Discern Misclassified flat-spectrum radio quasars from low-frequency peaked BL Lacertae objects

A sample of 312 low-frequency peaked BL Lacertae objects (LBLs) and 694 flat spectrum radio quasars (FSRQs) with the parameters both redshift and $\gamma$-ray photon spectral index ($\Gamma _\gamma$) is compiled from the active galactic nuclei (AGNs) Catalog Data Release 2 (4LAC-DR2) from Fermi-LAT. The multi-wavelength data of the sample sources are downloaded from the Space Science Data Center (SSDC), and then match the corresponding gamma-ray data from 4FGL-DR2. The synchrotron radiation peak frequency and Compton dominance (CD) parameters of the sources are obtained by using a log-parabolic to fit the average-state multi-wavelength spectral energy distribution. A support vector machine (SVM) in the $\log L_\gamma$-$\Gamma_\gamma$ frame is utilized to delineate the optimal boundary between FSRQs and LBLs sources. The 1$\sigma$ position of the Gaussian fitting on the histograms of the $\Gamma_\gamma$, $\log \nu^{syn}_{peak}$, and CD parameter distributions are also introduced. In the criterion, 25 FSRQ candidates are selected from LBL sample sources. The optical spectral identification result confirms that 8 out of 13 candidate sources available with the optical spectral data exhibit the relationship of $EW > 5 \mathring{\mathrm{A}}$.