CN114610898A - Method and system for constructing supply chain operation knowledge graph - Google Patents
Method and system for constructing supply chain operation knowledge graph Download PDFInfo
- Publication number
- CN114610898A CN114610898A CN202210222047.7A CN202210222047A CN114610898A CN 114610898 A CN114610898 A CN 114610898A CN 202210222047 A CN202210222047 A CN 202210222047A CN 114610898 A CN114610898 A CN 114610898A
- Authority
- CN
- China
- Prior art keywords
- knowledge
- data
- supply chain
- ontology
- chain operation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/242—Query formulation
- G06F16/243—Natural language query formulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/80—Information retrieval; Database structures therefor; File system structures therefor of semi-structured data, e.g. markup language structured data such as SGML, XML or HTML
- G06F16/83—Querying
- G06F16/835—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/80—Information retrieval; Database structures therefor; File system structures therefor of semi-structured data, e.g. markup language structured data such as SGML, XML or HTML
- G06F16/84—Mapping; Conversion
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to a method and a system for constructing a supply chain operation knowledge graph. The method comprises the following steps: acquiring data in the supply chain operation process; extracting knowledge from the acquired data in the supply chain operation process; constructing an ontology base according to the extracted knowledge, and performing knowledge fusion on the multi-source data of the supply chain based on the ontology base; carrying out knowledge processing on the data after knowledge fusion so as to ensure the warehousing quality of knowledge; storing the data after knowledge processing by adopting a graph database to form a supply chain operation knowledge graph; the method is characterized in that knowledge service application is realized on the basis of a supply chain operation knowledge map, and comprises a knowledge semantic retrieval service facing to the supply chain operation field and a knowledge accurate pushing service based on user portrait. The invention constructs the knowledge graph fusing multiple data sources, can realize the rapid query of data in different fields, can support the semantic retrieval function, meets the requirements of users on cross-stage knowledge correlation retrieval, improves the cooperative work efficiency and improves the level of auxiliary decision.
Description
Technical Field
The invention belongs to the technical field of information, and particularly relates to a method and a system for constructing a supply chain operation knowledge graph.
Background
Under the background of big data era, with the appearance of mass data and the fusion and cross application of multiple data sources, the traditional data management mode and the query mode are restricted to a certain extent. In recent years, a Knowledge Graph (Knowledge Graph) has been used as a new Knowledge representation method and data management model in important applications in fields such as natural language processing, question answering, information retrieval, and the like. The knowledge graph is a structured semantic knowledge base and is used for describing concepts in the physical world and the mutual relations thereof in a symbolic form; the basic composition unit is an entity-relation-entity triple, entities and related attribute-value pairs thereof, and the entities are mutually connected through relations to form a network knowledge structure.
With the release of google knowledge maps, the construction and application research of knowledge maps has attracted extensive attention in academia and industry. In China, the construction and research of knowledge maps are started, and a plurality of important research results are correspondingly obtained. Such as: knowing the cube and hundredth heart of dog; the GDM laboratory of the university of Compound Dan designs a Chinese knowledge map facing the book reading field.
Knowledge map (Knowledge Graph) is a series of different graphs displaying Knowledge development process and structure relationship in the book intelligence field, describing Knowledge resources and carriers thereof by using visualization technology, mining, analyzing, constructing, drawing and displaying Knowledge and mutual relation between Knowledge resources and Knowledge carriers.
The knowledge graph is a modern theory which achieves the aim of multi-discipline fusion by combining theories and methods of applying subjects such as mathematics, graphics, information visualization technology, information science and the like with methods such as metrology introduction analysis, co-occurrence analysis and the like and utilizing a visualized graph to vividly display the core structure, development history, frontier field and overall knowledge framework of the subjects.
The prior knowledge graph of the industry field is usually constructed manually, a uniform construction method is lacked, and the knowledge base aims at the specific industry field, so the description range is extremely limited. Aiming at the problems, the method is provided for fusing knowledge bases in different fields into a knowledge graph, aims to construct a multi-data fusion knowledge graph with consistent semantics and structure, and realizes the query and display of knowledge in different fields, thereby improving the data query efficiency.
The supply chain is a logistics network consisting of suppliers, manufacturers, warehouses, distribution centers, and distributors, among others. The same enterprise may constitute different constituent nodes of the network, but more often different enterprises constitute different nodes in the network. For example, in a supply chain, the same enterprise may occupy locations at both the manufacturer, the warehouse node, the distribution center node, and so on. In supply chains with finer division of work and higher professional requirements, the different nodes are essentially composed of different enterprises. The raw materials, work in process inventory, and product flows between the various member units of the supply chain constitute the flow of goods in the supply chain. Statistics indicate that enterprise supply chains can cost up to 25% of the operating costs of an enterprise.
The traditional supply chain mainly solves the related problems by means of personal experience judgment. Many times, the problem of the supply chain is a complex problem formed by multiple factors, the simple expert experience cannot give accurate judgment, and the technology of constructing the uniform supply chain knowledge graph is a main mode for solving the problem.
