NERSkill.Id: Annotated dataset of Indonesian's skill entity recognition

Abstract

NERSkill.Id is a manually annotated named entity recognition (NER) dataset focused on skill entities in the Indonesian language. The dataset comprises 418.868 tokens, each accompanied by corresponding tags following the BIO scheme. Notably, 15,51% of these tokens represent named entities, falling into three distinct categories: hard skill, soft skill, and technology. To construct this dataset, data were gathered from a job portal and subsequently processed using open-source libraries. Given the scarcity of annotated corpora for Indonesian, NERSkill.Id fills a significant void and offers immense value to multiple stakeholders. NLP researchers can harness the dataset's richness to advance skill entity recognition technology in the Indonesian language. Companies and recruiters can benefit by employing NERSkill.Id to enhance talent acquisition and job matching processes through accurate skill identification. Furthermore, educational institutions can leverage the dataset to adapt their courses and training programs to meet the evolving needs of the job market. This dataset can be effectively utilized for training and evaluating named entity recognition systems, empowering advancements in skill entity recognition for the Indonesian language.

Keywords

• Natural language processing
• Named entity recognition
• Text mining
• Skill entity recognition
• Indonesian skill entity