CardioBERTa.en_GP_enriched / training_data_stats.json
fernandogd's picture
Release CardioBERTa-based clinical concept disambiguation model
706a4a8 verified
Raw History Blame Contribute Delete
965 Bytes
{
"triplets": 4952020,
"malformed_rows": 0,
"unique_triplets": 4951542,
"duplicate_triplets": 478,
"unique_cuis": 477293,
"unique_anchors": 467097,
"unique_positives": 485607,
"unique_terms": 550651,
"unique_anchor_positive_pairs": 4915749,
"ambiguous_terms_across_cuis": 199127,
"self_pairs_exact": 0,
"self_pairs_normalized": 40,
"cuis_with_multiple_terms": 477293,
"terms_per_cui_mean": 10.059801840798,
"terms_per_cui_median": 7,
"terms_per_cui_p95": 28.0,
"terms_per_cui_max": 37,
"triplets_per_cui_mean": 10.375220252549273,
"triplets_per_cui_median": 7,
"triplets_per_cui_p95": 33.0,
"triplets_per_cui_max": 35,
"anchor_words_mean": 4.16947488095767,
"positive_words_mean": 4.407634056405265,
"cuis_common_with_synonyms": 83914,
"cuis_added_vs_synonyms": 393379,
"cuis_missing_vs_synonyms": 0,
"terms_common_with_synonyms": 165661,
"terms_added_vs_synonyms": 384990,
"terms_missing_vs_synonyms": 0
}