CardioBERTa.en_enriched / training_data_stats.json
fernandogd's picture
Release CardioBERTa-based clinical concept disambiguation model
4508ed4 verified
Raw History Blame Contribute Delete
928 Bytes
{
"triplets": 83914,
"malformed_rows": 0,
"unique_triplets": 83914,
"duplicate_triplets": 0,
"unique_cuis": 83914,
"unique_anchors": 83292,
"unique_positives": 83472,
"unique_terms": 165661,
"unique_anchor_positive_pairs": 83798,
"ambiguous_terms_across_cuis": 2089,
"self_pairs_exact": 0,
"self_pairs_normalized": 36,
"cuis_with_multiple_terms": 83878,
"terms_per_cui_mean": 1.9995709893462355,
"terms_per_cui_median": 2.0,
"terms_per_cui_p95": 2.0,
"terms_per_cui_max": 2,
"triplets_per_cui_mean": 1,
"triplets_per_cui_median": 1.0,
"triplets_per_cui_p95": 1.0,
"triplets_per_cui_max": 1,
"anchor_words_mean": 5.238303501203613,
"positive_words_mean": 5.761493910432109,
"cuis_common_with_synonyms": 83914,
"cuis_added_vs_synonyms": 0,
"cuis_missing_vs_synonyms": 0,
"terms_common_with_synonyms": 165661,
"terms_added_vs_synonyms": 0,
"terms_missing_vs_synonyms": 0
}