CardioBERTa.sv_GP_enriched / training_data_stats.json
fernandogd's picture
Release CardioBERTa-based clinical concept disambiguation model
5f1c11a verified
Raw History Blame Contribute Delete
977 Bytes
{
"triplets": 4915580,
"malformed_rows": 0,
"unique_triplets": 4862249,
"duplicate_triplets": 53331,
"unique_cuis": 476972,
"unique_anchors": 482198,
"unique_positives": 447784,
"unique_terms": 539362,
"unique_anchor_positive_pairs": 4814909,
"ambiguous_terms_across_cuis": 196356,
"self_pairs_exact": 0,
"self_pairs_normalized": 3863,
"cuis_with_multiple_terms": 476972,
"terms_per_cui_mean": 9.887574113365146,
"terms_per_cui_median": 7.0,
"terms_per_cui_p95": 28.0,
"terms_per_cui_max": 36,
"triplets_per_cui_mean": 10.305804114287632,
"triplets_per_cui_median": 7.0,
"triplets_per_cui_p95": 33.0,
"triplets_per_cui_max": 35,
"anchor_words_mean": 3.843145468083115,
"positive_words_mean": 3.9231073444028985,
"cuis_common_with_synonyms": 80858,
"cuis_added_vs_synonyms": 396114,
"cuis_missing_vs_synonyms": 0,
"terms_common_with_synonyms": 155374,
"terms_added_vs_synonyms": 383988,
"terms_missing_vs_synonyms": 0
}