README.md

metadata

license: mit
task_categories:
  - token-classification
language:
  - en
tags:
  - insurance
  - finance
size_categories:
  - n<1K

🧾 Payslips

Description
This repository contains the Payslips dataset, introduced in the paper:

_Training LayoutLM from Scratch for Efficient Named-Entity Recognition in the Insurance Domain_ (COLING 2025, FinNLP Workshop).

The dataset consists of payment slips (text + layout, no images) used in insurance for information verification. Insurers often request payslips to ensure that applicants’ declared income matches insurance premiums or coverage.

Content

• 611 annotated pages (485 train / 126 test)

• 9 entity types for extraction

• Entities cover salary-related and personal information relevant to the insurance domain

• Format: text + layout coordinates, no document images

Intended Use

• Named-Entity Recognition (NER)/ Key Information Extraction (KIE)
• Document Understanding in the insurance/financial domain

License This dataset is released under the MIT License. See LICENSE for details.

The license allows the commercial use, modification, distribution, and private use of the dataset, provided that the original copyright notice and this permission notice are included in all copies or substantial portions of the dataset.

The license does not provide any warranty, and the dataset is provided "as is".

The copyright notice and the permission notice shall be included in all copies or substantial portions of the dataset.