Ranking functions using document usage statistics

Abstract

Methods of providing a document relevance score to a document on a network are disclosed. Computer readable medium having stored thereon computer-executable instructions for performing a method of providing a document relevance score to a document on a network are also disclosed. Further, computing systems containing at least one application module, wherein the at least one application module comprises application code for performing methods of providing a document relevance score to a document on a network are disclosed.

1 . A method of determining a document relevance score for a document on a network, said method comprising the steps of:
 assigning an actual usage value (U A ) to one or more documents on a network comprising N documents, wherein the actual usage value (U A ) is based on actual usage data maintained and stored on a server;   if less than N documents are assigned an actual usage value (U A ), assigning a default usage value (U D ) to the documents that do not have actual usage data associated therewith; and   using the usage value for each document to determine the document relevance score of a given document on the network.   
     
     
         2 . The method of  claim 1 , further comprising the step of:
 retrieving actual usage data or an actual usage value (U A ) for a document from a data storage file on the server.   
     
     
         3 . The method of  claim 2 , further comprising the step of:
 storing actual usage data or an actual usage value (U A ) for a document in a data storage file.   
     
     
         4 . The method of  claim 2 , wherein the document relevance score for each document on the network is generated using a formula: 
       
         
           
             
               Score 
               = 
               
                 
                   ∑ 
                   
                     
                       
                         
                           wtf 
                           ′ 
                         
                          
                         
                           ( 
                           
                             
                               k 
                               1 
                             
                             + 
                             1 
                           
                           ) 
                         
                       
                       
                         
                           k 
                           1 
                         
                         + 
                         
                           wtf 
                           ′ 
                         
                       
                     
                     × 
                     
                       log 
                        
                       
                         ( 
                         
                           N 
                           n 
                         
                         ) 
                       
                     
                   
                 
                 + 
                 
                   
                     w 
                     cd 
                   
                    
                   
                     
                       k 
                       cd 
                     
                     
                       
                         k 
                         cd 
                       
                       + 
                       
                         
                           
                             
                               b 
                               cd 
                             
                              
                             
                               CD 
                               
                                 k 
                                 ew 
                               
                             
                           
                           + 
                           
                             
                               b 
                               ud 
                             
                              
                             UD 
                           
                         
                         
                           
                             b 
                             cd 
                           
                           + 
                           
                             b 
                             ud 
                           
                         
                       
                     
                   
                 
                 + 
                 
                   
                     w 
                     u 
                   
                    
                   
                     
                       
                         k 
                         u 
                       
                        
                       U 
                     
                     
                       
                         k 
                         u 
                       
                       + 
                       U 
                     
                   
                 
               
             
           
         
         wherein:
 wtf′ represents a weighted term frequency, 
 N represents a number of documents on the network, 
 n represents a number of documents containing a query term, 
 w cd  represents a weight of a query-independent component, 
 b cd  represents a weight of a click distance, 
 b ud  represents a weight of a URL depth, 
 CD represents a computed click distance or assigned biased click distance for a document, 
 k ew  represents a tuning constant related to edge weights, 
 UD represents a URL depth,
 U represents an actual usage value or a default usage value, 
 
 w u  and k u  represent tuning parameters for the usage value, and 
 k cd  and k 1  are constants. 
 
       
     
     
         5 . A method of ranking documents on a network, said method comprising the steps of:
 determining a document relevance score for each document on the network using the method of claim  12 ; and   ranking the documents in descending order based on the document relevance scores of each document.   
     
     
         6 . A method of ranking search results of a search query, said method comprising the steps of:
 determining a document relevance score for each document in the search results of a search query using the method of claim  12 ; and   ranking the documents in descending order based on the document relevance scores of each document.   
     
     
         7 . A computing system containing at least one application module usable on the computing system, wherein the at least one application module comprises application code for performing a method of determining a document relevance score for a document on a network, said method comprising the steps of:
 assigning an actual usage value (U A ) to one or more documents on a network comprising N documents, wherein the actual usage value (U A ) is based on actual usage data maintained and stored on a server;   if less than N documents are assigned an actual usage value (U A ), assigning a default usage value (U D ) to the documents that do not have actual usage data associated therewith; and   using the usage value for each document to determine the document relevance score of a given document on the network.   
     
     
         8 . The computing system of  claim 7 , wherein the actual usage value is dependent on one or more usage-related properties of a document or a folder containing a set of documents, said one or more usage-related properties comprising a total number of document or folder views by users within a given period of time, an average number of document or folder views per user within a given period of time, a total time spent on a particular document or folder within a given period of time, an average time spent on a particular document or folder within a given period of time, wherein the given period of time comprises last week, last month, last year, a lifetime of the document or folder, or any other period of time. 
     
     
         9 . A computer readable medium having stored thereon computer-executable instructions for performing a process for ranking documents on a network, said process comprising:
 for each document in a plurality of the documents, determining if there is an actual-usage metric for the document, the actual-usage having been generated by user interactions with the document and representing a degree of actual user interaction with the document, and if there is not an actual-usage metric for the document, assigning to the document a default actual-usage metric value that is based on the actual-usage metrics of the documents generated by user interactions with the documents;   storing the actual-usage metrics of the documents for use in ranking documents identified as satisfying a query of the documents;   receiving a query entered by a user, the query including a query string;   searching the documents to identify documents that match the query string;   ranking, relative to each other, the identified documents that match the query string based on the stored respective actual-usage metrics of the identified documents; and   returning to the user indicia of the identified documents and their relative rankings based on the stored actual-usage metrics.