ywhwang's picture
Update README.md
a50b644 verified
metadata
dataset_info:
  features:
    - name: sentences
      sequence: string
    - name: labels
      sequence: string
  splits:
    - name: test
      num_bytes: 1305703
      num_examples: 14
  download_size: 397229
  dataset_size: 1305703
configs:
  - config_name: default
    data_files:
      - split: test
        path: data/test-*
language:
  - ko
tags:
  - finance
pretty_name: MinDS-Clustering-ko
size_categories:
  - n<1K

MinDS-Clustering-ko

  • Utilized user questions and intents from the 'Finance/Insurance' category of AIHub's complaint Q&A dataset.
  • Constructed a dataset for clustering user intents based on their queries.
  • The intent labels are structured as follows:
    • Transaction inquiry and management
    • Account and card-related services
    • Financial transactions and incidents
    • Loan-related inquiries
    • Insurance and financial product inquiries
    • Insurance claims and compensation
    • Loss and theft reports
    • Product subscription-related inquiries
    • Product information and inquiries
    • Product cancellation and modification
    • Transfer and remittance services
    • Authentication and financial transaction services
    • Automobile and accident handling
    • Balance and account balance inquiries
    • Certificate and financial document issuance
    • Withdrawal and cash services
  • Applied a similar label structure for comparison with FinanceMTEB/MInDS-14-en dataset.