Lunch talk on Oct. 14, 2024

Machine Learning-based Study of High-redshift Quasars

Speaker: Guangping Ye (HUST)

Venue: SWIFAR Building 2111

Time: 12:45 PM, Monday, Oct. 14, 2024

Abstract: 

High-redshift ( > 5) quasars (which are driven by supermassive black holes at the centers of their host galaxies through the process of accretion) provide effective probes for the study of galaxy evolution and cosmology, hence it is critical to obtain a large sample of high-redshift quasars to study the intergalactic medium, circumgalactic medium and the co-evolution of supermassive black holes and their host galaxies. We present a machine learning search for high-redshift (5.0 < z < 6.5) quasars using the combined photometric data from the DESI Imaging Legacy Surveys and the WISE survey. We discuss the imputation of missing values for high-redshift quasars, analyze the feature selections, compare different machine learning algorithms, and investigate the selections of class ensemble for the training sample. The 11-class random forest model can achieve a precision of 96.43% and a recall of 91.53% for high-redshift quasars for the test set. Using MUSE and DESI-EDR public spectra for verification, it returns excellent results. Additionally, we estimate photometric redshift for the high-redshift quasar candidates using random forest regression model with a high precision. We also demonstrate that the deeper images and more photometric measurements from the future imaging surveys such as CSST, RST, EST and LSST would significantly improve the performance of the random forest model. Subsequently, we can probe the lunminosity function of high-redshift quasars at the faint end utilizing the gravitational effect via crocc-matching our high-redshift quasar candidates to the clusters with known lensing model.

Report PPT:  SWIFAR_Guangping Ye.pdf