System & Method For Recommending Content Sources

Abstract

A networked computer system identifies, optimizes and recommends content sources for users. The content sources can be used for providing news feeds, search results, etc. based on taking into net useful content contributed by such sources over other sources.

What is claimed is: 
     
         1 .- 18 . (canceled) 
     
     
         19 . A method of recommending entities to a user for inclusion as content contributors for a user datastream using a networked computing system comprising:
 a. automatically identifying with the networked computing system an information contribution provided within a first user datastream by each of a base set of content contributing entities (E 1 , E 2  . . . Em) for a first topic;   wherein a collective information contribution is computed for (E 1 , E 2  . . . Em) for said first topic based on measuring individual information units from (E 1 , E 2  . . . Em) mapped by the networked computing system to said first topic;   b. for a first candidate contributing entity (En) who is not part of the first user's datastream, automatically identifying with the networked computing system an information contribution for said first topic to the user datastream;
 wherein a first individual information contribution is computed for (En) for said first topic based on measuring individual information units from (En) mapped by the networked computing system to said first topic; 
   c. automatically comparing with the networked computing system said collective information contribution and said first individual information contribution for said first topic;
 wherein said comparing includes at least a first computation determining a first net information gain achieved over said collective information contribution by including said first individual information contribution for said first topic; and 
   d. automatically recommending to said user with the networked computing system a recommendation that said first candidate contributing entity (En) be included for contributing content to the user datastream when said first net information gain exceeds a target threshold.   
     
     
         20 . The method of  claim 19 , further including a step: controlling the user datastream to automatically include content from said first candidate contributing entity (En) in response to the user accepting said recommendation. 
     
     
         21 . The method of  claim 19 , further including a step: obtaining a target coverage value specified by the user for said first topic, and using said target coverage value to determine an optimal set of content contributing entities required to achieve said coverage value for said first topic within the datastream. 
     
     
         22 . The method of  claim 21 , wherein said optimal set of content contributing entities includes the smallest number of entities required to achieve said coverage value for said first topic within the datastream. 
     
     
         23 . The method of  claim 21 , wherein said optimal set of content contributing entities includes the smallest number of entities required both to achieve said coverage value for said first topic within the datastream and maintain a duplication rate below a user selectable value. 
     
     
         24 . The method of  claim 20 , wherein information duplication is minimized for said first topic by filtering messages that include the same content as prior messages shown within the datastream. 
     
     
         25 . The method of  claim 19 , wherein the individual information units include uniform resource locator links within electronic messages. 
     
     
         26 . The method of  claim 19 , wherein steps (b) and (c) are performed over a predetermined time window. 
     
     
         27 . The method of  claim 26 , wherein said predetermined time window can be selected by the user. 
     
     
         28 . The method of  claim 19 , further including a step: generating a visual display output for the user, indicating graphically a predicted datastream message feed rate or predicted change in datastream message feed rate achieved based on adding or removing said first candidate contributing entity (En). 
     
     
         29 . The method of  claim 19 , further including a step: generating a visual display output for the user, indicating graphically a value of a data duplication rate present within the user's datastream as calculated by the networked computing system. 
     
     
         30 . The method of  claim 19 , wherein individual ones of said base set of content contributing entities are determined automatically for the user in response to the user selecting said first topic for the user datastream. 
     
     
         31 . The method of  claim 19 , wherein individual ones of said base set of content contributing entities are determined by the user as part of a whitelist. 
     
     
         32 . A method of recommending entities to a user for inclusion as content contributors for a user datastream using a networked computing system comprising:
 a. automatically identifying with the networked computing system an information contribution provided within a first user datastream by each of a base set of content contributing entities (E 1 , E 2  . . . Em) for a first topic
 wherein a collective information contribution is computed for (E 1 , E 2  . . . Em) for said first topic based on measuring individual information units from (E 1 , E 2  . . . Em) mapped by the networked computing system to said first topic; 
   b. for a first candidate contributing entity (En) who is not part of the first user's datastream, automatically identifying with the networked computing system an information contribution for said first topic to the user datastream;
 wherein a first individual information contribution is computed for (En) for said first topic based on measuring individual information units from (En) mapped by the networked computing system to said first topic; 
   c. automatically comparing with the networked computing system said collective information contribution and said first individual information contribution for said first topic;
 wherein said comparing includes at least a first computation determining a first net information gain achieved over said collective information contribution by including said first individual information contribution for said first topic; 
   d. repeating steps (b) and (c) for a second candidate contributing entity (Ep) to determine a second net information gain achieved over said collective information contribution by including a second individual information contribution for said first topic from (Ep);   e. measuring an overlap of individual information units between said first individual information contribution for said first topic from (Ep) and said second individual information contribution for said first topic from (Ep);   f. comparing said first net information gain and said second net information gain; and   g. automatically recommending one or both of said first candidate contributing entity and said second candidate contributing entity to said user with the networked computing system based on results of step (e) and step (f).   
     
     
         33 . The method of  claim 32 , wherein said recommending step is also based on measuring second content outside said first topic contributed by said first candidate contributing entity and said second candidate contributing entity, which second content reduces a recommendation score for such entities.