Open Access
Table 4.
Explanation of the different types of data sets used for our ML approaches.
| Data set type | Description and purpose |
|---|---|
| Training set | To train the ML models. Light curves of |
| four theoretical SNe Ia models are used. | |
| Validation set | To find the training epoch for the FCNN |
| that has lowest validation loss (four | |
| SNe Ia models as in training process). | |
| (Corresponding) test set | To evaluate the performance of a ML |
| model using four theoretical models as | |
| in the training process. The term | |
| “corresponding” is used if all | |
| parameters (e.g., κ, γ, …) for the | |
| production of the test set are the same | |
| as for the training set. | |
| This data set does not test the | |
| generalizability to different SN Ia | |
| light-curve shapes. | |
| SNEMO15 data set | Final test set using light curves from the |
| empirical SNEMO15 model not used | |
| in the training process, which most | |
| importantly tests the generalizability of | |
| the trained ML models. | |
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.