CardioBERTa.en_GP_only_snomed / training_data_stats.json
fernandogd's picture
Release CardioBERTa-based clinical concept disambiguation model
8887aea verified
Raw
History Blame Contribute Delete
751 Bytes
{
"triplets": 3760671,
"malformed_rows": 0,
"unique_triplets": 3760637,
"duplicate_triplets": 34,
"unique_cuis": 477292,
"unique_anchors": 385676,
"unique_positives": 406841,
"unique_terms": 470264,
"unique_anchor_positive_pairs": 3740672,
"ambiguous_terms_across_cuis": 162282,
"self_pairs_exact": 0,
"self_pairs_normalized": 4,
"cuis_with_multiple_terms": 477292,
"terms_per_cui_mean": 8.879103358112015,
"terms_per_cui_median": 6.0,
"terms_per_cui_p95": 26.0,
"terms_per_cui_max": 36,
"triplets_per_cui_mean": 7.879182973944671,
"triplets_per_cui_median": 5.0,
"triplets_per_cui_p95": 25.0,
"triplets_per_cui_max": 35,
"anchor_words_mean": 4.134149730194426,
"positive_words_mean": 4.347405556082943
}