Search engine and method for implementing the same

Patent metadata

Field	Value
Publication number	US-2016371379-A1
Application number	US-201514954784-A
Country	US
Kind code	A1
Filing date	Nov 30, 2015
Priority date	Jun 18, 2015
Publication date	Dec 22, 2016
Grant date	—

Abstract

Search engine method includes: receiving a user query request input; searching candidate results matching with the query request; determining a semantic relatedness between the query request and each candidate result based on a click-escape model; sorting the candidate results according to the semantic relativity. The click-escape model has an escape dictionary, a non-escape dictionary, or a combination thereof. Sorting candidate results of a search in accordance with a semantic relatedness can enhance the sorting effect of the searched results, avoid searched results which do not match the user's query appearing in the forefront of the searched result list, and guarantee a good user experience.

First claim

What is claimed is:

1 . A method for implementing a search engine, comprising:
receiving a query request input by a user; searching candidate results that match the query request; determining a semantic relatedness between the query request and each of the candidate results, based on a click-escape model; and sorting the candidate results according to the semantic relatedness, wherein the click-escape model comprises an escape dictionary, a non-escape dictionary, or a combination thereof, the escape dictionary comprises a term corresponding to the searched result which is determined to escape and a preceding term and a following term of the term, and the non-escape dictionary comprises a term corresponding to the searched result which is determined not to escape and a preceding term and a following term of the term.

2 . The method of claim 1 , wherein determining the semantic relatedness between the query request and each of the candidate results comprises, for each of the candidate results:
determining the semantic relatedness between one or more elements of the candidate result and the query request, wherein the element comprises at least one of: a title of the candidate result, an anchor text and a core sentence of a text; and determining the semantic relatedness between the query request and the candidate result according to the determined semantic relatedness between one or more elements of the candidate result and the query request.

3 . The method of claim 2 , wherein determining the semantic relatedness between one or more elements of the candidate result and the query request further comprises:
calculating a theme matching similarity between the element of the candidate result and the query request by means of an inter-element text theme matching model, based on the click-escape model; determining an escape factor according to a matching condition between the element of the candidate result and the query request; and calculating the semantic relatedness between the element of the candidate result and the query request, based on the escape factor and the theme matching similarity.

4 . The method of claim 3 , wherein calculating the theme matching similarity between the element of the candidate result and the query request based on the click-escape model comprises:
determining an adjacent preceding term and an adjacent following term aligned with a term in the query request from the element of the candidate result by means of term aligning; adjusting similarity weights of the corresponding preceding term and the corresponding following term in the element of the candidate result according to the escape dictionary, the non-escape dictionary, or a combination thereof; and calculating the theme matching similarity between the element of the candidate result and the query request by the inter-element text theme matching model, according to the adjusted similarity weights.

5 . The method of claim 4 , wherein adjusting the similarity weights of the corresponding preceding term and the corresponding following term in the element of the candidate result according to the escape dictionary, the non-escape dictionary, or a combination thereof comprises:
if the non-escape dictionary comprises a corresponding term and the preceding term or the following term of the corresponding term in the element of the candidate result, reducing the similarity weight of the preceding term or the following term; and if the escape dictionary comprises the corresponding term and the preceding term or the following term of the corresponding term in the element of the candidate result, increasing the similarity weight of the preceding term or the following term.

6 . The method of claim 4 , wherein the inter-element text theme matching model is a vector space model and is expressed as:

Sim


(

Q
,
S

)

=

∑

w

1

k

=

w

2

l



(

Wgt


(

w

1

k

)

*

Wgt


(

w

2

l

)

)

∑

k
=

1

…



M





Wgt


(

w

1

k

)

2



∑

l

Assignees

• Baidu online network technology beijing co ltd

Inventors

• Fang Gaolin

Classifications

• G06F16/24578

  using ranking · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/24575

  using context · CPC title

• G06F40/30

  Semantic analysis · CPC title

• G06F16/951

  Indexing; Web crawling techniques · CPC title