A knowledge graph construction method for modular design of sewing equipment

Abstract

The present invention discloses a method for constructing a knowledge graph in the field of modular design of sewing equipment. The present invention extracts entities and entity relationships from the data in the field of modular design of sewing equipment to obtain a domain knowledge set; then adopts a graph-based method to perform entity linking to obtain a domain knowledge data layer; then adopts the word frequency-inverse document rate method and the K-means clustering algorithm in sequence to extract and cluster domain ontology terms, and then adopts a template-based method to extract the classification and non-classification relationships between the ontology of the domain ontology of modular design of sewing equipment, thereby forming a domain knowledge pattern layer; the domain knowledge data layer and the pattern layer are stored in a graph database to realize the visualization of the knowledge graph. In view of the inherent characteristics of the field of modular design of sewing equipment, the present invention realizes the entity extraction, linking and domain ontology construction of multi-source heterogeneous data of modular design of sewing equipment, and realizes the visual representation of the knowledge graph in the field of modular design of sewing equipment.

Description

Technical Field

[0001] The present invention relates to a text processing and knowledge graph construction method in the field of modular design of sewing equipment, and specifically to a knowledge graph construction method in the field of modular design of sewing equipment. Background Art

[0002] In the sewing equipment market, there are many product types, and user needs vary, primarily in terms of sewing range and automation configuration. The application of modular product design methods allows for modular production of products with various sewing ranges and specifications. Different module selections and combinations can be used to form product lines with different functional systems, giving the product design a high degree of independence, interchangeability, and versatility. The modular design approach can maximize the satisfaction of society's requirements for the product. At the same time, standardizing different functions and structures, modular production also shortens the product production cycle, ultimately increasing the company's market share, ensuring economic benefits, and laying the foundation for its long-term development.

[0003] The modular design process for sewing equipment involves not only numerous theoretical deductions and extensive data calculations, but also the design experience and knowledge of domain experts. Sewing machines have been produced in my country for many years, accumulating a wealth of experience and preserving a vast amount of historical sewing equipment design data. With the iteration and upgrade of sewing equipment, the complex structures and processes have made sewing equipment design even more difficult. This vast amount of multi-source, heterogeneous sewing equipment design data lacks effective organization, making it difficult to provide technical personnel with simple and easy-to-use knowledge acquisition services. To comprehensively retain expert knowledge and experience, shorten the design and development cycle, reduce development costs, and improve design quality delivery, efficient data knowledge organization and convenient and easy-to-use knowledge acquisition methods have become urgent issues to be addressed in the modular design of sewing equipment. Therefore, applying knowledge graph technology to the modular design process of sewing equipment is essential. Summary of the Invention

[0004] The purpose of the present invention is to address the above-mentioned problems existing in the prior art and provide a method for constructing a knowledge graph in the field of modular design of sewing equipment. For the multi-source heterogeneous data of modular design of sewing equipment, a semantic recognition method is adopted to realize the construction of data layer and pattern layer, construct a domain knowledge graph, and realize the visual expression of the domain knowledge graph.

[0005] In order to achieve the above objectives, the technical solutions of the present invention are as follows:

[0006] 1. A knowledge graph construction method for modular design of sewing equipment

[0007] Step 1: Extract sewing equipment modular design entities and entity relationships from the sewing equipment modular design domain data to obtain a knowledge set in the sewing equipment modular design domain;

[0008] Step 2: Based on the domain knowledge set of modular design of sewing equipment, a graph-based method is used to perform entity linking to obtain the domain knowledge data layer of modular design of sewing equipment;

[0009] Step 3: Based on the domain knowledge set of sewing equipment modular design, the word frequency-inverse document rate method and K-means clustering algorithm are used to extract and cluster domain ontology terms to obtain the sewing equipment modular design domain ontology. Then, the template-based method is used to extract the classification and non-classification relationships between the ontology of sewing equipment modular design domain ontology. The sewing equipment modular design domain ontology and the classification and non-classification relationships between ontologies constitute the knowledge model layer of sewing equipment modular design domain.

[0010] Step 4: Store the knowledge data layer and pattern layer in the field of modular design of sewing equipment in the graph database, and realize the visualization of the knowledge graph in the field of modular design of sewing equipment in the graph database.

[0011] The step 1 is specifically as follows:

[0012] The data in the field of modular design of sewing equipment is divided into structured data, semi-structured data, and unstructured data according to the data storage type. Entities are extracted from structured data and semi-structured data by constructing regular expressions, and entities are extracted from unstructured data by using a machine learning-based method, thereby obtaining entity extraction results for the data in the field of modular design of sewing equipment;

[0013] Then, the entity relationship extraction method based on dependency relationship is used to extract entity relationship of the sewing equipment modular design domain data, and the entity relationship of the sewing equipment modular design domain data is obtained;

[0014] The entity extraction results and entity relationship extraction results of the sewing equipment modular design domain data constitute the sewing equipment modular design domain knowledge set.

