Search

CN-115114454-B - Keyword retrieval method based on knowledge graph associated search recommendation

CN115114454BCN 115114454 BCN115114454 BCN 115114454BCN-115114454-B

Abstract

The invention discloses a keyword retrieval method based on knowledge graph association search recommendation. According to the method, the related search can be carried out on the keywords based on the knowledge graph, corresponding related object examples are recommended to the user according to the characteristics selected by the user, and the purpose of rapidly recommending useful information to the user is achieved. The method comprises the steps of A, receiving keywords, B, searching an object instance corresponding to the keywords, searching and obtaining a plurality of associated object instances of the object instance according to attribute similarity and/or relationship compactness, C, forming a knowledge graph by the object instance and the associated object instance, D, performing cluster analysis on object instance nodes in the graph according to time characteristics, node type characteristics, node attribute characteristics and relationship characteristics, setting feature options for the graph according to the clustering analysis, E, outputting the graph and feature options thereof to a user, F, selecting object instance nodes conforming to the features from the graph according to the features selected by the user, and recommending the object instance nodes to the user.

Inventors

  • CAI HUI
  • CAI CHUNYUAN
  • LU YOUFEI
  • LIANG XUEQING
  • WANG YONGCAI
  • ZHANG CHUNMEI
  • ZOU SHIRONG
  • FENG XINYAO
  • SHAO YANNING
  • CHEN ZHIMING
  • YANG JINGJING
  • PEI QIUGEN
  • QIAN ZHENGHAO
  • JIANG JIANG
  • PENG ZEWU
  • XIE HANYANG
  • WU HENG
  • WANG JUNFENG

Assignees

  • 广东电网有限责任公司

Dates

Publication Date
20260512
Application Date
20220708

Claims (4)

  1. 1. A keyword retrieval method based on knowledge graph association search recommendation is characterized by comprising the following steps: step A, receiving keywords input by a user; step B, searching an object instance corresponding to the keyword from a knowledge graph database, and searching a plurality of associated object instances of the object instance from the knowledge graph database according to attribute similarity and/or relationship compactness; Step C, forming a knowledge graph by the object instance corresponding to the keyword and the associated object instance; step D, performing cluster analysis on the object instance nodes in the knowledge graph according to the time feature, the node type feature, the node attribute feature and the relationship feature, and setting feature options for the knowledge graph according to a cluster analysis result, wherein the feature options comprise one or more of the time feature options, the node type feature options, the node attribute feature options and the relationship feature options; step E, outputting the knowledge graph and characteristic options thereof to a user; Step F, selecting object instance nodes conforming to the characteristics from the knowledge graph according to the characteristics selected by the user and recommending the object instance nodes to the user; g, according to the object instance node selected by the user, displaying a plurality of association expansion options of different types for the user to carry out association expansion on the selected object instance node; And H, receiving an associated expansion option selected by a user, wherein the option carries object instance node information selected by the user, searching an associated object instance of the object instance node from a knowledge graph database according to attribute similarity and/or relationship closeness, recommending the associated object instance of which the instance type belongs to the associated expansion option type to the user, and particularly, displaying the associated object instance as a node superposition in a knowledge graph of the keyword in the step H.
  2. 2. The method of claim 1, wherein step F comprises real-time linking recommending object instance nodes according with the feature to the user according to the feature selected by the user.
  3. 3. The method of claim 1, wherein step F is performed by highlighting object instance nodes corresponding to the features in the knowledge graph to recommend the object instance nodes to the user.
  4. 4. The keyword retrieval method based on knowledge graph association search recommendation of claim 1, wherein the plurality of association expansion options of different types are specifically an entity, an event, a document, a concept and all of the five types of association expansion options.

Description

Keyword retrieval method based on knowledge graph associated search recommendation Technical Field The invention mainly relates to the technical field of keyword retrieval, in particular to a keyword retrieval method based on knowledge-graph associated search recommendation. Background The knowledge graph is a semantic network for revealing the relation between entities, and the prior art in the power industry utilizes the knowledge graph technology to construct the knowledge graph of the power service domain so as to realize the visual display of each power equipment and the relation between the power equipment. However, the power business involves various links such as transmission, transformation, distribution, use and the like, and the number of related power equipment is usually very large, the relationship is complicated, one equipment is used as an object instance node, the object instance node of the constructed knowledge graph is very many, the relationship is complicated, and a user is difficult to quickly acquire useful information from the object instance node. Disclosure of Invention The invention aims to solve the technical problem of how to rapidly recommend useful information to a user and reduce the energy consumed by the user for acquiring the information. In order to solve the technical problems, the invention provides a keyword retrieval method based on knowledge graph association search recommendation, which comprises the following steps: step A, receiving keywords input by a user; step B, searching an object instance corresponding to the keyword from a knowledge graph database, and searching a plurality of associated object instances of the object instance from the knowledge graph database according to attribute similarity and/or relationship compactness; Step C, forming a knowledge graph by the object instance corresponding to the keyword and the associated object instance; step D, performing cluster analysis on the object instance nodes in the knowledge graph according to the time feature, the node type feature, the node attribute feature and the relationship feature, and setting feature options for the knowledge graph according to a cluster analysis result, wherein the feature options comprise one or more of the time feature options, the node type feature options, the node attribute feature options and the relationship feature options; step E, outputting the knowledge graph and characteristic options thereof to a user; And F, selecting object instance nodes conforming to the characteristics from the map according to the characteristics selected by the user and recommending the object instance nodes to the user. Further, in step F, according to the feature selected by the user, the object instance node conforming to the feature is recommended to the user in real time in a linkage manner. Further, step F specifically highlights the object instance nodes conforming to the features in the knowledge graph to recommend to the user. Further, the method also comprises the following steps: g, according to the object instance node selected by the user, displaying a plurality of association expansion options of different types for the user to carry out association expansion on the selected object instance node; And step H, receiving an associated expansion option selected by a user, wherein the option carries object instance node information selected by the user, searching an associated object instance of the object instance node from a knowledge graph database according to attribute similarity and/or relationship closeness, and recommending the object instance with the instance type belonging to the associated expansion option type to the user. Further, in the step H, specifically, the object instance is displayed as a node superposition in the knowledge graph of the keyword. Further, the plurality of different types of associated expansion options are specifically an entity, an event, a document, a concept and all of the five types of associated expansion options. According to the keyword searching method based on the knowledge graph associated search recommendation, the keyword can be subjected to the associated search based on the knowledge graph, the associated object instance conforming to the characteristics is recommended to the user according to the characteristics selected by the user, the purpose of rapidly recommending useful information to the user is achieved, and the energy consumed by the user for information acquisition is reduced. Drawings Fig. 1 is a flowchart of a keyword retrieval method based on knowledge-graph associated search recommendation provided by the invention. Fig. 2 is a schematic diagram of a retrieval process of a keyword retrieval method based on knowledge-graph related search recommendation. Detailed Description The invention is further described in detail below in connection with the detailed description. The keyword retrieval method based on the knowledge