Identifying and expanding implicitly temporally qualified queries

• US 8,156,111 B2
• Filed: 11/24/2008
• Issued: 04/10/2012
• Est. Priority Date: 11/24/2008
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Jones, Rosie, Metzler, Donald, Peng, Fuchun
• Primary Examiner(s): Perveen, Rehana
• Assistant Examiner(s): Bui, Tiffany Thuy
• Application Number:
  US12/277,059
• Publication Number:
  US 20100131538A1
• Time in Patent Office: 1,233 Days
• Field of Search: 707/721, 707/723, 707/725
• US Class Current: 707/721
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Methods and apparatus are described for identifying implicitly temporally qualified queries, i.e., queries for which a time period is implied but not explicitly stated, and for expanding such queries to include one or more temporal references.

Claims

1. A computer-implemented method for responding to a first search query that does not include an explicit temporal qualification, comprising:

receiving the first search query from a user device;

determining that the first search query is implicitly temporally qualified with reference to a first query term included in the first search query and query log data that relate a plurality of previous query terms included in previous search queries by a plurality of users to explicit temporal references included in the previous search queries, wherein the query log data represent frequencies with which the previous query terms were associated with particular ones of the explicit temporal references in the previous search queries, and wherein the previous query terms are included in the query log data only where the frequency with which each of the previous query terms was associated with at least one of the explicit temporal references in the previous search queries exceeds a programmable threshold, and wherein the first query term is associated with a first plurality of the explicit temporal references in the query log data;

expanding the first search query to include one or more of the explicit temporal references;

generating first search results using the expanded first search query, wherein the first search results are biased in accordance with the frequency with which each of the first plurality of explicit temporal references is associated with the first query term in the query log data; and

transmitting the first search results to the user device.

2. The method of claim 1 further comprising including each of the previous query terms in the query log data only where the previous queries including the previous query terms had been reformulated from original queries including the previous query terms but not explicitly including any of the explicit temporal references.

3. The method of claim 1 wherein expanding the first search query comprises expanding the first search query to include each of the first plurality of explicit temporal references.

4. The method of claim 1 wherein the first search results comprise one or more of organic search results or sponsored search results.

5. A computer-implemented method for presenting search results on a user device in response to a first search query that does not include an explicit temporal qualification, comprising:

providing a search interface on the user device in which a user enters the first search query; and

providing a search results interface on the user device in which search results are presented that are responsive to a first expanded version of the first search query, the first expanded version of the first search query including one or more of a plurality of explicit temporal references not included in the first search query based upon query log data that relate a plurality of previous query terms included in previous search queries by a plurality of users to the explicit temporal references included in the previous search queries, wherein the query log data represent frequencies with which the previous query terms were associated with particular ones of the explicit temporal references in the previous search queries, and wherein the previous query terms are included in the query log data only where the frequency with which each of the previous query terms was associated with at least one of the explicit temporal references in the previous search queries exceeds a programmable threshold, and wherein the first query term is associated with a first plurality of the explicit temporal references in the query log data, the search results being biased in accordance with the frequency with which each of the first plurality of explicit temporal references is associated with the first query term in the query log data.

6. The method of claim 5 further comprising providing an indication in the search results interface that the search results are responsive to the first expanded version of the first search query rather than the first search query.

7. The method of claim 6 further comprising providing a mechanism in the search results interface by which the user requests further search results responsive to the first search query.

8. The method of claim 6 further comprising providing a mechanism in the search results interface by which the user requests further search results responsive to a second expanded version of the first search query including a different temporal reference than the first expanded version of the first search query.

9. The method of claim 5 wherein the search results comprise one or more of organic search results or sponsored search results.

10. A computer-implemented method for responding to a first search query that does not include an explicit temporal qualification, comprising:

receiving the first search query from a user device;

determining that the first search query is implicitly temporally qualified with reference to a first query term included in the first search query and query log data that relate a plurality of previous query terms included in previous search queries by a plurality of users to explicit temporal references included in the previous search queries, wherein the query log data represent frequencies with which the previous query terms were associated with particular ones of the explicit temporal references in the previous search queries, and wherein the previous query terms are included in the query log data only where the frequency with which each of the previous query terms was associated with at least one of the explicit temporal references in the previous search queries exceeds a programmable threshold, and wherein the first query term is associated with a first plurality of the explicit temporal references in the query log data;

generating first search results responsive to the first search query, the first search results being ranked in accordance with a predetermined ranking function derived with reference to the query log data, and wherein the first search results are biased in accordance with the frequency with which each of the first plurality of explicit temporal references is associated with the first query term in the query log data; and

transmitting the first search results to the user device.

11. The method of claim 10 further comprising including each of the previous query terms in the query log data only where the previous queries including the previous query terms had been reformulated from original queries including the previous query terms but not explicitly including any of the specific temporal references.

12. The method of claim 10 wherein the first search results comprise one or more of organic search results or sponsored search results.

13. The method of claim 10 further comprising deriving the predetermined ranking function by training the predetermined ranking function on a temporally qualified query set represented by the query log data.

14. A system for responding to a first search query that does not include an explicit temporal qualification, the system comprising:

at least one data store having query log data stored therein, the query log data relating a plurality of previous query terms included in previous search queries by a plurality of users to explicit temporal references included in the previous search queries, wherein the query log data represent frequencies with which the previous query terms were associated with particular ones of the explicit temporal references in the previous search queries, and wherein the previous query terms are included in the query log data only where the frequency with which each of the previous query terms was associated with at least one of the explicit temporal references in the previous search queries exceeds a programmable threshold, and wherein a first query term included in the first search query is associated with a first plurality of the explicit temporal references in the query log data; and

at least one computing device configured to;

receive the first search query from a user device;

determine that the first search query is implicitly temporally qualified with reference to the first query term included in the first search query and the query log data;

generate first search results responsive to the first search query and taking the query log data into account, wherein the first search results are biased in accordance with the frequency with which each of the first plurality of explicit temporal references is associated with the first query term in the query log data; and

transmit the first search results to the user device.