Disclosure of Invention
The invention provides a method and a system for constructing a supply chain operation knowledge graph aiming at related problems in the supply chain operation.
The technical scheme adopted by the invention is as follows:
a supply chain operation knowledge graph construction method comprises the following steps:
acquiring data in the supply chain operation process, wherein the data comprises structured data, semi-structured data and unstructured data;
performing knowledge extraction on the acquired data in the supply chain operation process, wherein the knowledge extraction comprises entity extraction, relationship extraction and attribute extraction;
constructing an ontology base according to the extracted knowledge, and performing knowledge fusion on the multi-source data of the supply chain based on the ontology base;
carrying out knowledge processing on the data after knowledge fusion so as to ensure the warehousing quality of knowledge;
and storing the data after knowledge processing by adopting a database to form a supply chain operation knowledge map.
Further, the building of an ontology base according to the extracted knowledge and the knowledge fusion of the multi-source data of the supply chain based on the ontology base comprise: firstly, a domain ontology base is constructed, then the domain ontology bases are fused to form a global ontology base, and then the knowledge bases of all the domains are subjected to entity alignment and entity linkage.
Further, the building of the domain ontology library includes:
extracting a relation mode from a relation database of the field;
mapping the relationship schema to a domain ontology model using a transformation rule, the transformation rule comprising: converting the table name in the relation mode into a concept name in the ontology; converting the relationship between the tables into the relationship between the concepts in the ontology; converting the field names in the relational mode into attribute names of the ontology;
and evaluating and checking the domain ontology model, and establishing an ontology library in the domain after the evaluation and the check are passed.
Further, the fusing the domain ontology library to form a global ontology library includes:
firstly, detecting ontologies in different fields by adopting a similarity detection rule, wherein the similarity detection rule comprises semantic similarity detection, concept similarity detection, attribute similarity detection and data format similarity detection, so that the same or similar concepts in different fields are unified;
secondly, eliminating ambiguity of the same or similar concepts by adopting a conflict resolution rule, and eliminating redundant and wrong concepts;
and finally, mapping the rest domain ontologies to a global ontology library through conflict resolution and entity disambiguation processing.
Further, the entity link is realized by adopting a constraint vector-based embedding conversion algorithm, and the method comprises the following steps: projecting the entities and the relations in the knowledge graph to a low-dimensional vector space in an embedding mode; calculating head and tail entities and a loss function value of a relation in the vector space through vector translation conversion operation in the vector space, and realizing relation linkage of the head and tail entities; and adding relation semantic constraint conditions to enable the predicted relation between the entities to meet the semantic type of the relation.
Furthermore, the knowledge service application is realized on the basis of the supply chain operation knowledge graph, and comprises a knowledge semantic retrieval service facing the supply chain operation field and a precise knowledge pushing service based on user portrait.
Further, the knowledge semantic retrieval service facing the supply chain operation field comprises the following steps:
intention recognition: through a natural language processing technology, intention classification is carried out on retrieval and query contents of a user, and the retrieval and query contents are positioned to a body position corresponding to a supply chain operation knowledge graph according to predicted intention types;
information extraction: through natural language processing technology, the identification of entities and the relation thereof in the retrieval query content of a user is realized;
semantic matching: and according to the positioned body and the entities and the relations extracted from the user retrieval query content, efficiently matching the entities and the relations corresponding to the entities in the supply chain operation knowledge graph through a natural language processing technology, and feeding back the entities and the relations to the user in an answer mode according to the matched entities.
Further, the accurate knowledge pushing service based on the user portrait comprises the following steps:
acquiring user portrait data comprising a user background and a retrieval habit, and displaying historical consulting traces of a user in an all-around and three-dimensional manner;
based on user portrait data, the push service of accurate information is realized; the push service not only realizes that the same field returns a plurality of retrieval results of the same type, but also returns other related business information according to semantic analysis, and realizes that the user is guided from fuzzy retrieval to accurate positioning from the information correlation angle.
A supply chain operation knowledge graph construction system adopting the method comprises the following steps:
the data acquisition module is used for acquiring data in the supply chain operation process, wherein the data comprises structured data, semi-structured data and unstructured data;
the knowledge extraction module is used for extracting the knowledge of the acquired data in the supply chain operation process, and comprises entity extraction, relationship extraction and attribute extraction;
the knowledge fusion module is used for constructing an ontology base according to the extracted knowledge and carrying out knowledge fusion on the multi-source data of the supply chain based on the ontology base;
the knowledge processing module is used for carrying out knowledge processing on the data subjected to knowledge fusion so as to ensure the warehousing quality of knowledge;
and the knowledge storage module is used for storing the data after the knowledge processing by adopting a database to form a supply chain operation knowledge map.
Furthermore, the system also comprises a knowledge service application module which is used for realizing knowledge service application based on the supply chain operation knowledge map, and the knowledge service application module comprises a knowledge semantic retrieval service facing the supply chain operation field and a knowledge accurate pushing service based on the user portrait.
