Query Distance is a measure of similarity used in algorithms designed to categorize and cluster search queries based on their characteristics. In general, this metric helps quantify how closely related two queries are.
The concept is used within clustering algorithms, where historical queries are analyzed based on various features such as the broad nature of the query (informational/navigational), query length, mean document staytime, and query distance. Query distance may be measured using string-based heuristics (like fuzzy matching) or through comparing the semantic proximity of query embeddings.
By analyzing query distance, search systems can group related user queries into clusters, helping to identify query categories and refine relevance scoring, particularly when used in conjunction with implicit feedback metrics like historical click fractions.