LAMOST Galaxy Spectra
Volume: 522, Astronomical Data Analysis Software and Systems XXVII
Page: 409
Authors: Zhang, J.; Zhang, X.; Wu, Y.; Chen, X.
Abstract: Using the k-means cluster analysis algorithm, we carry out an unsupervised classification of galaxy spectra of LAMOST DR3 data release. The cluster analysis procedure on LAMOST galaxy spectra has two steps. The first step k-mean cluster utilizes the continuum normalized galaxy spectra, and all the galaxy spectra are classified into three main types: emission line galaxy spectra group, absorbing line galaxy spectra group and the composite galaxy spectra group. The second step k-means clustering clusters the three groups respectively and builds 22 sub-classes in all. We find correlation between the clustering classes and the characteristics of the galaxy templates built by Anne Kinney. There is correlation between clustering classes and morphological types with some intrinsic scatter. The clustering galaxy classes have been applied in the processing and analysis pipeline of the LAMOST data, including fast determination of the galaxy type, and the galaxy identification and redshift measurement.
Back to Volume