Search

Search is the act of locating information within a collection of data by expressing a query. It applies to web content, databases, document stores, and local files, and typically returns a ranked list of items that meet the user’s information need.

Web search systems commonly combine crawling, indexing, and ranking. A crawler gathers documents by following links, an index stores representations of content in a structure such as an inverted index, and a query processor matches user requests to the index and computes relevance scores. The results are presented to the user in an ordered list, often with snippets, metadata, and filters.

Search algorithms are based on different retrieval models. Boolean retrieval uses explicit conjunctions and disjunctions; vector space models rank documents by measuring similarity between query vectors and document vectors using metrics like TF-IDF; probabilistic models estimate the probability that a document is relevant. Modern systems often blend these approaches with machine-learned ranking, incorporating signals such as content quality, freshness, link structure, and user interactions. Inverted indexes and efficient scoring are foundational for handling large corpora.

Different domains feature specialized search. Database search relies on structured queries (SQL) and full-text search with dedicated indexes; enterprise and desktop search index files, emails, and metadata. Search quality is evaluated using metrics such as precision, recall, mean reciprocal rank, and NDCG.

Historically, search progressed from simple keyword matching to scalable, AI-assisted systems. Its development has significantly influenced information access, commerce, and decision making, while raising considerations around privacy, bias, transparency, and spam.