Named Entity Recognition in User-Generated Text: A Systematic Literature Review

Abstract

Named Entity Recognition (NER) in social media has received much research attention in the field of natural language processing (NLP) and information extraction. Research on this topic has grown dramatically in recent years. Hence, one of the objectives of this systematic literature review (SLR) is to present the outline techniques, approaches, and methods used to handle NER on X based on English datasets prepared for WNUT (Workshop on User-generated Text). This study could be used to develop more accurate models in the future. This SLR focuses on articles that had been published over the course of eight years, i.e., from July 2015 to the end of 2023. A total of 67 out of 316 articles published during the period were selected having met the set chosen criteria. Based on the analysis of the selected articles, challenges were identified and discussed. In this SLR, we aim to provide a better understanding of current viewpoints and highlight opportunities for research in NER in User-generated Text specifically for English usage on X. It can aid in identifying named entities, such as names, locations, companies, and groups, within a specific informal social media context like X. This research is notable for being the first systematic review that emphasizes the dearth of NER on X based on English datasets prepared for WNUT.The main contribution of this systematic review is a comprehensive study on NER in X messages for social media, entailing its challenges and opportunities. Moreover, new possible research directions are suggested for the researchers.

Keywords

• Named entity recognition
• NER
• user-generated text
• WNUT
• X
• systematic literature review