Classification of the <i>Fermi</i>-LAT blazar candidates of uncertain type using extreme gradient boosting

Abstract Machine learning based approaches are emerging as very powerful tools for many applications including source classification in astrophysics research due to the availability of huge high quality data from different surveys observational astronomy. The Large Area Telescope on board Fermi satellite (Fermi-LAT) has discovered more than 6500 energy gamma-ray sources sky its survey over a decade. A significant fraction observed by Fermi-LAT either remains unassociated or been identified Blazar Candidates Uncertain type (BCUs). We explore potential eXtreme Gradient Boosting (XGBoost)- supervised machine algorithm identify blazar subclasses among sample 112 BCUs 4FGL catalog whose X-ray counterparts available within 95% uncertainty regions observations. have used information multi-wavelength observations IR, optical, UV, and γ-ray wavebands along with redshift measurements reported literature classification. Among uncertain blazars, 62 classified BL Lacertae objects (BL Lacs) 6 Flat Spectrum Radio Quasars (FSRQs). This indicates improvement respect multi-perceptron neural network literature. Our study suggests that spectral index, IR color indices most important features identifying using XGBoost classifier. also importance BCU candidates.