README.md

metadata

dataset_info:
  features:
    - name: text
      dtype: string
    - name: label
      dtype: int64
  splits:
    - name: train
      num_bytes: 580321
      num_examples: 10000
  download_size: 342886
  dataset_size: 580321
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: mit
task_categories:
  - text-classification
language:
  - uz
pretty_name: Uzbek Sentiment Analysis Benchmark
size_categories:
  - 1K<n<10K

Original Dataset: https://huggingface.co/datasets/risqaliyevds/uzbek-sentiment-analysis

Binary sentiment analysis dataset for benchmarking purposes. It was annotated using GPT-4o API. Label 1 is for Positive, 0 for Negative.