Dataset for "Identifying and Characterizing Ultracool Spectral Blend Binaries with Machine Learning Methods"

singles_templates_250608.h5 can be read as a dataframe table. The dataframe contains the 1006 true singles used to create the synthetic singles and synthetic binaries.It contains the following columns: 'NAME', 'WAVE', 'FLUX', 'UNCERTAINTY', 'SPT', 'J_SNR' The NAME column contains strings with the name of each single The WAVE column contains numpy arrays with the wavelength grid The FLUX column contains numpy arrays with the normalized flux The UNCERTAINTY column contains numpy arrays with the uncertainty The SPT column contains strings with the spectral type The J_SNR column contains integers with the Signal to Noise ratio on the J-band (between 1.2 and 1.35 microns) SinDF_low/mid/hig.pickle can be read as a dataframe table. These files contain the synthetic singles used as train/test data for the RF models. 'system_interpolated_flux': numpy arrays with the normalized flux 'system_interpolated_noise': numpy arrays with the normalized flux 'SPT': string of the spectral type 'SPT_NUM': float of the spectral type 'J_SNR': integer of the Signal to Noise ratio on the J-band (between 1.2 and 1.35 microns) 'SNR_CLASS': string indicating if J_SNR falls in the 'low', 'mid', 'hig' category 'WAVE': numpy arrays with the wavelength grid (length=409) 'flux_0' - 'flux_408': float - the individual points of the flux array. Can be used as features for the RF models BinDF_low/mid/hig.pickle files contain the synthetic binaries used as train/test data for the RF models. 'system_interpolated_flux': numpy arrays with the normalized flux 'system_interpolated_noise': numpy arrays with the normalized flux 'primary_type': integer of the primary spectral type 'secondary_type': integer of the primary spectral type 'J_SNR': integer of the Signal to Noise ratio on the J-band (between 1.2 and 1.35 microns) 'SNR_CLASS': string indicating if J_SNR falls in the 'low', 'mid', 'hig' category 'WAVE': numpy arrays with the wavelength grid (length=409) 'flux_0' - 'flux_408': float - the individual points of the flux array. Can be used as features for the RF models the _d files contain the same synthetic single/binary systems as their non-_d counterparts (e.g., SinDF_low = SinDF_low_d), except for the extra columns. Because of this, the _d versions are larger in size.If you don’t plan to train BId or BClass 2, 4, 6, or 8, you can skip downloading the _d files. 'difference_spectrum': numpy array of the flux of the standard MINUS the flux of the binary (length=409) 'diff_0' - 'diff_408': float - the individual points of the difference_spectrum array. Can be used as features for the RF models BId1- BId8 contain the Binary Identification models. BClass1 - BClass8 contain the Bindary Classification models. You can open the example_notebook.ipynb to learn how to open and use all of these files.