CardioBERTa.cs_P_enriched / training_data_stats.json
fernandogd's picture
Release CardioBERTa-based clinical concept disambiguation model
1a9132c verified
Raw History Blame Contribute Delete
971 Bytes
{
"triplets": 1592861,
"malformed_rows": 0,
"unique_triplets": 1591745,
"duplicate_triplets": 1116,
"unique_cuis": 476184,
"unique_anchors": 395754,
"unique_positives": 394884,
"unique_terms": 526263,
"unique_anchor_positive_pairs": 1572775,
"ambiguous_terms_across_cuis": 192127,
"self_pairs_exact": 0,
"self_pairs_normalized": 185,
"cuis_with_multiple_terms": 476175,
"terms_per_cui_mean": 3.918397090200427,
"terms_per_cui_median": 3.0,
"terms_per_cui_p95": 9.0,
"terms_per_cui_max": 36,
"triplets_per_cui_mean": 3.345053592728861,
"triplets_per_cui_median": 2.0,
"triplets_per_cui_p95": 9.0,
"triplets_per_cui_max": 35,
"anchor_words_mean": 3.955929613443985,
"positive_words_mean": 4.184335607438439,
"cuis_common_with_synonyms": 68973,
"cuis_added_vs_synonyms": 407211,
"cuis_missing_vs_synonyms": 0,
"terms_common_with_synonyms": 135148,
"terms_added_vs_synonyms": 391115,
"terms_missing_vs_synonyms": 0
}