Systems and methods for analyzing and clustering search queries

Abstract

Computerized systems and methods are disclosed for analyzing search query relationships and managing electronic content. In accordance with one implementation, log data pertaining to a plurality of queries may be received over an electronic network. A click graph may be generated representing one or more relationships between the queries. Further, temporal similarities may be identified between the queries, for example, by looking at peaks in frequency of queries over time. A pair of search queries may be evaluated based on the generated click graph and the identified temporal similarities to determine whether the queries in the pair are related.

1 - 39 . (canceled) 
     
     
         40 . A method for determining search query relationships, the method comprising the following operations performed by one or more processors:
 receiving, over an electronic network, log data relating to a plurality of search queries received from users;   generating a click graph representing relationships among a plurality of queries and a plurality of visited query results associated with each of the plurality of queries;   computing a numeric value quantifying a first layer relationship between a first query and a second query in the plurality of queries of the click graph, where a positive numeric value for the first layer relationship indicates that at least one of the plurality of visited query results is associated with both the first query and the second query;   determining a degree of temporal similarity between the first query and the second query based on a temporal distance between peaks in frequency of occurrence for the first query and the second query; and   designating the first query and the second query as related based on the computed numeric value being positive or the temporal distance being below a first threshold value.   
     
     
         41 . The method of  claim 40 , wherein the click graph represents at least one second layer relationship between the first query and the second query in the plurality of queries, the second layer relationship indicating that each of the first and second queries has a first layer relationship to at least one of the plurality of visited query results. 
     
     
         42 . The method of  claim 40 , wherein the click graph represents at least one third layer relationship between the first query and the second query in the plurality of queries, the third layer relationship indicating that each of the first and second queries has at least one of a first layer relationship and a second layer relationship to at least one of the plurality of visited query results. 
     
     
         43 . The method of  claim 40 , wherein the temporal distance is a Euclidean distance. 
     
     
         44 . The method of  claim 40 , wherein determining a degree of temporal similarity comprises computing a time series for each of the first and second queries, the time series indicating a query frequency at each of a plurality of times. 
     
     
         45 . The method of  claim 44 , further comprising performing at least one of normalizing, coarsening, and flattening on each computed time series. 
     
     
         46 . The method of  claim 44 , further comprising normalizing, coarsening, and flattening each computed time series. 
     
     
         47 . The method of  claim 41 , further comprising:
 computing a second numeric value quantifying the second layer relationship between the first query and the second query, where a positive numeric value for the second layer relationship indicates that each of the first and second queries has a first layer relationship to at least one of the plurality of visited query results; and   designating the first query and the second query as related based on the computed second numeric value being positive and the temporal distance being below a second threshold value greater than the first threshold value.   
     
     
         48 . The method of  claim 42 , further comprising:
 computing a third numeric value quantifying the third layer relationship between the first query and the second query, where a positive numeric value for the third layer relationship indicates that each of the first and second queries has at least one of a first layer relationship and a second layer relationship to at least one of the plurality of visited query results; and   designating the first query and the second query as related based on the computed third numeric value being positive and the temporal distance being below a third threshold value greater than the first threshold value.   
     
     
         49 . A system for determining search query relationships, the system comprising:
 one or more storage devices that store instructions; and   one or more processors that are configured by executing the instructions to:
 receive, over an electronic network, log data relating to a plurality of search queries received from users; 
 generate a click graph representing relationships among a plurality of queries and a plurality of visited query results associated with each of the plurality of queries, wherein the click graph represents at least one first layer relationship between a first query and a second query in the plurality of queries, the first layer relationship indicating that at least one of the plurality of visited query results is associated with both the first query and the second query; 
 compute a numeric value quantifying the first layer relationship between the first query and the second query, where a positive numeric value for the first layer relationship indicates that at least one of the plurality of visited query results is associated with both the first query and the second query; 
 determine a degree of temporal similarity between the first query and the second query based on a temporal distance between peaks in frequency of occurrence for the first query and the second query; and 
 designate the first query and the second query as related based on the computed numeric value being positive or the temporal distance being below a threshold value. 
   
