Promoting websites based on location

A computer-implemented method to associate a location with a website, the method comprising:

• retrieving index data having page content for websites;
• organizing the index data based on websites;
• extracting anchors within each website pointing to targets;
• for each website, obtaining anchor geographic locations for each anchor from a location database;
• grouping the anchors based on the obtained anchor geographic locations;
• counting a number of anchors having each of the obtained anchor geographic locations;
• counting a number of users that visit the website;
• identifying a user geographic location for each user;
• grouping the users based on the identified user geographic location;
• selecting the user geographic location having more of the user visits to the website than any other user geographic location;
• selecting the obtained anchor geographic location having more of the anchors than any other obtained anchor location;
• associating the website with one of:
• the selected anchor geographic location if anchors and anchor geographic locations are available or user geographic location if anchors and anchor geographic locations are unavailable;
• ranking the websites based on the associated location using a Kullback-Leibler divergence, wherein a website is identified as unexpectedly popular when a probability that the users visiting the website from at least one geographic location other than the associated location is greater than a popularity expectation of the associated location, wherein the popularity expectation represents a probability that the users that visit the website are geographically located at the associated location;
• and
• increasing the rank of the website that is identified as unexpectedly popular to be higher in rank than the ranks assigned to other websites that are not identified as unexpectedly popular.

A computer system having hardware computer-readable media and processors configured to perform a method to associate a location with a website, the method comprising:

• retrieving index data having page content for websites;
• organizing the index data based on websites;
• extracting anchors within each website pointing to targets;
• for each website, obtaining anchor geographic locations for each anchor from a location database;
• grouping the anchors based on the obtained anchor geographic locations;
• counting a number of anchors having each of the obtained anchor geographic locations;
• counting a number of users that visit the website;
• identifying a user geographic location for each user;
• grouping the users based on the identified user geographic location;
• selecting the user geographic location having more of the user visits to the website than any other user geographic location;
• selecting the obtained anchor geographic location having more of the anchors than any other obtained anchor location;
• associating the website with one of:
• the selected anchor geographic location if anchors and anchor geographic locations are available or user geographic location if anchors and anchor geographic locations are unavailable;
• ranking the websites based on the associated location using a Kullback-Leibler divergence, wherein a website is identified as unexpectedly popular when a probability that the users visiting the website from at least one geographic location other than the associated location is greater than a popularity expectation of the associated location, wherein the popularity expectation represents a probability that the users that visit the website are geographically located at the associated location;
• and
• increasing the rank of the website that is identified as unexpectedly popular to be higher in rank than the ranks assigned to other websites that are not identified as unexpectedly popular.