Relevance

Relevance is the quality of being closely connected to or important for a particular matter. In everyday use, something is relevant if it helps address the issue at hand or supports a stated goal. Relevance is often relative, varying with the user’s needs, context, and information goals.

In information retrieval and search, relevance refers to how well a document or result satisfies a user’s information need. Relevance judgments are typically subjective and may be categorized as highly relevant, relevant, partially relevant, or not relevant. System performance is evaluated with metrics such as precision, recall, and F1, which compare retrieved results to a set of relevance judgments.

In logic and philosophy, relevance concerns whether a premise bears on the conclusion of an argument. Relevance logic studies systems where implication requires a substantive connection between premises and conclusion, rather than mere truth-functional relation.

In statistics and machine learning, relevance can describe the informativeness of a feature for predicting an outcome. Measures such as mutual information or correlation assess feature relevance, guiding feature selection and model simplification.

In communication theory, relevance theory, proposed by Sperber and Wilson, posits that readers and listeners seek the most relevant balance of cognitive effect and processing effort, shaping how information is interpreted and communicated.

In law and evidence, relevance refers to information that has probative value for a matter at issue, subject to rules aimed at limiting prejudice, confusion, or extraneous weight.

Relevance is a practical criterion across disciplines: it guides selection, attention, and inference, and it remains contingent on goals, context, and the available information.