Table 2
Summary of the previous classifiers included in this work.
| Name | Algorithm used | Type | Input | Output | Authors |
|---|---|---|---|---|---|
| CNN1 | Convolutional Neural Network | Classification only | Array of fluxes and errors | conf_*_CNN1 | Rodrigues et al. (2023) |
| CNN1NE | Convolutional Neural Network | Classification only | Array of fluxes | conf_*_CNN1NE | Rodrigues et al. (2023) |
| CNN2 | Convolutional Neural Network | Classification only | 2D image including fluxes and errors | conf_*_CNN2 | Rodrigues et al. (2023) |
| RF | Random Forest | Classification only | Array of fluxes | conf_*_RF | Rodrigues et al. (2023) |
| LGBM | LightGBM | Classification only | Array of fluxes | conf_*_LGBM | Rodrigues et al. (2023) |
| ANN1 | Artificial Neural Network | Classification only | Array of magnitudes + hybridisation | conf_*_ANN1 | Martínez-Solaeche et al. (2023) |
| ANN2 | Artificial Neural Network | Classification only | Array of fluxes + hybridisation | conf_*_ANN2 | Martínez-Solaeche et al. (2023) |
| SQUEzE | Random Forest | Classification and redshift | Specific line metrics | conf_SQUEZE_**, z_SQUEZE_**, class_SQUEZE | Pérez-Ràfols et al. (2023a) |
| QPz | Template-based fitting | Redshift only | Array of fluxes | zphot_QPz | Queiroz, private communication |
| LePhare | Template-based fitting | Redshift only | Array of fluxes | zphot_LePhare | Queiroz, private communication |
Notes. A single * in the output columns represents the different classes and should be replaced by star, gal, and lqso (for low-z quasars) and hqso (for high-z quasars). In the same column, the double ** represents alternative classifications and should be replaced with numbers 0 to 4 (0 being the most confident classification and 4 being the least confident). We refer the reader to the respective papers for details about the architectures and hyperparameters of the codes.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.