System and method for generating substitutable queries

1. A method for generating one or more related queries with respect to a given query, the method comprising:

receiving a query comprising one or more terms;

generating a corpus of candidate reformulations based on the received query;

identifying one or more queries or terms from the corpus of candidate reformulations related to the received query;

generating with a processor one or more related queries on the basis of the identified one or more queries or terms from the corpus of candidate reformulations and segmenting the received query into one or more segments;

calculating a relevance factor by assigning a likelihood ratio score to each of the one or more related queries, the likelihood ratio indicating a degree of relatedness of the one or more related queries to the query, the likelihood ratio calculated by comparing a plurality of pair independence functions to each other, the pair independence functions being a function of a number of occurrences of the one or more related queries in the corpus of candidate reformulations; and

comparing the relevance factor for each of the one or more related queries to a threshold and selecting therefrom at least one related query that exceeds the threshold.

2. The method of claim 1 wherein segmenting the received query comprises segmenting the received query into a plurality of terms or phrases, generating one or more related queries comprising identifying at least one other query from the candidate reformulations with a same number of segments as that of the query, selecting at least one segment from the at least one other query with the same number of segments as that of the query that differs therefrom, the selected at least one segment identified as a candidate segment-level reformulation, and replacing at least one of the plurality of terms or phrases of the query with the candidate segment-level reformulation.

3. The method of claim 1 wherein generating the corpus of candidate reformulations comprises generating a pair of successive queries of one or more terms.

4. The method of claim 1 wherein generating the corpus of candidate reformulations comprises generating a pair of successive segments of one or more terms.

5. The method of claim 1 wherein identifying comprises identifying one or more queries or terms from the corpus of candidate reformulations related to the one or more segments of the received query.

6. The method of claim 1 wherein receiving a query comprises receiving from a user of a client device.

7. The method of claim 1 wherein generating a corpus of candidate reformulations comprises:

identifying one or more pairs of queries issued successively by one or more users of client devices; and

generating a corpus of candidate reformulations on the basis of the one or more identified pairs of queries.

8. The method of claim 1 wherein generating a corpus of candidate reformulations comprises:

segmenting one or more queries issued successively by one or more users of client devices;

identifying one or more differing segments between the successively issued queries; and

generating a corpus of candidate reformulations on the basis of the one or more segments.

9. The method of claim 1 , wherein generating a corpus of candidate formulations based on the received query comprise:

maintaining a query log including queries generated by one or more users and a time that the queries were generated; and

identifying as candidate reformulations from the query log one or more other queries received within a duration of time from the query.

10. The method of claim 9 , wherein the query is received from a user and the one or more queries identified comprise a least one successive query from the user within the duration of time.

11. The method of claim 1 , comprising identifying search results in response to the query using at least one of the selected queries.

12. The method of claim 1 wherein the relevance factor is calculated further based on a score generated by linear regression for each of the one or more related queries.

13. The method of claim 1 wherein the relevance factor is calculated further based on a query substitutable type of each of the one or more related queries.