The invention has the following beneficial effects:
1) the invention fully utilizes knowledge in different fields, constructs the knowledge graph of multiple data sources under the condition of fusing multiple data sources, and can realize the rapid query of data in different fields.
2) The supply chain operation knowledge graph can support the semantic retrieval function, meet the requirements of users on cross-stage knowledge association retrieval, improve the cooperative work efficiency and improve the level of auxiliary decision.
3) The accurate pushing based on the user portrait can realize accurate information pushing service according to the historical retrieval habit of the user, not only can realize that the same field returns a plurality of retrieval results of the same type, but also returns other related business information according to semantic analysis, and guides the user from fuzzy retrieval to accurate positioning in the angle of information association, thereby solving the problems of single returned result, difficult description of a retrieval object and the like, providing support for cross-business retrieval of the user, and improving the working efficiency.
Drawings
Fig. 1 is a supply chain operational knowledge graph general technology roadmap.
FIG. 2 is a knowledge graph construction flow diagram.
FIG. 3 is a schematic diagram of a process for constructing a domain ontology library from relational data.
FIG. 4 is a diagram of a global ontology library building process.
FIG. 5 is a diagram of a different repository entity alignment process.
Detailed Description
The present invention will be described in further detail below with reference to specific examples and the accompanying drawings.
1. General technical route
The technical route of the construction and application of the supply chain knowledge graph is shown in figure 1 and comprises a data source, a knowledge graph construction, a knowledge service system and a knowledge service application. The data source is the basis for supporting the application of the knowledge service, and the data of the plate is structured, unstructured and semi-structured data generated in three stages of planning design, construction and production, operation and maintenance in Wangming Ridge and a Mediterranean coal factory. The knowledge graph construction realizes semantic integration of a multi-source heterogeneous database by mainly fusing data generated in supply chain operation, and forms a uniform supply chain operation knowledge graph. The knowledge service system is based on the supply chain operation knowledge map, and supports supply chain operation-oriented knowledge service application, wherein the support application comprises knowledge semantic retrieval, knowledge accurate pushing and the like.
2. Supply chain operation knowledge graph construction
The supply chain operation knowledge graph construction process is shown in fig. 2, and a structured semantic knowledge base is formed.
The supply chain operation knowledge graph is based on data of a data resource pool, and is constructed by knowledge acquisition, knowledge fusion, knowledge processing and knowledge storage, so that data support is provided for knowledge graph analysis. As shown in fig. 2, the main implementation steps are as follows:
1) and (6) acquiring data. Various financial, management, logistics, production, design, research and development and other aspects of data in the operation process are collected and collected to form a data base constructed by the knowledge graph of the coal preparation plant, wherein the data comprises structured data, semi-structured data and non-structured data.
2) And (5) extracting knowledge. The purpose of the stage is to automatically extract information from a heterogeneous data source to obtain a structured knowledge unit, and extract related information constructed by a map from processed BIM model data in a data resource pool of an MIM big data platform through an interface service. The key technology of knowledge extraction comprises entity extraction, relationship extraction and attribute extraction, and is an important link for map construction.
3) And constructing a supply chain operation knowledge body by combining the extracted knowledge and the domain category thereof with expert domain knowledge and experience.
4) And (4) knowledge fusion. The integration and complementation of knowledge in a coal plant design knowledge base, a construction knowledge base and an operation and maintenance knowledge base are realized through technologies such as coreference resolution, entity disambiguation and entity linking. Aiming at the same entity, the complete description of the entity is obtained from different stages of coal plant operation, so that entity association is established for various data.
5) And (5) knowledge processing. Data after knowledge fusion cannot be directly stored, and the storage quality of knowledge is ensured through a certain degree of quality evaluation, so that a guarantee is provided for knowledge service.
6) And (4) storing knowledge. And storing the data after knowledge processing by using a Neo4j graph database, so that the data can be conveniently inquired in the form of a graph data structure, and data support is provided for upper-layer application of the knowledge graph.
3 supply chain multi-source data fusion
In order to fully utilize knowledge in different fields and realize rapid query of data in different fields, the invention constructs the knowledge graph of multiple data sources under the condition of fusing multiple data sources. The method comprises the steps of firstly constructing ontology bases in different fields, then mapping the ontology bases in the different fields to form a global ontology base, and then performing entity alignment and entity linkage on knowledge bases in all the fields to enrich and expand a constructed multi-data fusion knowledge graph.
3.1) data Source
The ontology database data sources used to construct the knowledge graph can be derived from structured data, semi-structured data, and unstructured data, as well as some existing general knowledge graph libraries, and the like.
1) The data is structured. It mainly refers to tables, excel tables and other data with structures in relational databases.
2) Semi-structured data. The method mainly refers to the intermediate between structured data and unstructured data, and common related web pages such as XML and HTML belong to semi-structured data. The semi-structured data mainly comes from Wikipedia, Baidu encyclopedia and the like.
3) There is no structured data. It mainly refers to plain text data, images, and sound data.
