EN: Decision-support signal for suggesting candidate ICD-10 subgroups from Russian clinical notes. Not a substitute for clinician judgment; not validated for autonomous diagnosis.
RU: Вспомогательный сигнал для предложения кандидатных 3-символьных кодов МКБ-10 по русскому клиническому тексту. Не заменяет врача и не предназначен для автономных клинических решений.

Training data / Обучающие данные

Source CSV: datasets/subgroups/group_P.csv
SHA-256: eaa1a9f6e52dba1c8167c8a5c40d1d455bc2e6072842e6cb6bc3c1c09fd67d4c
Produced by ml/build_subgroup_datasets.ipynb (iterative multi-label stratification by parse_id).
Splits: train=159 · val=39 · test=38
Labels: 25 (ordered, includes P_OTHER for rare codes collapsed during dataset build).

Metrics (test split)

metric	value
macro_f1	0.7660
micro_f1	0.7414
weighted_f1	0.7773
subset_accuracy	0.6053
hit@1	0.9211
hit@3	0.9474
recall@3	0.9474
mrr	0.9430

Full per-label breakdown in metrics.json.

Limitations / Ограничения

Russian only; heavy reliance on clinical abbreviations (АД, ТТГ, УЗИ, etc.).
Training text had PII redacted (*ДАТА*, *ГОРОД*, ...); model may behave differently on non-redacted input.
Small chapters (train rows < 250) were trained with heavy regularization; some labels may have low support.
Rare labels without positives in train are kept in the label map (see label_map.json → rare_label_ids) for interface stability but will effectively never fire.

Inference

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

repo = "Dmitry43243242/icd10-ru-subgroup-p"
tok = AutoTokenizer.from_pretrained(repo)
mdl = AutoModelForSequenceClassification.from_pretrained(repo)
mdl.eval()

text = "жалобы пациента..."
inp = tok(text, return_tensors="pt", truncation=True, max_length=512)
with torch.no_grad():
    probs = torch.sigmoid(mdl(**inp).logits)[0]
preds = [mdl.config.id2label[i] for i, p in enumerate(probs.tolist()) if p >= 0.5]
top3 = sorted(
    [(mdl.config.id2label[i], p) for i, p in enumerate(probs.tolist())],
    key=lambda x: -x[1],
)[:3]
print(preds, top3)

Citation / Ссылка

Built as part of the ai-app ICD-10 classification pipeline. Upstream model: ai-forever/ruBert-base (ai-forever).