US 12,405,935 B2
Data fusion method and apparatus based on knowledge graph, and electronic device
Yu Kai Ji, Shanghai (CN)
Assigned to SIEMENS LTD., CHINA, Beijing (CN)
Appl. No. 18/043,085
Filed by Siemens Ltd., China, Beijing (CN)
PCT Filed Aug. 31, 2020, PCT No. PCT/CN2020/112645
§ 371(c)(1), (2) Date Feb. 27, 2023,
PCT Pub. No. WO2022/041226, PCT Pub. Date Mar. 3, 2022.
Prior Publication US 2023/0315703 A1, Oct. 5, 2023
Int. Cl. G06F 16/22 (2019.01); G06F 16/24 (2019.01); G06F 16/2452 (2019.01); G06F 16/2457 (2019.01); G06F 16/28 (2019.01)
CPC G06F 16/2228 (2019.01) [G06F 16/24522 (2019.01); G06F 16/2457 (2019.01); G06F 16/288 (2019.01)] 17 Claims
OG exemplary drawing
 
1. A method comprising:
mapping an original data structure to a standard data structure by acquiring data fields in an original data structure, determining semantics of the data fields in the original data structure, and mapping the original data structure to a standard data structure according to the semantics of the data fields;
wherein mapping the original data structure to the standard data structure includes mapping a key value structure contained in response body information of a restful response;
acquiring standard entities and determining attributes contained in the standard entities, wherein the attributes are related to data fields in the standard data structure;
establishing associative relationships between the standard entities and generating a knowledge graph with the standard entities as nodes and the associative relationships between the standard entities as edges;
determining an address of the data source in the graphical database and establishing an associative relationship between an entity and the data source;
loading the knowledge graph and generating a query mode containing multiple queryable instances, wherein the queryable instances correspond to individual edges and/or nodes in the knowledge graph; and
storing the queryable instances in an internal memory or a cache to provide user access by data query;
wherein generating a query mode containing multiple queryable instances comprises:
determining the nodes and edges in the knowledge graph, and for any node, determining a root type of the node in a graph query language comprising GraphQL;
traversing the edges and converting the edges to link types in the graph query language according to the root type as well as defining a query depth indicating a minimum value of edges passed through when going from one node to another; and
traversing the nodes, and for any node, generating a queryable instance corresponding to the node according to the link types.