3.2) ontology library construction
Ontology (ontology) is a specification for modeling concepts, is an abstract model for describing the objective world, and gives a clear definition to concepts and their connections in a formalized way. The ontology defines data patterns in the knowledge graph, and therefore, the result of ontology construction research can assist the construction of the knowledge graph to a great extent. The method of ontology construction is different for different application fields and different requirements. The invention uses Ontology language OWL (Web Ontology language) to construct corresponding domain Ontology base from a plurality of data sources, and then forms a global Ontology base through mapping.
3.2.1) Domain ontology library construction
The main data sources for constructing the domain ontology library of the embodiment are an environment monitoring database, an air pollution detection database and a medical health database. In addition, website data and the like in the related fields are also used. The following description focuses on the process of obtaining the domain ontology base from the relational database, as shown in fig. 3.
Firstly, a relational database in the domain is created for a specific domain, and the database contains detailed information of expression methods and specific applications in the domain, so that a relational pattern can be extracted from the relational database in the domain, information and field information in the table in the relational database are analyzed, and a corresponding conceptual model is established.
Secondly, since the relationship schema includes the relationship between tables and fields and the relationship between tables and tables, the ontology library includes the relationship between concepts and attributes. Therefore, the relationship schema is mapped to the ontology model using a certain rule. A series of conversion rules are designed, such as: converting the table name in the relation mode into a concept name in the ontology; the relationship between the tables is converted into the relationship between concepts in the ontology; converting the field names in the relational schema into the attribute names of the ontology, and the like. Through the conversion rule, the domain ontology model can be obtained.
And finally, evaluating and verifying the domain ontology model. The key points of the part are that the constructed domain ontology model is checked to see whether the constructed domain ontology model meets the construction principle of an ontology library, whether terms in the ontology model are correct, whether concepts and relations in the ontology model are complete and the like. After the ontology model is evaluated, an ontology library in the field can be established.
Relational databases may have complete data schemas, including complete table structures and integrity constraints. Thus, relationship names in a database may be converted into concepts in an ontology, and partial field names may be converted into attributes in the ontology, examples are shown in table 1 and table 2, where table 1 converts relationship names into OWL language of ontology concepts, and table 2 converts field names into OWL language of attribute names.
TABLE 1 TABLE 2
In addition, to expand and perfect the domain ontology base, non-relational data needs to be collected and populated. In this embodiment, the semi-structured data in the industrial field is subjected to structured processing, corresponding knowledge is acquired from a corresponding encyclopedic website through a web crawler technology, the semi-structured data is converted into structured data, and finally, the relationship data is converted into a rule of an ontology for conversion.
3.2.2) Global ontology library construction
In order to facilitate the construction of a multi-data fusion knowledge graph, ontology bases in multiple fields need to be fused to construct a global ontology base. The process is shown in fig. 4. On the basis of the constructed domain ontology base, the ontology bases of a plurality of domains are fused together to form a global ontology base through rules such as similarity detection and conflict resolution. The method comprises the following steps:
firstly, knowledge fusion is carried out on ontology bases in different fields, and for the same or similar concepts and attributes and the like, similarity detection rules are adopted to detect the ontologies in the different fields. Such as: semantic similarity detection, concept similarity detection, attribute similarity detection, data format similarity detection, and the like. Through the similarity detection, the same or similar ontologies in different fields can be unified, but the conflict between the same or similar ontologies cannot be solved.
Secondly, conflict resolution rules are adopted to resolve the similar concepts or attributes existing above. The ambiguity of the concept can be eliminated through the conflict resolution rule, and the redundant and wrong concepts can be eliminated, so that the quality of the global ontology base is ensured. The method mainly eliminates the similar or similar concepts or attributes in the above to unify the concepts or attributes and combine the concepts or attributes into a global ontology.
And finally, mapping the rest domain ontologies to a global ontology library through conflict resolution, entity disambiguation and other processing, and combining the domain ontologies with the processed domain ontology libraries to realize the construction of the global ontology.
3.3) physical alignment
Entity alignment, also known as entity matching or entity resolution, is the process of determining whether 2 entities in the same or different datasets point to the same object in the real world. The purpose of entity alignment is: the method comprises the steps of finding out entities which have different entity names in different knowledge bases but represent the same thing in the real world, combining the entities, identifying the entities by using unique identifiers, and finally adding the entities into a corresponding knowledge graph. Entity alignment procedures for different knowledge bases. The method comprises the steps of giving different knowledge bases, carrying out entity matching algorithm calculation under the action of priori alignment data, adjustment parameters and relevant external data, and finally obtaining an alignment result between entities. The different knowledge base entity alignment process is shown in fig. 5.
Although entity disambiguation is performed on entities of ontology bases in different fields when a global ontology base is constructed, entity alignment is used for enriching and expanding a knowledge graph, and the knowledge graph is filled by extracting the entities and the relationships among the entities from the existing general knowledge graph and related data thereof by using an entity alignment method. At present, there are many algorithms related to entity alignment, and the following are commonly used: an entity alignment method based on a traditional probability model, an entity alignment method based on machine learning, an entity alignment method based on similarity propagation, an entity alignment method based on an LDA model, an entity alignment method based on a CRF model, an entity alignment method based on a Markov logic network and the like.
