# US 2010/0306213 A1 · Merging Search Results

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2010/0306213 A1
- Document ID: US20100306213A1
- Status: Published application
- Kind: A1
- Application number: 12/473,158
- Filed: 2009-05-27
- Published: 2010-12-02
- Earliest claimed: 2009-05-27
- Assignees: Microsoft Corporation
- Inventors: Michael J. Taylor; Filiip Radlinski; Milad Shokouhi
- Web page: https://patentgrep.com/patent/US20100306213A1

## CPC

- G06F 16/9536: Search customisation based on social or collaborative filtering
- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Merging search results is required, for example, where an information retrieval system issues a query to multiple sources and obtains multiple results lists. In an embodiment a search engine at an Enterprise domain sends a query to the Enterprise search engine and also to a public Internet search engine. In embodiments, results lists obtained from different sources are merged using a merging model which is learnt using a machine learning process and updates when click-through data is observed for example. In examples, user information available in the Enterprise domain is used to influence the merging process to improve the relevance of results. In some examples, the user information is used for query modification. In an embodiment a user is able to impersonate a user of a specified group in order to promote particular results.

## Claims

### Claim 1

A method at an information retrieval system arranged to retrieve results from a plurality of different sources comprising:
at an input, receiving a query;
at an output, issuing the query to the plurality of sources at least one of which is a public domain search engine and at least one of which is a private domain search engine;
at the input receiving a results list from each of the plurality of sources;
arranging a processor to form a merged results list from the received results lists using a merging model;
arranging a user interface to present the merged results list;
observing user behavior in response to the presented merged results list; and
using the observed user behavior to update the merging model.

### Claim 2 (depends on 1)

A method as claimed in claim 1 wherein the query is received from a user and wherein the method further comprises receiving user information for that user and wherein the merging model takes into account the user information.

### Claim 3 (depends on 1)

A method as claimed in claim 1 which further comprises storing the merging model using a data structure comprising a plurality of weighted event variables describing a merged results list presentation event.

### Claim 4 (depends on 1)

A method as claimed in claim 1 wherein the step of arranging the processor to form a merged results list from the received results lists comprises using the merging model to determine a probability value for each result being the probability that the result will be clicked by a user and using the probability values to form the merged results list.

### Claim 5 (depends on 1)

A method as claimed in claim 1 which further comprises storing the merging model using a data structure comprising a plurality of event variables selected from any of: a search engine rank, textual features, textual features obtained from a snippet, an impersonating user setting, query frequency, a number of results from each of the sources, user information, past click history.

### Claim 6 (depends on 1)

A method as claimed in claim 1 which further comprises arranging the processor to select a subset of the received results lists from which to form the merged results list where that subset may comprise only one of the received results lists.

### Claim 7 (depends on 6)

A method as claimed in claim 6 which comprises arranging the processor to make the selection on the basis of probability values obtained from the merging model.

### Claim 8 (depends on 1)

A method as claimed in claim 1 which further comprises receiving a user input specifying a change to the plurality of different sources; arranging the processor to update the merged results list according to the change; and arranging the user interface to present the updated merged results list.

### Claim 9 (depends on 1)

A method as claimed in claim 1 which further comprises receiving a user input specifying an impersonating user condition; storing the merging model using a data structure comprising a plurality of event variables including an impersonating user event variable; and updating the merging model to a higher degree than for observed user behavior where no impersonating user condition applies.

### Claim 10 (depends on 1)

A method as claimed in claim 1 wherein the query is received from a user and wherein the method further comprises receiving user information for that user; modifying the received query before issuing it to the plurality of sources; and where the modification is on the basis of the user information.

### Claim 11 (depends on 1)

A method as claimed in claim 1 which further comprises modifying the received query before issuing it to the plurality of sources; and where the modification is source-specific.

### Claim 12 (depends on 1)

A method as claimed in claim 1 which further comprises arranging an evaluator to evaluate the merging model at intervals and modifying the merging model on the basis of the evaluation.

### Claim 13

A method at an information retrieval system arranged to retrieve results from a plurality of different sources comprising:
at an input, receiving a query from a user in a private domain;
at the input, receiving information about the user;
at an output, issuing the query to the plurality of sources at least one of which is in a public domain and at least one of which is in the private domain to which the information retrieval system has access;
at the input receiving a results list from each of the plurality of sources;
arranging a processor to form a merged results list from the received results lists using a merging model which takes into account the user information;
arranging a user interface to present the merged results list;
observing user behavior in response to the presented merged results list; and
using the observed user behavior to update the merging model.

### Claim 14 (depends on 13)

A method as claimed in claim 13 which further comprises storing the merging model using a data structure comprising a plurality of weighted event variables describing a merged results list presentation event.

### Claim 15 (depends on 13)

A method as claimed in claim 13 wherein the step of arranging the processor to form a merged results list from the received results lists comprises using the merging model to determine a probability value for each result being the probability that the result will be clicked by a user and using the probability values to form the merged results list.

### Claim 16

An information retrieval system comprising
an input arranged to receive a query;
an output, arranged to issue the query to a plurality of sources;
the input being arranged to receive a results list from each of the plurality of sources;
a memory storing a merging model in a data structure comprising a plurality of weighted event variables describing a merged results list presentation event;
a processor arranged to form a merged results list from the received results lists using the merging model;
a user interface arranged to present the merged results list;
and wherein the processor is also arranged to observe user behavior in response to the presented merged results list and use the observed user behavior to update the merging model.

### Claim 17 (depends on 16)

An information retrieval system as claimed in claim 16 wherein the processor is arranged to observe both explicit and implicit feedback.

### Claim 18 (depends on 16)

An information retrieval system as claimed in claim 16 wherein the sources are independent of one another and wherein at least one of the sources is in a public domain and at least one of the sources is in a private domain.

### Claim 19 (depends on 16)

An information retrieval system as claimed in claim 16 wherein the input is arranged to receive user information and wherein the memory stores the merging model such that it takes into account the user information

### Claim 20 (depends on 16)

An information retrieval system as claimed in claim 16 wherein the processor is arranged to modify the received query before it is issued to the plurality of sources such that the modification is on the basis of user information.
