INFERRING VIEW SEQUENCE AND RELEVANCE DATA

• US 20110010323A1
• Filed: 07/08/2009
• Published: 01/13/2011
• Est. Priority Date: 07/08/2009
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Li, Xiaodong, Gloy, Nikolas Clemens, Wang, Kuansan
• Granted Patent: US 8,489,533 B2
• Time in Patent Office: Days
• US Class Current: 706/46
• CPC Class Codes: G06F 9/451 Execution arrangements for user interfaces

Abstract

A system described herein includes a receiver component that receives action sequence data pertaining to a plurality of selectable graphical items displayed to a user on a computer display screen, wherein the action sequence data is indicative of an order in which the user performed an action with respect to a first subset of graphical items in the plurality of selectable graphical items. The system also includes a determiner component that infers a view sequence with respect to a second subset of graphical items in the plurality of graphical items, wherein the view sequence is an order that the user viewed graphical items in the second subset of graphical items, and wherein the second subset of graphical items comprises the first subset of graphical items and at least one other graphical item in the plurality of graphical items.

Claims

1. A method comprising the following computer-executable acts:

receiving action sequence data pertaining to a graphical user interface displayed to a user on a display, wherein the graphical user interface comprises a plurality of graphical items, and wherein the action sequence data is indicative of a sequence of actions of the user with respect to a first subset of graphical items in the plurality of graphical items;

inferring view sequence data with respect to a second subset of graphical items in the plurality of graphical items based at least in part upon the action sequence data, wherein the view sequence data is indicative of an order in which the user viewed the second subset of graphical items, and wherein the second subset of the graphical items comprises the first subset of the graphical items and at least one additional graphical item in the plurality of graphical items.

2. The method of claim 1, wherein the action sequence data comprises an indication that the user selected at least one graphical item in the first subset of graphical items and a time corresponding to the user selecting the at least one graphical item.

3. The method of claim 1, wherein the action sequence data comprises an indication that the user hovered over at least one graphical item in the first subset of graphical items and a time corresponding to the user hovering over the at least one graphical item.

4. The method of claim 1, wherein the action sequence data comprises an indication of an order of selections of the user of search results displayed to the user responsive to receipt of a query issued by the user, and wherein the view sequence data is indicative of an order that a subset of the search results displayed to the user were viewed by the user.

5. The method of claim 4, further comprising inferring a measure of relevance for at least one search result displayed to the user with respect to the query based at least in part upon the view sequence data.

6. The method of claim 5, wherein the measure of relevance is inferred based at least in part upon a dwell time corresponding to a search result selected by the user.

7. The method of claim 1, wherein the view sequence data is inferred based at least in part upon an amount of time between actions of the user in the action sequence data.

8. The method of claim 1, further comprising altering a design of the graphical user interface based at least in part upon inferred view sequence data with respect to multiple users of the graphical user interface, wherein the view sequence data is inferred based at least in part upon action sequence data pertaining to the multiple users of the graphical user interface.

9. The method of claim 1 configured for execution over search engine logs.

10. The method of claim 1, wherein a Markov Model is utilized in connection with inferring the view sequence data.

11. The method of claim 1, wherein inferring the view sequence data comprises:

generating a transition probability matrix, wherein a number of graphical items in the plurality of graphical items is n, wherein the transition probability matrix is an n×

n matrix, and wherein each element of the transition probability matrix comprises a probability that a jth graphical item is selected after an ith graphical item is selected, wherein the view sequence data is inferred based at least in part upon the transition probability matrix.

12. The method of claim 11, wherein inferring the view sequence data further comprises:

computing a skip matrix and a view transition matrix based at least in part upon the transition probability matrix, wherein the skip matrix is an n×

n matrix, and wherein an element j, j in the skip matrix is a probability that the jth graphical item is viewed and not selected by the user, wherein the view transition matrix is an n×

n matrix, wherein element i, j in the view transition matrix is a probability that the user views the jth graphical item immediately subsequent to viewing the ith graphical item.

13. The method of claim 12, wherein the view sequence data is inferred based at least in part upon the following algorithm:

C=(1−

α

)V(I−

α

SV)<sup>−

1</sup>(I−

S),

where C is the transition probability matrix, V is the view transition matrix, S is the skip matrix, α

is the average view rate, and I is the identity matrix.

14. The method of claim 13, further comprising utilizing a Viterbi algorithm to compute the view sequence data.

15. A system comprising the following computer-executable components:

a receiver component that receives action sequence data pertaining to a plurality of selectable graphical items displayed to a user on a computer display screen, wherein the action sequence data is indicative of an order in which the user performed an action with respect to a first subset of graphical items in the plurality of selectable graphical items; and

a determiner component that infers a view sequence with respect to a second subset of graphical items in the plurality of graphical items, wherein the view sequence is an order that the user viewed graphical items in the second subset of graphical items, and wherein the second subset of graphical items comprises the first subset of graphical items and at least one other graphical item in the plurality of graphical items.

16. The system of claim 15, wherein the action sequence data is included in a search engine log, wherein the search engine log comprises data indicating which search results were selected by the user, time that the search results were selected by the user, which search results were hovered over by the user, and time that the search results were hovered over by the user.

17. The system of claim 16, wherein the plurality of graphical items are search results displayed to the user responsive to receipt of a search query issued by the user, and further comprising a relevance assignor component that assigns a measure of relevance to at least one of the search results with respect to the query based at least in part upon the view sequence.

18. The system of claim 15, wherein the determiner component infers a probability that at least one search result was viewed and skipped by the user in connection with inferring the view sequence of the user.

19. The system of claim 15, wherein the plurality of graphical items comprises a first selectable search result and a second selectable search result, and wherein the determiner component infers a probability that the user viewed the first search result and then immediately subsequently viewed the second search result.

20. A computer-readable medium comprising instructions that, when executed by a processor, cause the processor to perform the following acts:

receive a search log, wherein the search log comprises a query issued by a user, a plurality of search results displayed to the user upon receipt of the query, a first subset of search results selected by the user, and a time corresponding to when each of the first subset of search results were selected by the user;

infer a view sequence of the user based at least in part upon the received search log, wherein the view sequence is indicative of an order that a second subset of search results in the plurality of search results were viewed by the user, wherein the view sequence of the user includes at least one search result not selected by the user; and

update a relevance measure for at least one search result with respect to the query based at least in part upon the view sequence.