The invention adopts a similarity propagation-based entity alignment method, the algorithm takes the entity alignment problem as an optimization problem of a global matching scoring objective function for modeling, belongs to a binary classification problem, and can obtain an approximate solution of the problem by a greedy optimization algorithm. The basic process is as follows: (1) extracting entities in the open link data and encyclopedic data in the industry field to obtain a synonymous name set of the entities; (2) matching the entities with the entities in the constructed knowledge graph by an entity alignment method, and taking the result as a candidate entity set for entity combination; (3) and comparing the entities in the candidate entity sets, and combining the entities into one entity if the entities have the same upper-layer concept.
3.4) entity linking
Entity linking refers to an operation of linking an entity object extracted from a text to a corresponding correct entity object in a knowledge graph. The entity link prediction means that the relation among the missing entities is predicted in a given knowledge graph, so that the knowledge graph is enriched and expanded. The basic idea is to select a group of candidate entity objects from the knowledge graph or other related text data according to the head (tail) entity and the relation of a given triple, then calculate the correct tail (head) entity through an entity link prediction algorithm, and add the obtained triple to the corresponding knowledge graph. At the present stage, related knowledge graph entity link prediction algorithms are more. Commonly used are: vector embedding based conversion algorithm, tensor decomposition based algorithm, path based reasoning algorithm, combined text reasoning algorithm and the like.
On the basis of the prior art, the invention provides a constraint vector embedding conversion algorithm, and a better entity link prediction result is obtained. The basic idea is as follows: and projecting the entities and the relations in the knowledge map to a low-dimensional vector space in an embedding (embedding) mode, and calculating loss function values of the head and tail entities and the relations in the vector space through vector translation conversion operation in the vector space to realize the relation linkage of the head and tail entities. Based on the constraint embedded conversion algorithm, on the basis of the original vector embedded conversion algorithm, a relationship semantic constraint condition is added, so that the predicted relationship between entities needs to meet the semantic type of the relationship. Such as: for a relationship "birth to", the head entity is typically a person or animal, and the tail entity is typically a time or place.
4 supply chain operation knowledge graph application
4.1) knowledge semantic retrieval service facing supply chain operation field
The fragmentation phenomenon of the supply chain operation data is prominent, the multi-source heterogeneous data generated by each system is lack of semantic association, the cross-stage knowledge association retrieval requirement is restrained, and the data utilization rate is low. Aiming at the problem, the invention provides a semantic retrieval service facing the supply chain operation field through the constructed supply chain operation knowledge graph so as to meet the requirement of coal preparation plant workers on cross-phase associated knowledge acquisition in different stages of design, implementation and operation. The semantic retrieval service oriented to the supply chain operation field is beneficial to improving the knowledge retrieval efficiency of related business personnel and improving the work quality and efficiency.
The supply chain has independent operation of different business units of different enterprises and overall deployment based on the supply chain in operation and operation. Each individual independently operated has a certain self-regulation, and unified cooperation is lacked among individuals in overall coordination, which is also a main reason for various problems caused in the initial stage of supply chain operation. While most of these problems are common and common, the same problem may have resulted in certain solutions in different supply chains and can guide the solution of other supply chain problems. Based on this situation, it becomes important to construct a supply chain problem library and a solution retrieval system based on the knowledge graph.
Knowledge-graph is a formal representation of the knowledge of the objective world, representing entities, events and relationships between objects by means of triples. Compared with the traditional search technology based on the key words, the semantic search maps the key words input by the user search into the concept and entity of the objective world in the knowledge map under the knowledge support of the knowledge map, and the search result directly displays the structured information content meeting the user requirement. The semantic search can accurately capture the search intention of the user by using the knowledge graph, and answers meeting the search intention of the user are directly given by using the knowledge graph, not only the content containing the keywords.
In the invention, the supply chain operation knowledge graph covers unstructured text data, picture data, video monitoring data, standard data, internet of things collected data, other structured data and the like generated in three stages. In the process of knowledge graph construction, semantic integration of multi-source heterogeneous data from different services is realized by applying technologies such as information extraction, space-time association, entity linkage, ontology matching and the like. Knowledge showing strong correlation in time and space is connected together, and knowledge showing semantic correlation in different stages is connected together. The supply chain operation knowledge graph can support the semantic retrieval function, meet the requirements of users on cross-stage knowledge association retrieval, improve the cooperative work efficiency and improve the auxiliary decision level.
The main implementation steps of the semantic retrieval function based on the supply chain operation knowledge graph are as follows:
1) and identifying the intention. And performing intention classification on the retrieval query content of the user through a natural language processing technology. And positioning to an ontology position corresponding to the supply chain operation knowledge graph according to the predicted intention type.
