Query categorizer

• US 20080016046A1
• Filed: 07/26/2007
• Published: 01/17/2008
• Est. Priority Date: 07/14/2006
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Guha, Sudipta
• Granted Patent: US 7,664,744 B2
• Time in Patent Office: Days
• US Class Current: 707/3
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9538 Presentation of query results
• CPC Class Codes: G06F 16/954 Navigation, e.g. using categorised browsing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

In response to a user'"'"'s submission of a set of query terms through an Internet search engine'"'"'s user interface, the Internet search engine automatically ranks multiple verticals based on the estimated probabilities that those verticals will contain relevant content. After the verticals have been ranked, a set of the highest-ranked verticals is selected. Search results from each of the verticals in the selected set are presented to the user. Thus, the user may obtain a sample from verticals that are the most likely to contain relevant content. The search engine may rank a set of verticals based on historical search result selection data for similar sets of query terms. The historical search result selection data for a set of query terms indicates, for each vertical, the number of times that any user selected a search result from that vertical after conducting a search based on that set of query terms.

Claims

1. A method comprising performing a machine-executed operation involving instructions, wherein said instructions are instructions which, when executed by one or more processors, cause the one or more processors to perform certain steps including:

receiving a first set of query terms;

for each vertical in a plurality of verticals, determining a number of previous searches that both (a) were performed based on the first set of query terms and (b) produced a set of search results, in the vertical, from which at least one search result was selected by a user;

ranking verticals in the plurality of verticals based at least in part on said numbers determined for each vertical in the plurality of verticals, thereby generating a ranked list of verticals;

selecting one or more selected verticals from the ranked list of verticals based on ranks of verticals in the ranked list of verticals; and

presenting, for each particular vertical of the one or more selected verticals, one or more search results that are (a) generated based on the first set of query terms and (b) within the particular vertical;

wherein the machine-executed operation is at least one of (a) sending said instructions over transmission media, (b) receiving said instructions over transmission media, (c) storing said instructions onto a machine-readable storage medium, or (d) executing the instructions.

2. The method of claim 1, wherein the certain steps further include:

for each vertical in the plurality of verticals, determining a number of previous searches that were performed, in that vertical, based on a second set of query terms;

wherein ranking verticals in the plurality of verticals comprises ranking verticals in the plurality of verticals based at least in part on the numbers of previous searches that were performed, in that vertical, based on the second set of query terms.

3. The method of claim 2, wherein the certain steps further include:

determining the second set of query terms by selecting, from one or more previously submitted sets of query terms, a particular set of query terms that has a proximity, relative to the first set of query terms, that is above a specified threshold;

wherein the particular set of query terms differs from the first set of query terms.

4. The method of claim 3, wherein the proximity of the particular set of query terms relative to the first set of query terms is determined based at least in part on an extent to which a set of search results produced by a search for the particular set of query terms is similar to a set of search results produced by a search for the first set of query terms.

5. The method of claim 1, wherein ranking verticals in the plurality of verticals comprises:

determining, for each vertical in the plurality of verticals, a first value that is based on a number of search results that were selected in that vertical; and

generating a first sorted list of values by sorting a list of said first values.

6. The method of claim 5, wherein ranking verticals in the plurality of verticals comprises:

determining, for each vertical in the plurality of verticals, a second value that is based on a number of searches that were performed relative to that vertical; and

generating a second sorted list of values by sorting a list of said second values.

7. The method of claim 6, wherein ranking verticals in the plurality of verticals comprises:

setting a first pointer to point to a particular value in the first sorted list of values;

setting a second pointer to point to a particular value in the second sorted list of values;

determining a first vertical that corresponds to a value to which the first pointer points;

determining a second vertical that corresponds to a value to which the second pointer points; and

adding, to the ranked list of verticals, one of;

(a) the first vertical, and (b) the second vertical.

8. The method of claim 7, wherein adding, to the ranked list of verticals, one of the first vertical and the second vertical comprises:

determining a first delta based on a difference between (a) a value, in the second sorted list, that corresponds to the first vertical and (b) a value, in the first sorted list, that corresponds to the first vertical;

determining a second delta based on a difference between (a) a value, in the first sorted list, that corresponds to the first vertical, and (b) a value, in the second sorted list, that corresponds to the first vertical;

determining a third delta based on a difference between the first delta and the second delta;

determining whether the third delta is greater than a specified value;

if the third delta is not greater than the specified value, then adding the first vertical to the ranked list of verticals; and

if the third delta is greater than the specified value, then adding the second vertical to the ranked list of verticals.

9. The method of claim 7, wherein ranking verticals in the plurality of verticals comprises:

if the first vertical is added to the ranked list of verticals, then advancing the first pointer in the first sorted list; and

if the second vertical is added to the ranked list of verticals, then advancing the second pointer in the second sorted list.

10. The method of claim 7, wherein ranking verticals in the plurality of verticals comprises:

if the first vertical is added to the ranked list of verticals, then removing, from the first sorted list and the second sorted list, values that correspond to the first vertical; and

if the second vertical is added to the ranked list of verticals, then removing, from the first sorted list and the second sorted list, values that correspond to the second vertical.

11. The method of claim 1, wherein presenting, for each particular vertical of the one or more selected verticals, the one or more search results comprises:

presenting, on a page, two or more search results from a first vertical of the one or more selected verticals; and

presenting, on the same page, two or more search results from a second vertical of the one or more selected verticals;

wherein the first vertical is differs from the second vertical.

12. The method of claim 1, wherein presenting, for each particular vertical of the one or more selected verticals, the one or more search results comprises:

presenting, on a page, a first section of search results that indicates that the first section contains search results that were obtained from a first vertical of the one or more selected verticals; and

presenting, on the same page, a second section of search results that indicates that the first section contains search results that were obtained from a second vertical of the one or more selected verticals;

wherein the first vertical differs from the second vertical.

13. The method of claim 1, wherein the plurality of verticals comprises two or more of:

a “

web”

vertical, a “

video”

vertical, an “

images”

vertical, a “

local”

vertical, a “

shopping”

vertical, an “

answers”

vertical, an “

audio”

vertical, a “

directory”

vertical, a “

jobs”

vertical, and a “

news”

vertical.

14. The method of claim 1, wherein selecting one or more selected verticals from the ranked list of verticals based on ranks of verticals in the ranked list of verticals comprises selecting, from the ranked list of verticals, a specified number of verticals that are more highly ranked, in the ranked list of verticals, than any vertical that is not selected from the ranked list of verticals.