The way we cite: common metadata used across disciplines for defining bibliographic references (2205.13419v3)

Abstract: Current citation practices observed in articles are very noisy, confusing, and not standardised, making identifying the cited works problematic for hu-mans and any reference extraction software. In this work, we want to investigate such citation practices for referencing different types of entities and, in particular, to understand the most used metadata in bibliographic refer-ences. We identified 36 types of cited entities (the most cited ones were articles, books, and proceeding papers) within the 34,140 bibliographic references extracted from a vast set of journal articles on 27 different subject ar-eas. The analysis of such bibliographic references, grouped by the particular type of cited entities, enabled us to highlight the most used metadata for de-fining bibliographic references across the subject areas. However, we also noticed that, in some cases, bibliographic references did not provide the essential elements to identify the work they refer to easily.