CN-121996660-A - Construction method and system based on knowledge graph of material database

CN121996660ACN 121996660 ACN121996660 ACN 121996660ACN-121996660-A

Abstract

The invention relates to the technical field of material informatics and discloses a construction method based on a knowledge graph of a material database, which comprises the following steps of S1, collecting multi-source basic data, carrying out standardization processing, identifying knowledge graph entities based on the standardized data, establishing entity attributes, defining explicit association relations among the entities and generating corresponding relation data, S2, carrying out association mining on the basis of the explicit association relations and the entity attributes to obtain hidden association relations for representing potential relations among the entities, and S3, fusing the entities, the explicit association relations and the hidden association relations to construct the knowledge graph of the material database. By taking a raw material-formula-process-performance-application scene as an entity skeleton, the relation among the entities is subjected to explicit modeling, and a computable association network is constructed, so that research and development personnel can realize cross-link retrieval, reasoning and positioning based on unified semantics, and the efficiency and consistency of formula development and performance analysis are improved.

Inventors

ZHANG XIN
JIANG ZHIYING
FANG ZHAOHUA

Assignees

房兆华
张鑫

Dates

Publication Date: 20260508
Application Date: 20260206

Claims (10)

1. The construction method based on the knowledge graph of the material database is characterized by being suitable for the formula design and the performance prediction of high polymer fine processing products, and comprises the following steps: S1, collecting multi-source basic data and carrying out standardization processing, identifying knowledge graph entities and establishing entity attributes based on the standardized data, defining explicit association relations among the entities and generating corresponding relation data; S2, carrying out association mining on the basis of the explicit association relationship and the entity attribute to obtain a hidden association relationship used for representing potential association between the entities; And S3, fusing the entity, the explicit association relationship and the hidden association relationship to construct a knowledge graph of the material database.
2. The method for constructing knowledge graph based on material database according to claim 1, wherein the polymer fine processed product is at least one selected from the group consisting of plastics, rubber, adhesives, inks, paints, oil products, surface treatment agents, metal processing liquids, cleaning agents and textile chemicals.
3. The method for constructing a knowledge-graph based on a material database according to claim 1, wherein the knowledge-graph entity comprises a raw material fruit body, a formula fruit body, a process fruit body, a main performance fruit body, an auxiliary performance fruit body and an application scene fruit body.
4. The method for building a knowledge graph based on a material database as claimed in claim 1, wherein the step S1 further comprises filtering, de-duplicating and aligning the attributes and the attribute values of the entities.
5. The method for constructing knowledge graph based on material database according to claim 3, wherein, The raw material sub-entity at least comprises chemical names, molecular structures, physical and chemical property attributes and compatibility attributes with other raw materials; the formula sub-entity at least comprises a component proportion attribute, a core performance index attribute, a data credibility grading attribute and a data source identification attribute; The process fruit body at least comprises a processing technological parameter attribute and an equipment requirement attribute, and an adaptation relation is established with the formula fruit body; the main performance sub-entity at least comprises a quantization index attribute and a detection standard attribute of the main performance; the auxiliary performance sub-entity at least comprises a quantization index attribute and a detection standard attribute of the auxiliary performance; The application scene sub-entity at least comprises an environmental condition attribute and a core performance requirement attribute.
6. The method for constructing a knowledge graph based on a material database according to claim 1, wherein the entity association relationship includes a composition relationship of raw materials-formulas, an adaptation relationship of formulas-processes, a mapping relationship of formulas-performances, a matching relationship of performances-application scenes, and a synergy/antagonism relationship of raw materials-raw materials.
7. The method for constructing a knowledge graph based on a material database according to claim 1, wherein the mining hidden association relationship comprises a similar formula recommendation relationship, a substitute raw material matching relationship and a process parameter optimization direction relationship.
8. The method for constructing a knowledge graph based on a material database according to claim 7, wherein the similar formula recommendation relationship is obtained by an association rule mining algorithm; the matching relation of the alternative raw materials is obtained by calculating the similarity of the raw materials through a graph embedding algorithm; The process parameter optimization direction relationship is obtained based on potential correlations between historical data mining process parameters and performance.
9. The method for building knowledge graph based on material database according to claim 3, wherein in step S3, the method for building knowledge graph based on material database comprises constructing knowledge graph by using the relationship data between entities obtained in step S1 and step S2, using each entity as node in the knowledge graph, using the association relationship between the sub-entities as the side of the knowledge graph, and storing the knowledge represented by the raw material sub-entity, the recipe sub-entity, the process sub-entity, the main performance sub-entity, the auxiliary performance sub-entity, the application scene sub-entity and their relationship side by using extensible markup language XML to obtain the knowledge graph based on material database.
10. A construction system based on a knowledge graph of a material database, which is applied to the construction method based on the knowledge graph of the material database as claimed in any one of claims 1 to 9, and is characterized by comprising: The data acquisition module is used for acquiring multi-source basic data and carrying out standardized processing to obtain a data source for constructing a knowledge graph; The entity identification module is used for carrying out entity identification on the data source and establishing entity attributes; the entity association module is used for defining explicit association relations among entities and mining hidden association relations; the knowledge graph construction module is used for fusing the entity, the explicit association relationship and the hidden association relationship, and constructing and storing the knowledge graph of the material database.

