2021ApJ...908..177K

We utilize machine-learning methods to distinguish BL Lacertae (BL Lac) objects from flat spectrum radio quasars (FSRQs) within a sample of likely X-ray blazar counterparts to Fermi 3FGL unassociated gamma-ray sources. From our previous work, we have extracted 84 sources that were classified as >=99% likely to be blazars. We then utilize Swift-XRT, Fermi, and The Wide-field Infrared Survey Explorer (WISE) data together to distinguish the specific type of blazar, FSRQs, or BL Lac objects. Various X-ray and gamma-ray parameters can be used to differentiate between these subclasses. These are also known to occupy different parameter space on the WISE color-color diagram. Using all these data together would provide more robust results for the classified sources. We utilized a random forest classifier to calculate the probability for each blazar to be associated with a BL Lac object or an FSRQ. Based on P_bll, which is the probability for each source to be a BL Lac object, we placed our sources into five different categories based on this value as follows: P_bll >= 99%: highly likely BL Lac object, P_bll >= 90%: likely BL Lac object, P_bll <= 1%: highly likely FSRQ, P_bll <= 10%: likely FSRQ, and 90% < P_bll < 10%: ambiguous. Our results categorize the 84 blazar candidates as 50 likely BL Lac objects and the other 34 as being ambiguous. A small subset of these sources have been listed as associated sources in the most recent Fermi catalog, 4FGL, and in these cases our results are in agreement on the classification.