Supernova scores for active anomaly detection

Data description for (link to article) The dataset consists of the following files: -------------------------------------------------------------------------------- sn_clf_train_data.csv Training dataset for the SN classifier.Each object is described by initial set features and includes the following metadata: object identifier (oid) SN score real–bogus score -------------------------------------------------------------------------------- features.zip archive containing dr23 features. features/├── augmented/│ ├── feature_augmented.name│ ├── dr23_feature_{FIELDID}.dat│ └── ...└── initial/ ├── feature_initial.name ├── dr23_feature_{FIELDID}.dat └── ...-------------------------------------------------------------------------------- oids.zip archive containing object identifiers. oids/├── dr23_oid_{FIELDID}.dat└── ... -------------------------------------------------------------------------------- features/initial/dr23_feature_{FIELDID}.dat contains light curve feature data for objects, where each object is represented by 47 feature values. These values are encoded as little-endian single-precision IEEE-754 floating-point numbers (32-bit floats). Feature names are listed in the plain text file features/initial/feature_initial.name, with one name per line.oids/dr23_oid_{FIELDID}.dat contains ZTF DR23 object identifiers, encoded as little-endian 64-bit unsigned integers. features/augmented/dr23_feature_{FIELDID}.dat contains the same features as features/initial/dr23_feature_{FIELDID}.dat, but with an additional column representing the binary supernova classifier prediction. Each object in this file corresponds to 48 features: the original 47 features plus 1 additional feature. Feature names for this file are provided in features/augmented/feature_augmented.name. The files oids/dr23_oid_{FIELDID}.dat, features/initial/dr23_feature_{FIELDID}.dat, and features/augmented/dr23_feature_{FIELDID}.dat share the same object order. Below is a sample Python script for accessing the data using NumPy: import numpy as np # Load object IDsoid = np.memmap('oids/dr23_oid_{FIELDID}.dat', mode='r', dtype=np.uint64) # Load features and reshapefeature = np.memmap('features/initial/dr23_feature_{FIELDID}.dat', mode='r', dtype=np.float32).reshape(oid.shape[0], -1) # Print dataset informationprint(f'Number of objects: {len(oid)}')print(f'Features shape: {feature.shape}')