Description

Construction method and system based on knowledge graph of material database Technical Field The invention relates to the technical field of material informatics, in particular to a method and a system for constructing a knowledge graph based on a material database. Background The knowledge graph is a graph structure database and is characterized in that data of different sources, different types and different structures are fused, and after the ontology is extracted, the ontology is associated into a graph through the relationship between the ontologies. The essence is that the data in the field is systemized and relational, and then the visualization is realized in a graphical mode. Knowledge maps can be used to present knowledge resources, mine, analyze, build and display associations between knowledge. At present, the application of the knowledge graph is mainly concentrated in the fields of search engines and intelligent questions and answers, and the application of the knowledge graph in the professional field is less. In the process of material research and industrialization, the material data sources are scattered, multiple dimensions such as the own data, the public channel data, the client enterprise data, the document unstructured data and the like in enterprises are covered, the collection mode, the format standard and the information integrity of the data from different sources are huge, the traditional relational database is limited by the table structure and the fixed mode, the capability of expressing complex attributes and dynamic association is insufficient, the inherent relation among the material components, the structure, the process and the performance cannot be fully mined, and the improvement of the material research and development efficiency and the reduction of the cost are restricted. Therefore, the knowledge graph oriented to the material field is constructed, the multi-source heterogeneous data are integrated, and the relation between the data is deeply mined. Disclosure of Invention In order to solve the problem of lack of knowledge graphs of a material database in the prior art, the invention discloses a construction method and a construction system based on knowledge graphs of the material database, which are used for carrying out explicit modeling on relationships such as raw material-formula composition, formula-process adaptation, formula-performance mapping, performance-scene matching, raw material-raw material synergy/antagonism and the like by taking a raw material-formula-process-performance-application scene as an entity skeleton, so as to construct a computable association network, thereby enabling research and development personnel to realize cross-link retrieval, reasoning and positioning based on unified semantics and improving the efficiency and consistency of formula development and performance analysis. In order to achieve the above purpose, the invention adopts the following technical scheme: the construction method based on the knowledge graph of the material database is suitable for the formula design and the performance prediction of high polymer fine processing products, and comprises the following steps: S1, collecting multi-source basic data and carrying out standardization processing, identifying knowledge graph entities and establishing entity attributes based on the standardized data, defining explicit association relations among the entities and generating corresponding relation data; S2, carrying out association mining on the basis of the explicit association relationship and the entity attribute to obtain a hidden association relationship used for representing potential association between the entities; And S3, fusing the entity, the explicit association relationship and the hidden association relationship to construct a knowledge graph of the material database. Optionally, the polymer fine processing product is at least one selected from plastics, rubber, adhesives, printing inks, paints, oil products, surface treatment agents, metal processing fluids, cleaning agents and textile chemicals. Optionally, the knowledge-graph entity includes a raw material fruit body, a formula fruit body, a process fruit body, a main performance fruit body, an auxiliary performance fruit body and an application scene fruit body. Optionally, step S1 further comprises filtering, de-duplicating and aligning the attributes and attribute values of the entity. Optionally, the raw material fruiting body at least comprises chemical names, molecular structures, physical and chemical property attributes and compatibility attributes with other raw materials; the formula sub-entity at least comprises a component proportion attribute, a core performance index attribute, a data credibility grading attribute and a data source identification attribute; The process fruit body at least comprises a processing technological parameter attribute and an equipment requirement attribute, an