Classification of CSC2.1.1 unknown X-ray sources from "Self-Organizing Maps for the Exploration and Classification of X-Ray Sources in the Chandra Source Catalog"

Classifcation of a sample of unknown X-ray detections extracted from the Chandra Source Catalog (CSC) version 2.1.1 based on the methodology presented in "Self-Organizing Maps for the Exploration and Classification of X-Ray Sources in the Chandra Source Catalog".The classification is based on a classified built on the two-dimensional mapping of a sample of X-ray detections from CSC2.1.1 in a feature space generated by spectral and time variability properties. The two available tables are: 1. dataset_per_detection_classification.csv List of per-detection classification. Columns: id <---- identifier of the detectionbmu_x <---- x coordinate of the Best Matching Unit (BMU) on the SOM grid.bmu_y <---- y coordinate of the Best Matching Unit (BMU) on the SOM grid.pred <---- Predicted source class for the detection.conf <---- Posterior confidence of the predicted class.abstain <---- Boolean flag indicating whether the classifier abstained for the detection.source_name <---- CSC source name associated with the detection.truncated_R <---- Boolean flag indicating whether the hexagonal neighbourhood used for classification extends beyond the SOM boundary. 2. dataset_per_source_classification.csv List of per-source classification.Columns: source_name <---- CSC source named_detections <---- Total number of detections associated with the source.n_classified <---- Number of detections with a valid, non-abstained classification.n_abstained <---- Number of detections for which the classifier abstained.final_classification <---- Source-level classification obtained from the most frequent valid detection-level classification. If multiple classes are tied, the tie is resolved using the average confidence of the tied classes.agreement_count <---- Number of valid detections assigned to the source-level final_classification.agreement_rate <---- Fraction of valid, non-abstained detections assigned to the final_classification.n_unique_classes <---- Number of distinct predicted classes.all_classes <---- Class-count mapping for the valid detection-level classifications, listing each predicted class and the number of detections assigned to it.avg_confidence_final_class <---- Mean confidence of the valid detections assigned to the final_classification.avg_confidence_all <---- Mean confidence of all valid, non-abstained detection-level classifications for the source.heterogeneity_score <---- Score used to order sources according to the disagreement and class diversity among their valid detection-level classifications. It is calculated as: n_classified × (1 − agreement_rate)^2 × n_unique_classes.