2) And (5) information extraction. The method mainly realizes the identification of entities and the relation thereof in the retrieval query content of a user through a natural language processing technology.
3) And (6) semantic matching. And according to the positioned body and the entity and the relation extracted from the user retrieval query content, efficiently matching the entity and the relation corresponding to the supply chain operation knowledge graph through a natural language processing technology, and feeding back the entity and the relation to the user in an answer mode according to the matched entity.
The three steps correspond to intention recognition, information extraction and semantic matching in the knowledge service system of fig. 1. The association analysis in the "knowledge service system" of fig. 1 refers to (); knowledge classification management means (); historical behavioral analysis refers to ().
Through a semantic retrieval mode, the supply chain production end collects the full life cycle information scattered in each system, and constructs a knowledge base of categories such as equipment, processes, plants, trends and the like through knowledge graph classification. When the coal preparation plant carries out equipment installation, maintenance and plant area reconstruction and expansion, the knowledge base becomes the main scientific basis for carrying out the work.
Through a semantic retrieval mode, the supply chain collects the full life cycle information scattered in each system, and a knowledge base of categories such as equipment, processes, plants, trends and the like is constructed through knowledge graph classification. When equipment installation, maintenance and repair are carried out in a supply chain production chain and a factory area is reconstructed, the knowledge base becomes a main scientific basis for carrying out development work.
4.2) accurate push service of knowledge based on user portrait
The most important task in the process of realizing the digital transformation of the supply chain is to break a data island of each service stage and realize the associated management of each service unit data of the supply chain. The invention realizes data fusion through the knowledge graph, integrates various data and is convenient for unified management and analysis of users.
There are numerous intellectual materials of the supply chain that can be applied to different stages of organization, production, marketing, funding, quality control, logistics, and after-sales services of the supply chain. However, the original data island mode scatters the data everywhere, and does not form good series connection and combing, and the data can not be obtained effectively quickly when the data support is needed.
However, the following problems are often faced in the semantic retrieval process based on knowledge graph:
1. the retrieval result is not necessarily the content desired by the user. Due to the fact that data fusion of all service stages is achieved, the data range is widened to other fields, the returned result may be related data of other fields, and users can achieve accurate information query only through multiple times of retrieval.
2. The user cannot accurately describe the retrieved object. The semantic retrieval process is to analyze the semantics of the input field, then to inquire the information most similar to the field from the library and return, and in the cross query process of the business data, if the input field is inconsistent with the retrieval object, the result desired by the user cannot be accurately returned; in order to improve the working efficiency of three major business stages of a coal preparation plant, the above problems need to be solved urgently.
Accurate pushing based on user portrait aims at historical retrieval habits of users, and therefore accurate information pushing service is achieved. The user accurate portrait data comprises basic information such as user backgrounds, retrieval habits and the like of three major business stages of coal preparation plant design, implementation and operation and maintenance, can show the historical consulting traces of the user in an all-round and three-dimensional manner, and provides objective basis for analyzing the consulting preferences of the user by the system. The user background refers to the service attribute of the user, and the retrieval habit refers to the preference of the user on a retrieval object, including data category and data content.
The pushing service can not only return a plurality of retrieval results of the same type in the same field, but also return other related business information according to semantic analysis, and guide the user from fuzzy retrieval to accurate positioning from the angle of information association, thereby solving the problems of single return result, difficult description of a retrieval object and the like, providing support for cross-business retrieval of the user, and further improving the working efficiency of the coal preparation plant in three business stages of design, construction and operation and maintenance.
The supply chain needs a great deal of professional knowledge to support in the aspects of industrial chain operation, supplier selection, logistics cooperation, production process management, quality inspection management, financial chain credit and fund payment, industrial chain brand construction and the like. In the process, the accurate knowledge pushing service based on the user portrait carries out preposed collection and arrangement according to the knowledge requirements possibly appearing in each stage, and actively pushes the knowledge to the coal preparation plant management, operation, maintenance, safety supervision and post personnel training. Therefore, the solution can be quickly generated when the preposed training of different personnel and related problems occur, and the support of various knowledge on the operation aspect of the coal preparation plant is ensured. The accurate knowledge pushing service oriented to the supply chain operation field is beneficial to providing individualized knowledge association for related business personnel and improving the working efficiency.
The particular embodiments of the present invention disclosed above are illustrative only and are not intended to be limiting, since various alternatives, modifications, and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The invention should not be limited to the disclosure of the embodiments in the present specification, but the scope of the invention is defined by the appended claims.
Claims (10)
1. A supply chain operation knowledge graph construction method is characterized by comprising the following steps:
acquiring data in the supply chain operation process, wherein the data comprises structured data, semi-structured data and unstructured data;
performing knowledge extraction on the acquired data in the supply chain operation process, wherein the knowledge extraction comprises entity extraction, relationship extraction and attribute extraction;
constructing an ontology base according to the extracted knowledge, and performing knowledge fusion on the multi-source data of the supply chain based on the ontology base;
carrying out knowledge processing on the data after knowledge fusion so as to ensure the warehousing quality of knowledge;
and storing the data after knowledge processing by adopting a database to form a supply chain operation knowledge map.
