CardioBERTa.nl_GP_enriched / training_data_stats.json
fernandogd's picture
Release CardioBERTa-based clinical concept disambiguation model
18a8881 verified
Raw History Blame Contribute Delete
975 Bytes
{
"triplets": 4898584,
"malformed_rows": 0,
"unique_triplets": 4872237,
"duplicate_triplets": 26347,
"unique_cuis": 476972,
"unique_anchors": 481326,
"unique_positives": 463560,
"unique_terms": 543296,
"unique_anchor_positive_pairs": 4839774,
"ambiguous_terms_across_cuis": 197704,
"self_pairs_exact": 0,
"self_pairs_normalized": 1770,
"cuis_with_multiple_terms": 476971,
"terms_per_cui_mean": 9.938308747683303,
"terms_per_cui_median": 7.0,
"terms_per_cui_p95": 28.0,
"terms_per_cui_max": 37,
"triplets_per_cui_mean": 10.270170995362411,
"triplets_per_cui_median": 7.0,
"triplets_per_cui_p95": 33.0,
"triplets_per_cui_max": 35,
"anchor_words_mean": 4.16920706065263,
"positive_words_mean": 4.394558509152849,
"cuis_common_with_synonyms": 79985,
"cuis_added_vs_synonyms": 396987,
"cuis_missing_vs_synonyms": 0,
"terms_common_with_synonyms": 156019,
"terms_added_vs_synonyms": 387277,
"terms_missing_vs_synonyms": 0
}