ESCO / README.md
Daniel Duckworth
more strings
3a89c38
|
Raw
History Blame Contribute Delete
300 kB
metadata
license: cc
task_categories:
  - text-classification
  - zero-shot-classification
  - feature-extraction
  - text2text-generation
  - sentence-similarity
task_ids:
  - multi-label-classification
  - multi-class-classification
annotations_creators:
  - expert-generated
language_creators:
  - expert-generated
language:
  - en
multilinguality:
  - monolingual
source_datasets:
  - original
pretty_name: Multilingual classification of European Skills, Competences, and Occupations
size_categories:
  - n<1K
dataset_info:
  - config_name: esco_en
    features:
      - name: ISCO_CODE_1
        dtype:
          class_label:
            names:
              '0': '0'
              '1': '1'
              '2': '2'
              '3': '3'
              '4': '4'
              '5': '5'
              '6': '6'
              '7': '7'
              '8': '8'
              '9': '9'
      - name: ISCO_LABEL_1
        dtype: string
      - name: ISCO_DEFINITION_1
        dtype: string
      - name: ISCO_CODE_2
        dtype: string
      - name: ISCO_LABEL_2
        dtype: string
      - name: ISCO_DEFINITION_2
        dtype: string
      - name: ISCO_CODE_3
        dtype: string
      - name: ISCO_LABEL_3
        dtype: string
      - name: ISCO_CODE_4
        dtype: string
      - name: ISCO_LABEL_4
        dtype: string
      - name: ISCO_DEFINITION_4
        dtype: string
      - name: ISCO_CODES
        sequence: string
      - name: ISCO_LABELS
        sequence: string
      - name: ESCO_CODE
        dtype: string
      - name: ESCO_LABELS
        sequence: string
      - name: ESCO_OCCUPATION
        dtype: string
      - name: ESCO_DESCRIPTION
        dtype: string
      - name: LANGUAGE
        dtype: string
    splits:
      - name: train
        num_examples: 2413
      - name: test
        num_examples: 604
configs:
  - config_name: esco_en
    data_files:
      - split: train
        path: data/en/train.parquet
      - split: test
        path: data/en/test.parquet
    default: true
train-eval-index:
  - config: esco_en
    task: text-classification
    task_id: multi_class_classification
    splits:
      train_split: train
      eval_split: test
    col_mapping:
      ESCO_DESCRIPTION: text
      ISCO_CODE_4: target
    metrics:
      - type: accuracy
        name: Accuracy
      - type: f1
        name: F1 macro
        args:
          average: macro
      - type: f1
        name: F1 micro
        args:
          average: micro
      - type: f1
        name: F1 weighted
        args:
          average: weighted
      - type: precision
        name: Precision macro
        args:
          average: macro
      - type: precision
        name: Precision micro
        args:
          average: micro
      - type: precision
        name: Precision weighted
        args:
          average: weighted
      - type: recall
        name: Recall macro
        args:
          average: macro
      - type: recall
        name: Recall micro
        args:
          average: micro
      - type: recall
        name: Recall weighted
        args:
          average: weighted
      - type: danieldux/isco_hierarchical_accuracy
        name: ISCO Hierarchical Accuracy