Query expansion for web search
AssigneeMicrosoft Corporation
InventorsJun Xu, Hang Li
Life of this patent
Estimating the term…
Abstract
Systems, methods, and devices are described for retrieving query results based at least in part on a query and one or more similar queries. Upon receiving a query, one or more similar queries may be identified and/or calculated. In one embodiment, the similar queries may be determined based at least in part on click-through data corresponding to previously submitted queries. Information associated with the query and each of the similar queries may be retrieved, ranked, and or combined. The combined query results may then be re-ranked based at least in part on a responsiveness and/or relevance to the previously submitted query. The re-ranked query results may then be output to a user that submitted the original query.
Claims20
A method comprising:
- receiving a query that includes one or more terms, the query being associated with first resource identifiers, and the first resource identifiers being associated with first click-through data and ranking scores;
- identifying a similar query including a set of terms different from the one or more terms, the similar query being associated with second resource identifiers and the second resource identifiers being associated with second click-through data;
- determining a query similarity score between the query and the similar query, the query similarity score being based at least in part on a correlation between the first click-through data and the second click-through data;
- identifying a resource identifier of the second resource identifiers that is not included in the first resource identifiers;
- determining a ranking score between the resource identifier and the similar query;
- determining a combined ranking score associated with the resource identifier based on the ranking score and the query similarity score;
- and
- combining the resource identifier with the first resource identifiers to generate a set of query results, the set of query results being ranked based at least in part on the combined ranking score associated with the resource identifier and the ranking scores associated with the first resource identifiers.
A computer-readable storage medium including computer-executable instructions that, when executed by a processor, perform operations comprising:
- receiving a query requesting information relevant to the query;
- identifying one or more queries similar to the query when the one or more similar queries are stored in a repository;
- calculating the one or more similar queries, wherein the calculating comprises determining a Pearson correlation based on click-through information of resource identifiers of the query and the one or more queries to determine a number of shared resource identifiers that corresponds to both the query and the one or more queries when the one or more similar queries are not stored in the repository;
- and
- providing a combined list of ranked information corresponding to the query and to the one or more similar queries.
A system comprising:
- a processor;
- a memory comprising components that are executed by the processor, the components comprising:
- a similar query finder to receive a query and retrieve one or more queries similar to the query;
- a query similarity calculator to calculate the one or more similar queries based at least in part on a Pearson correlation based on click-through data, wherein the one or more queries are determined to be similar to the query when the Pearson correlation is above a threshold;
- a retrieval interface to:
- retrieve information responsive to the query and to the one or more similar queries;
- rank the information into a set of search results;
- and
- re-rank the set of search results based at least in part on the query;
- and
- a re-ranker configured to re-rank the set of search results into a set of re-ranked search results based at least in part on a relevance to the query.
Dependencies are inferred from the claim text.