Hugging Face's logo

Datasets:
Mykes
/
patient_queries_ner_SDDCS

Modalities:
Text
Formats:
parquet
Size:
1K - 10K
Libraries:
Datasets
pandas
Dataset card Data Studio Files Community
2
patient_queries_ner_SDDCS / README.md
Mykes's picture
Mykes
Update README.md
98867cd verified
preview code
|
raw
history blame contribute delete
3.76 kB
metadata
dataset_info:
  features:
    - name: sentence
      dtype: string
    - name: word_labels
      sequence: string
  splits:
    - name: train
      num_bytes: 3906887
      num_examples: 9016
  download_size: 1113278
  dataset_size: 3906887
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

image/jpeg

NER Dataset for Symptoms, Diseases, Drugs, Cities, Subway Stations, Gender, and Age

Dataset Overview

This dataset contains Named Entity Recognition (NER) annotations for Russian text, specifically focusing on extracting entities related to medical and geographical contexts. The dataset is annotated with seven entity types:

  • SYMPTOMS: Physical or mental characteristics indicating disease or condition.
  • DISEASES: Diagnosed medical conditions.
  • DRUGS: Medications or substances used for treatment.
  • CITIES: Names of urban areas/geographical locations.
  • SUBWAY STATIONS: Names of metro stations.
  • GENDER: Gender of individuals (e.g., мужчина, женщина).
  • AGE: Age of individuals, typically expressed numerically.

The dataset consists of 9,020 sentences. Each word in a sentence is assigned a label that corresponds to one of the entity types or O (non-entity token). This dataset is useful for training and evaluating Named Entity Recognition models in the medical and geographical domains.

Dataset Example

Here is an example sentence from the dataset, along with the corresponding entity labels:

Sentence: 

паразитолог в москве рядом с авиамоторная для ребенка с лямблиоз, изменения аппетита и обложенности языка

Word Labels: 

['B-SPECIALITY', 'O', 'B-CITY', 'O', 'O', 'B-SUBWAY', 'O', 'B-AGE', 'O', 'B-DISEASE', 'O', 'B-SYMPTOM', 'I-SYMPTOM', 'O', 'B-SYMPTOM', 'I-SYMPTOM']

In this example:

  • "паразитолог" (speciality) is tagged as B-SPECIALITY
  • "москве" (city) is tagged as B-CITY
  • "авиамоторная" (subway) is tagged as B-SUBWAY
  • "ребенка" (child) is tagged as B-AGE
  • "лямблиоз" (disease) is tagged as B-DISEASE
  • изменения аппетита (symptom) is tagged as B-SYMPTOM and I-SYMPTOM
  • обложенности языка (symptom) is tagged as B-SYMPTOM and I-SYMPTOM

Dataset Structure

The dataset contains two columns:

  • sentence: A string representing the input sentence.
  • word_labels: A list of labels assigned to each word in the sentence, using BIO (Beginning, Inside, Outside) tagging format.

Entity Types

Entity Type Description

  • SYMPTOMS Symptoms or signs of conditions (e.g., "тяжесть в желудке", "снижения трудоспособности").
  • DISEASES Diagnosed medical conditions (e.g., "легкое когнитивное расстройство").
  • DRUGS Names of medications or treatments (e.g. "цефепим").
  • CITIES Urban areas or geographical locations ("Москва", "Тула").
  • SUBWAY STATIONS Names of subway/metro stations ("Шоссе Энтузиастов", "Проспект Мира").
  • GENDER Gender references (e.g., "мужчина", "женщина").
  • AGE a word potentially indicative of age (e.g. "пожилой", "ребенок").

Labeling Format

The dataset uses the BIO labeling format:

  • B- indicates the beginning of an entity.
  • I- indicates the inside of an entity.
  • O indicates a token that is outside of any entity.

For example:

  • B-SYMPTOM marks the first token of a symptom entity.
  • I-SYMPTOM marks subsequent tokens of the same symptom entity.
  • O marks a token that does not belong to any entity.