README.md

metadata

dataset_info:
  features:
    - name: Text
      dtype: string
    - name: Label
      dtype: string
    - name: Encoding
      dtype: int64
    - name: Classification Score
      list:
        - name: label
          dtype: string
        - name: score
          dtype: float64
  splits:
    - name: train
      num_bytes: 3916705
      num_examples: 24746
    - name: validation
      num_bytes: 1295577
      num_examples: 8249
    - name: test
      num_bytes: 1299961
      num_examples: 8249
  download_size: 2594115
  dataset_size: 6512243
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*

Label is used to give a context to the related text using the following map :

• 0 --> "PATIENT"
• 1 --> "DOCTOR"
• 2 --> "NEUTRAL"

The "Classification Score" field reports the predicted outcomes of the patient-doctor-classifier.