     
     
         50 . The system of  claim 49 , wherein the click graph represents at least one second layer relationship between the first query and the second query in the plurality of queries, the second layer relationship indicating that each of the first and second queries has a first layer relationship to at least one of the plurality of visited query results. 
     
     
         51 . The system of  claim 49 , wherein the click graph represents at least one third layer relationship between the first query and the second query in the plurality of queries, the third layer relationship indicating that each of the first and second queries has at least one of a first layer relationship and a second layer relationship to at least one of the plurality of visited query results. 
     
     
         52 . The system of  claim 49 , wherein the temporal distance is a Euclidean distance. 
     
     
         53 . The system of  claim 49 , wherein to determine a degree of temporal similarity comprises computing a time series for each of the first and second queries, the time series indicating a query frequency at each of a plurality of times. 
     
     
         54 . The system of  claim 53 , wherein the one or more processors are further configured to perform at least one of normalizing, coarsening, and flattening on each computed time series. 
     
     
         55 . The system of  claim 53 , wherein the one or more processors are further configured to perform normalizing, coarsening, and flattening each computed time series. 
     
     
         56 . The system of  claim 50 , wherein the one or more processors are further configured to:
 compute a second numeric value quantifying the second layer relationship between the first query and the second query, where a positive numeric value for the second layer relationship indicates that each of the first and second queries has a first layer relationship to at least one of the plurality of visited query results; and   designate the first query and the second query as related based on the computed second numeric value being positive and the temporal distance being below a second threshold value greater than the first threshold value.   
     
     
         57 . The system of  claim 51 , wherein the one or more processors are further configured to:
 compute a third numeric value quantifying the third layer relationship between the first query and the second query, where a positive numeric value for the third layer relationship indicates that each of the first and second queries has at least one of a first layer relationship and a second layer relationship to at least one of the plurality of visited query results; and   designate the first query and the second query as related based on the computed third numeric value being positive and the temporal distance being below a third threshold value greater than the first threshold value.   
     
     
         58 . A computer-readable storage medium including instructions which, when executed by at least one processor, cause the at least one processor to perform the operations of:
 receiving, over an electronic network, log data relating to a plurality of search queries received from users;   generating a click graph representing relationships among a plurality of queries and a plurality of visited query results associated with each of the plurality of queries;   computing a numeric value quantifying a first layer relationship between a first query and a second query in the plurality of queries of the cock graph, where a positive numeric value for the first layer relationship indicates that at least one of the plurality of visited query results is associated with both the first query and the second query;   determining a degree of temporal similarity between the first query and the second query based on a temporal distance between peaks in frequency of occurrence for the first query and the second query; and   designating the first query and the second query as related based on the computed numeric value being positive or the temporal distance being below a threshold value.   
     
     
         59 . The computer-readable storage medium of  claim 58 , wherein the instructions further cause the at least one processor to perform the operations of:
 computing a second numeric value quantifying a second layer relationship between the first query and the second query, where a positive numeric value for the second layer relationship indicates that each of the first and second queries has a first layer relationship to at least one of the plurality of visited query results; and   designating the first query and the second query as related based on the computed second numeric value being positive and the temporal distance being below a second threshold value greater than the first threshold value.   
     
     
         60 . The computer-readable storage medium of  claim 58 , wherein the instructions further cause the at least one processor to perform the operations of:
 computing a third numeric value quantifying the third layer relationship between the first query and the second query, where a positive numeric value for the third layer relationship indicates that each of the first and second queries has at least one of a first layer relationship and a second layer relationship to at least one of the plurality of visited query results; and   designating the first query and the second query as related based on the computed third numeric value being positive and the temporal distance being below a third threshold value greater than the first threshold value.