CardioBERTa.cs_GP_enriched / training_data_stats.json
fernandogd's picture
Release CardioBERTa-based clinical concept disambiguation model
3e9dfa0 verified
Raw History Blame Contribute Delete
970 Bytes
{
"triplets": 4689093,
"malformed_rows": 0,
"unique_triplets": 4685695,
"duplicate_triplets": 3398,
"unique_cuis": 476969,
"unique_anchors": 465737,
"unique_positives": 446760,
"unique_terms": 526548,
"unique_anchor_positive_pairs": 4618651,
"ambiguous_terms_across_cuis": 191608,
"self_pairs_exact": 0,
"self_pairs_normalized": 185,
"cuis_with_multiple_terms": 476969,
"terms_per_cui_mean": 9.776270575236547,
"terms_per_cui_median": 7,
"terms_per_cui_p95": 27.0,
"terms_per_cui_max": 36,
"triplets_per_cui_mean": 9.83102256121467,
"triplets_per_cui_median": 7,
"triplets_per_cui_p95": 31.0,
"triplets_per_cui_max": 35,
"anchor_words_mean": 3.8381947212392675,
"positive_words_mean": 3.9683945274704513,
"cuis_common_with_synonyms": 68973,
"cuis_added_vs_synonyms": 407996,
"cuis_missing_vs_synonyms": 0,
"terms_common_with_synonyms": 135148,
"terms_added_vs_synonyms": 391400,
"terms_missing_vs_synonyms": 0
}