US 12,066,989 B2
Method and apparatus for constructing knowledge graph
Xian Yi Ye, Beijing (CN)
Assigned to SIEMENS LTD., CHINA, Beijing (CN)
Appl. No. 18/043,717
Filed by Siemens Ltd., China, Beijing (CN)
PCT Filed Sep. 10, 2020, PCT No. PCT/CN2020/114572
§ 371(c)(1), (2) Date Mar. 1, 2023,
PCT Pub. No. WO2022/051996, PCT Pub. Date Mar. 17, 2022.
Prior Publication US 2023/0350861 A1, Nov. 2, 2023
Int. Cl. G06F 16/21 (2019.01); G06F 16/22 (2019.01)
CPC G06F 16/212 (2019.01) [G06F 16/2282 (2019.01)] 10 Claims
OG exemplary drawing
 
1. A method for constructing an application based on a knowledge graph representing data structures, the method comprising:
acquiring training data relating to a field as user data in different subsystems of an industrial system, wherein the user data is data to be subjected to data structure identification;
using the training data to train a classification model generating, based on an attribute value, an entity attribute to which the attribute value belongs;
for each column of data in the user data, respectively providing each attribute value included in the column of data into the classification model, to obtain an entity attribute to which each attribute value in the column of data belongs, and determining an entity attribute described by the column of data according to the entity attribute to which each attribute value in the column of data belongs;
subjecting the user data to data structure identification to determine a mapping relationship between the columns of data in the user data;
determining a mapping relationship between the entity attributes described by the columns of data in the user data according to the entity attributes described by the columns of data in the user data and the mapping relationship between the columns of data in the user data;
determining a mapping relationship between entities corresponding to the entity attributes described by the columns of data in the user data according to the mapping relationship between the entity attributes described by the columns of data in the user data and the entities corresponding to the entity attributes;
generating a knowledge graph comprising the entities and the mapping relationship between the entities;
generating a relationship mapping table according to the entity attributes described by columns of data in the user data as determined and the mapping relationship between the entity attributes described by the columns of data in the user data as determined; and
displaying the knowledge graph during development of the application to indicate which attribute values have a mapping relationship with an entity attribute.