2. The method of claim 1, wherein the building an ontology base according to the extracted knowledge and performing knowledge fusion on the supply chain multi-source data based on the ontology base comprises: firstly, a domain ontology base is constructed, then the domain ontology bases are fused to form a global ontology base, and then entity alignment and entity linking are carried out on knowledge bases of all domains.
3. The method of claim 2, wherein the building a domain ontology library comprises:
extracting a relation mode from a relation database of the field;
mapping the relationship schema to a domain ontology model using a transformation rule, the transformation rule comprising: converting the table name in the relation mode into a concept name in the ontology; converting the relationship between the tables into the relationship between the concepts in the ontology; converting the field names in the relational mode into attribute names of the ontology;
and evaluating and checking the domain ontology model, and establishing an ontology library in the domain after the evaluation and the check are passed.
4. The method of claim 2, wherein fusing the domain ontology library to form a global ontology library comprises:
firstly, detecting ontologies in different fields by adopting a similarity detection rule, wherein the similarity detection rule comprises semantic similarity detection, concept similarity detection, attribute similarity detection and data format similarity detection, so that the same or similar concepts in different fields are unified;
secondly, eliminating ambiguity of the same or similar concepts by adopting a conflict resolution rule, and eliminating redundant and wrong concepts;
and finally, mapping the rest domain ontologies to a global ontology library through conflict resolution and entity disambiguation processing.
5. The method of claim 2, wherein the entity link is implemented by using a constraint vector embedding based transformation algorithm, comprising the steps of: projecting the entities and the relations in the knowledge graph to a low-dimensional vector space in an embedding mode; calculating head and tail entities and loss function values of the relation in the vector space through vector translation conversion operation in the vector space, and realizing the relation linkage of the head and tail entities; and adding relation semantic constraint conditions to enable the predicted relation between the entities to meet the semantic type of the relation.
6. The method of claim 1, wherein knowledge service applications are implemented based on the supply chain operation knowledge graph, and comprise a knowledge semantic retrieval service facing the supply chain operation field and a knowledge precision push service based on user portrait.
7. The method according to claim 6, wherein the knowledge semantic retrieval service facing the supply chain operation domain comprises the following steps:
intention recognition: through a natural language processing technology, intention classification is carried out on retrieval query contents of a user, and the retrieval query contents are positioned to an ontology position corresponding to a supply chain operation knowledge graph according to predicted intention types;
information extraction: through natural language processing technology, the identification of entities and the relation thereof in the retrieval query content of a user is realized;
semantic matching: and according to the positioned body and the entities and the relations extracted from the user retrieval query content, efficiently matching the entities and the relations corresponding to the entities in the supply chain operation knowledge graph through a natural language processing technology, and feeding back the entities and the relations to the user in an answer mode according to the matched entities.
8. The method of claim 6, wherein the user-portrait-based knowledge precision push service comprises the steps of:
acquiring user portrait data which comprises a user background and a retrieval habit, and displaying history look-up traces of a user in an all-dimensional and three-dimensional manner;
based on user portrait data, the push service of accurate information is realized; the push service not only realizes that the same field returns a plurality of retrieval results of the same type, but also returns other related business information according to semantic analysis, and realizes that the user is guided from fuzzy retrieval to accurate positioning from the information correlation angle.
9. A supply chain operation knowledge graph construction system adopting the method of any one of claims 1 to 8, which is characterized by comprising the following steps:
the data acquisition module is used for acquiring data in the supply chain operation process, wherein the data comprises structured data, semi-structured data and unstructured data;
the knowledge extraction module is used for extracting the knowledge of the acquired data in the supply chain operation process, and comprises entity extraction, relationship extraction and attribute extraction;
the knowledge fusion module is used for constructing an ontology base according to the extracted knowledge and carrying out knowledge fusion on the multi-source data of the supply chain based on the ontology base;
the knowledge processing module is used for carrying out knowledge processing on the data subjected to knowledge fusion so as to ensure the warehousing quality of knowledge;
and the knowledge storage module is used for storing the data after the knowledge processing by adopting a database to form a supply chain operation knowledge map.
