Irelated

Irelated is a term used in information science and data analysis to denote a metric for semantic relatedness between information entities within a corpus, knowledge graph, or retrieval system. The concept is intended to capture how much two items share information content and contextual meaning beyond simple co-occurrence, supporting tasks such as ranking, clustering, and recommendation.

Origin and usage of the term are informal. Irelated has appeared in discussions and experimental work since the early 2020s, often as a label for information-content–based similarity ideas. There is no single standardized definition or formula, and implementations vary by domain and dataset. In practice, Irelated is frequently described as a context-sensitive, information-content–driven measure that can be adapted to specific taxonomies, ontologies, or corpora.

Definition and interpretation. An Irelated score typically ranges from near 0 to 1, where higher values indicate greater shared information content or conceptual proximity. The measure is commonly conceptually related to established information-content approaches, but tailored to emphasize information overlap rather than lexical similarity alone. Depending on the method, the score may be computed using the information content of shared concepts (for example, through a lowest common ancestor in a taxonomy), or through a combination of shared features, co-occurrence signals, and hierarchical distance.

Applications and variants. Irelated is used in knowledge graphs for link ranking, in natural language processing for sense disambiguation, and in recommender systems for context-aware suggestions. Variants may be intrinsic (dependent only on the item's content) or contextual (dependent on user, task, or domain context).

See also semantic similarity, information content, knowledge graphs, and relatedness measures.