ESTABLISHING SEARCH RESULTS AND DEEPLINKS USING TRAILS

• US 20110264673A1
• Filed: 04/27/2010
• Published: 10/27/2011
• Est. Priority Date: 04/27/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Huang, Jeff, White, Ryen W., Dandekar, Nikhil, Singla, Adish, Bailey, Peter
• Granted Patent: US 10,289,735 B2
• Time in Patent Office: Days
• US Class Current: 707/750
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9538 Presentation of query results

Abstract

Search and browse trails are temporally-ordered sequences of web pages visited by a user during post-search query navigation beginning with a page associated with one of the search results. The trails can provide useful information for a number of search-related purposes. For example, these trails can be used to leverage the post-query behavior of other users to help the current user search more effectively and allow them to make more informed search interaction decisions. The trails can also be used to establish search results and refine search result rankings, select and evaluate deeplinks, and recommend multi-step trails as an alternative to or enhancement for existing search result presentation techniques.

Claims

1. A computer-implemented process for establishing search results based in part on search and browse trail data, said process comprising:

using a computer to perform the following process actions;

accessing trails found within said trail data, wherein each trail comprises a sequence of URLs each of which was selected in turn by a prior user subsequent to selecting a search result presented in response to a prior search query;

identifying one or more trails that match a current search query to a prescribed degree; and

employing the identified trail or trails to select one or more URLs found therein for inclusion in search results found by a search engine in response to the current search query.

2. The process of claim 1, wherein the process action of identifying one or more trails that match a current search query to a prescribed degree, comprises the actions of:

for each response to the current search query, computing a score for each trail having a URL associated with the search result under consideration as its starting point, based on the frequency that the trail is followed when a search query associated with the trail includes a term common to the current search query;

identifying those trails having a score that exceeds a prescribed score threshold up to a prescribed number of trails; and

designating the identified trails as matching the current search query.

3. The process of claim 2, wherein the process action of computing a score for each trail having a URL associated with the search result under consideration as its starting point, comprises the actions of:

for each term in the current search query,

computing a value representing the frequency that the trail under consideration is followed when the search query associated with the trail includes the term under consideration, and

weighting the computed frequency value based on one or more trail selection criteria;

summing the weighted frequency values computed for each term of the current search query; and

designating the summed weighted frequency values as the score for the trail under consideration with respect to the search result under consideration.

4. The process of claim 3, wherein the trail selection criteria comprises the frequency that the current search term under consideration is found in the search result under consideration, and at least one of:

trail length in terms of the number of sequential URLs comprising the trail;

or

domain diversity in terms of the number of URLs comprising the trail that is outside the domain of the URL associated with the search result under consideration;

or

trail strength in terms of the trail length and ease of navigation;

or

trail breadth in terms of the number of branches from the first URL;

or

trail depth in terms of the maximum number of nodes in a branch from the first URL;

or

a number of URLs in the trail having the current search query in their title or URL string.

5. The process of claim 1, wherein the process action of employing the identified trail or trails to select one or more URLs found therein for inclusion in search results found by a search engine in response to the current search query, comprises the actions of:

selecting one or more URLs found in the identified trails to be included in the search results; and

adding the selected one or more URLs directly to the search results with an arbitrary rank.

6. The process of claim 5, wherein the process action of selecting one or more URLs found in the identified trails, comprises an action of selecting a trail URL whenever at least one of a number of inclusion factors exists, said inclusion factors comprising:

the trail URL exceeds a prescribed click count threshold, wherein the click count refers to the number of times a URL appears in the one or more identified trails;

or

the trail URL is the first URL in the trail;

or

the trail URL is the last URL in the trail;

or

the trail URL is the URL in an identified trail on which a user associated with that trail spent the most time in comparison to the other URLs in that trail;

or

the trail URL is a URL in an identified trail on which a user associated with the trail spent an amount of time equaling or exceeding a prescribed threshold.

7. The process of claim 5, wherein the process action of selecting one or more URLs found in the identified trails, comprises an action of selecting a trail URL whenever at least one of a number of inclusion factors exists, said inclusion factors comprising:

the trail URL in an identified trail is outside the domain of the first URL in that trail;

or

the trail URL in an identified trail exhibits a depth which equals or exceeds a prescribed depth threshold;

or

the trail URL in an identified trail exhibits a breadth which equals or exceeds a prescribed breadth threshold.

8. The process of claim 5, wherein the process action of selecting one or more URLs found in the identified trails, comprises an action of selecting only those URLs with a title or URL string, or both, which includes at least one term found in a search query associated with the trail that includes the URL.

