Determination method and device for recommended search keyword

Abstract

The embodiment of the invention discloses a determination method and device for a recommended search keyword. The method comprises the steps that a target search term input by a user is obtained; whether the target search term is an entity is judged; if yes, an entity related to the target search term is determined according to a pre-established knowledge graph representing the relation between entities; and a first recommended search keyword corresponding to the target search term is determined according to the determined entity. By use of the embodiment, information relevance between the given recommended search keyword and a keyword input by the user is improved.

Description

technical field [0001] The invention relates to the field of search technology, in particular to a method and device for determining recommended search keywords. Background technique [0002] In today's information age, the Internet has developed rapidly and provided many conveniences for people's production and life, and people are increasingly using the Internet to obtain corresponding services. [0003] At present, when a user uses the search service, the search result page will display the search result, and will also give recommended search keywords of "you can also search" on the search result page according to the search term entered by the user. Taking the video website as an example, the existing recommended search keyword determination method is to check the query (query) behavior under the user session (session) and the same doc (document, document) in the search log of the user. Click on the query behavior to get it by mining. [0004] However, the existing det...

AI Technical Summary

Problems solved by technology

[0004] However, the existing determination methods are obtained based on the user's historical behavior, and do not consider when the search term is a named entity ("entity" for short, which refers to things that exist objectively and can be distinguished from each other) that widely exists in natural language. , the important relationship between the search term and some entities, ignoring those entities that have an important relationship with the search term, resulting in a low correlation between the recommended search keywords and the keywords entered by the user

For example, when the search entity is "Li Chen", the current recommended search keywords for "You can also search" do not have the entity "Fan Bingbing" that has a significant relationship with "Li Chen", but are based on the user's historical behavior. In this way, the information correlation between the recommended search keywords and the keywords entered by the user is not high, and the user's search experience is not high.

Images