Dutch Named Entity Recognition and De-identification Methods for the Human Resource Domain

TL;DR

This paper evaluates Dutch text de-identification methods in the HR domain, updating NER models with BERTje, and introduces a new dataset for recognizing job titles to improve privacy protection.

Contribution

It updates existing de-identification methods with state-of-the-art NER models and introduces a new dataset for recognizing job titles in Dutch HR texts.

Findings

BERTje-based NER with CoNLL 2002 corpus achieves high recall for persons and locations.

DEDUCE performs best for gender de-identification.

New dataset effectively recognizes Dutch job titles in texts.

Abstract

The human resource (HR) domain contains various types of privacy-sensitive textual data, such as e-mail correspondence and performance appraisal. Doing research on these documents brings several challenges, one of them anonymisation. In this paper, we evaluate the current Dutch text de-identification methods for the HR domain in four steps. First, by updating one of these methods with the latest named entity recognition (NER) models. The result is that the NER model based on the CoNLL 2002 corpus in combination with the BERTje transformer give the best combination for suppressing persons (recall 0.94) and locations (recall 0.82). For suppressing gender, DEDUCE is performing best (recall 0.53). Second NER evaluation is based on both strict de-identification of entities (a person must be suppressed as a person) and third evaluation on a loose sense of de-identification (no matter what how a person is suppressed, as long it is suppressed). In the fourth and last step a new kind of NER dataset is tested for recognising job titles in texts.