10. The system of claim 9, further comprising a knowledge service application module, configured to implement knowledge service applications based on the supply chain operation knowledge graph, including a supply chain operation domain oriented knowledge semantic retrieval service, and a user portrait-based knowledge precision pushing service.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210222047.7A CN114610898A (en) | 2022-03-09 | 2022-03-09 | Method and system for constructing supply chain operation knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210222047.7A CN114610898A (en) | 2022-03-09 | 2022-03-09 | Method and system for constructing supply chain operation knowledge graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114610898A true CN114610898A (en) | 2022-06-10 |
Family
ID=81861303
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210222047.7A Pending CN114610898A (en) | 2022-03-09 | 2022-03-09 | Method and system for constructing supply chain operation knowledge graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114610898A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116127047A (en) * | 2023-04-04 | 2023-05-16 | 北京大学深圳研究生院 | Method and device for establishing enterprise information base |
| CN116720819A (en) * | 2023-08-10 | 2023-09-08 | 福建省闽清双棱纸业有限公司 | Impregnated paper raw material management system integrating knowledge graph and neural network |
| WO2023246007A1 (en) * | 2022-06-23 | 2023-12-28 | 广州大学 | Value chain knowledge discovery method under personalized customization |
| CN117473199A (en) * | 2023-11-17 | 2024-01-30 | 广东永锢电子机械科技有限公司 | Information pushing method and system applied to supply chain logistics system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109284394A (en) * | 2018-09-12 | 2019-01-29 | 青岛大学 | A method of Company Knowledge map is constructed from multi-source data integration visual angle |
| CN113722564A (en) * | 2021-09-09 | 2021-11-30 | 国网电子商务有限公司 | Visualization method and device for energy and material supply chain based on space map convolution |
-
2022
- 2022-03-09 CN CN202210222047.7A patent/CN114610898A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109284394A (en) * | 2018-09-12 | 2019-01-29 | 青岛大学 | A method of Company Knowledge map is constructed from multi-source data integration visual angle |
| CN113722564A (en) * | 2021-09-09 | 2021-11-30 | 国网电子商务有限公司 | Visualization method and device for energy and material supply chain based on space map convolution |
Non-Patent Citations (3)
| Title |
|---|
| 吴运兵: "基于多数据源的知识图谱构建方法研究", 福州大学学报(自然科学版), pages 329 - 335 * |
| 李肖俊: "多源异构数据情境中学术知识图谱模型构建研究", 现代情报, pages 88 - 97 * |
| 黄恒琪: "知识图谱研究综述", 计算机应用系统, pages 1 - 12 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023246007A1 (en) * | 2022-06-23 | 2023-12-28 | 广州大学 | Value chain knowledge discovery method under personalized customization |
| CN116127047A (en) * | 2023-04-04 | 2023-05-16 | 北京大学深圳研究生院 | Method and device for establishing enterprise information base |
| CN116720819A (en) * | 2023-08-10 | 2023-09-08 | 福建省闽清双棱纸业有限公司 | Impregnated paper raw material management system integrating knowledge graph and neural network |
| CN116720819B (en) * | 2023-08-10 | 2023-10-27 | 福建省闽清双棱纸业有限公司 | Impregnated paper raw material management system integrating knowledge graph and neural network |
| CN117473199A (en) * | 2023-11-17 | 2024-01-30 | 广东永锢电子机械科技有限公司 | Information pushing method and system applied to supply chain logistics system |
| CN117473199B (en) * | 2023-11-17 | 2024-03-19 | 广东永锢电子机械科技有限公司 | Information pushing method and system applied to supply chain logistics system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Industrial big data analytics: challenges, methodologies, and applications | |
| CN112612902A (en) | Knowledge graph construction method and device for power grid main device | |
| CN114610898A (en) | Method and system for constructing supply chain operation knowledge graph | |
| Hao et al. | Integrating and navigating engineering design decision-related knowledge using decision knowledge graph | |
| CN109255031A (en) | The data processing method of knowledge based map | |
| CN108701254A (en) | System and method for the tracking of dynamic family, reconstruction and life cycle management | |
| Pardillo et al. | Using ontologies for the design of data warehouses | |
| Masmoudi et al. | Knowledge hypergraph-based approach for data integration and querying: Application to Earth Observation | |
| CN111191047A (en) | Knowledge graph construction method for human-computer cooperation disassembly task | |
| CN115438199A (en) | Knowledge platform system based on smart city scene data middling platform technology | |
| Zhu et al. | Automatic real-time mining software process activities from SVN logs using a naive Bayes classifier | |
| Garcia-Buendia et al. | A bibliometric study of lean supply chain management research: 1996–2020 | |
| Slimani | A study investigating typical concepts and guidelines for ontology building | |
| Boukhari et al. | The role of user requirements in data repository design | |
| CN115221337A (en) | Data weaving processing method and device, electronic equipment and readable storage medium | |
| Chudasama et al. | InterpretME: A tool for interpretations of machine learning models over knowledge graphs | |
| Nica et al. | New research directions in knowledge discovery and allied spheres | |
| Аndroshchuk et al. | Model of Explicit Knowledge Management in Organizational and Technical Systems | |
| Liu et al. | Research on the framework of decision support system based on ERP systems | |
| Palshikar et al. | RINX: A system for information and knowledge extraction from resumes | |
| Eckhardt et al. | Uncertainty issues and algorithms in automating process connecting web and user | |
| Khan et al. | Cloud-based big data management and analytics for scholarly resources: Current trends, challenges and scope for future research | |
| Sonnenschein et al. | Validating and constructing behavioral models for simulation and projection using automated knowledge extraction | |
| Tomingas et al. | Discovering Data Lineage from Data Warehouse Procedures | |
| Repta et al. | Towards Document Flow Discovery in e-health Systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |