Seasonality
Part of the Query Understanding series by Daniel Tunkelang.
Overview
The same query can carry very different intent depending on when it is submitted. Search systems that are aware of temporal context can interpret these variations correctly and surface more relevant results. Seasonal awareness operates at multiple timescales — time of year, day of week, and even time of day can all shift what the most relevant interpretation of a query is. Beyond improving relevance, seasonal signals are also useful for surfacing timely content and aligning results with promotional calendars. Detecting seasonal intent requires learning which queries have strong temporal associations, as opposed to queries whose meaning is stable year-round, and then acting on those associations without over-fitting to historical patterns that may not repeat exactly.
Note: Article content behind Medium paywall — accessible at source URL with Medium account.