9. The process of claim 1, wherein the process action of employing the identified trail or trails to select one or more URLs found therein for inclusion in search results found by a search engine in response to the current search query, comprises the actions of:

selecting one or more URLs found in the identified trails, wherein one or more of the selected URLs are to be potentially included in the search results; and

subjecting the selected URLs along with the search results to a re-ranking operation to produce revised search results that potentially include at least one of the selected URLs.

10. The process of claim 9, further comprising, prior to subjecting the selected URLs along with the search results to the re-ranking operation, performing a process action of adding the current search query to the title or URL string, or both, of the selected URLs.

11. The process of claim 9, wherein the process action of selecting one or more URLs found in the identified trails, comprises an action of selecting a trail URL whenever at least one of a number of inclusion factors exists, said inclusion factors comprising:

the trail URL exceeds a prescribed click count threshold, wherein the click count refers to the number of times a URL appears in the one or more identified trails;

or

the trail URL is the first URL in the trail;

or

the trail URL is the last URL in the trail;

or

the trail URL is the URL in an identified trail on which a user associated with that trail spent the most time in comparison to the other URLs in that trail;

or

the trail URL is a URL in an identified trail on which a user associated with the trail spent an amount of time equaling or exceeding a prescribed threshold.

12. The process of claim 9, wherein the process action of selecting one or more URLs found in the identified trails, comprises an action of selecting a trail URL whenever at least one of a number of inclusion factors exists, said inclusion factors comprising:

the trail URL in an identified trail is outside the domain of the first URL in that trail;

or

the trail URL in an identified trail exhibits a depth which equals or exceeds a prescribed depth threshold;

or

the trail URL in an identified trail exhibits a breadth which equals or exceeds a prescribed breadth threshold.

13. The process of claim 9, wherein the process action of selecting one or more URLs found in the identified trails, comprises an action of selecting only those URLs with a title or URL string, or both, which includes at least one term found in a search query associated with the trail that includes the URL.

14. A computer-implemented process for establishing a set of one or more deeplinks for a search result presented by a search engine using search and browse trail data, said process comprising:

using a computer to perform the following process actions;

accessing navigational trails within said search and browse data, wherein each navigational trail comprises a sequence of URLs in the same domain which a prior user visited, starting with a first navigational URL corresponding to said search result and ending at a destination URL, wherein a destination URL is a URL associated with a site where the prior user dwelled for more than a prescribed period of time;

computing a time to destination for each of the accessed navigational trails, wherein said time to destination is defined as the sum of an effective time the prior user associated with the navigational trail dwelled on each URL in the sequence of URLs preceding the destination URL;

computing a total time to destination for each group of navigational trails ending in the same destination URL, wherein said computing of the total time to destination for each group comprises computing the sum of the time to destination computed for each of the navigational trails making up the group; and

establishing the set of deeplinks based on the computed total times to destination.

15. The process of claim 14, wherein the process action of establishing the set of deeplinks based on the computed total times to destination, comprises an action of designating the destination URL of each of said groups of navigational trails as a deeplink whenever that group'"'"'s total time to destination exceeds a prescribed deeplink threshold.

16. The process of claim 14, wherein the process action of establishing the set of deeplinks based on the computed total times to destination, comprises an action of designating as deeplinks a prescribed number of the destination URLs whose associated groups of navigational trails exhibit a higher total time to destination in comparison to the other groups of navigational trails.

17. The process of claim 14, wherein the process action of establishing the set of deeplinks based on the computed total times to destination, comprises an action of designating the destination URL of each of said groups of navigational trails as a deeplink whenever that group'"'"'s total time to destination exceeds a prescribed deeplink threshold, up to a prescribed number.

18. The process of claim 14, wherein a navigational trail further comprises one or more intermediate destination URLs, wherein an intermediate destination URL is a URL associated with a site where a prior user dwelled for more than the prescribed period of time, and wherein the effective time of an intermediate destination URL is deemed to be said prescribed period of time for the purpose of computing the time to destination for the navigational trail which includes the intermediate destination URL.

19. The process of claim 14, wherein the process action of computing the total time to destination for each group of navigational trails further comprises an action of dividing the sum of the times to destination computed for each of the navigational trails in the group under consideration by the number of prior users who navigated via a plurality of URLs from the first navigational URL to the destination URL associated with the group, multiplied by the sum of the number of prior users who navigated via a plurality of URLs from the first navigational URL to the destination URL associated with the group and a number of prior users who employed a deeplink associated with said search result to reach the destination URL associated with the group.

20. A computer-implemented process for presenting search results based on search and browse trail data, said process comprising:

using a computer to perform the following process actions;

accessing trails found within said trail data, wherein each trail comprises a sequence of URLs each of which was selected in turn by a prior user subsequent to selecting a search result presented in response to a prior search query;

inputting a current search query;

identifying one or more trails that match the current search query to a prescribed degree; and

presenting the identified trail or trails.