[0015] The step 2 is specifically as follows:

[0016] Firstly, according to the domain knowledge set of modular design of sewing equipment, each target entity word and the corresponding alternative link entity set of each target entity word are determined. For each target entity word and its corresponding alternative link entity set, a graph-based method is used to perform entity linking on the current target entity word and its corresponding alternative link entity set to obtain an entity link graph. Then, according to the entity link graph, the comprehensive similarity between the current target entity word and each alternative link entity in the alternative link entity set is calculated respectively. Then, the alternative link entity with a comprehensive similarity greater than the comprehensive similarity threshold is selected as the target link entity of the current target entity word. Finally, the domain knowledge data layer of modular design of sewing equipment is composed of each target entity word, the corresponding target link entity and the corresponding entity relationship.

[0017] The calculation formula for the comprehensive similarity between the current target entity word and each candidate link entity in the candidate link entity set is as follows:

[0018] w(v i )=α1×w1(v i )+α2×w2(v i )+α2×w3(v i )

[0019] α1+α2+α3=1

[0020]

[0021] Among them, w(v i ) represents the comprehensive similarity between the current target entity word item and the i-th candidate link entity in the candidate link entity set, α1, α2 and α3 are the important correlation coefficient, sentence structure similarity coefficient and word node similarity coefficient respectively; w1(v i ) represents the word node v corresponding to the i-th candidate link entity in the current entity link graph i The important correlation, w2(v i ) represents the word node v corresponding to the i-th candidate link entity i The sentence structure similarity with the current target entity word item, w3(v i ) represents the word node v corresponding to the current target entity word item and the i-th candidate link entity i The word node similarity of Represents all word nodes v corresponding to the i-th candidate link entity i The set of nodes that indicate the relationship, V(v j ) represents the word node v jThe total number of pointed relationships to other word nodes in the current entity link graph, N represents the total number of word nodes in the current entity link graph, ε represents the damping coefficient; H(item) represents the lexical order annotation in the sentence where the current target entity word item is located, H(v i ) represents the word node v corresponding to the i-th candidate link entity i The order of words in the sentence is marked; Represents the word frequency vector of the current target entity word item in the corresponding alternative link entity set, Represents the word node v corresponding to the i-th candidate link entity i The term frequency vector in the corresponding set of candidate link entities, cos() represents the cosine distance calculation function.

[0022] The step three is specifically as follows:

[0023] Firstly, based on the domain knowledge set of modular design of sewing equipment, the word frequency-inverse document rate method is used to extract ontology terms and obtain the domain ontology term set;

[0024] Then, the K-means clustering algorithm is used to integrate and cluster the domain ontology term sets to obtain multiple sewing equipment modular design domain ontologies;

[0025] Finally, a template-based method is used to extract the classification relationships and non-classification relationships between multiple sewing equipment modular design domain ontologies. The sewing equipment modular design domain knowledge model layer is composed of the sewing equipment modular design domain ontology and the classification relationships and non-classification relationships between ontologies.

[0026] 2. A storage medium storing a computer program, wherein the computer program implements the method when executed by a processor.

[0027] The computer program described herein is an instruction corresponding to the implementation of the method.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] The method of the present invention, in view of the characteristics of the modular design field of sewing equipment, such as a large number of product types, a long history and abundant experience data, starts from the perspective of comprehensively retaining expert knowledge and experience, shortening the design and development cycle, reducing development costs and improving design quality. It adopts a graph-based method to realize modular design entity extraction from multi-source heterogeneous data in the modular design field of sewing equipment, realizes the construction of data layer and pattern layer in the modular design field of sewing equipment, and further realizes visual representation of the modular design field of sewing equipment.

[0030] This method effectively leverages the information provided by a large amount of historical sewing equipment design data. During entity extraction, the similarity between candidate link entities and entity referents is comprehensively calculated using key relevance, sentence structure similarity, and word node similarity, resulting in a more closely matched sewing equipment modular design entity. Furthermore, this method further implements a visual representation of the sewing equipment modular design knowledge graph. Furthermore, this visual representation can be dynamically updated, expanded, and enriched, offering excellent maintainability and scalability, facilitating future services and applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Schematic diagram of the method of the present invention.

[0032] Figure 2 This is an entity link flow chart for the modular design field of sewing equipment of the present invention.

[0033] Figure 3 This is a schematic diagram of the partial visualization results of the knowledge graph in the field of modular design of sewing equipment implemented by the present invention using Neo4j. DETAILED DESCRIPTION

[0034] The present invention will be further described in detail below with reference to specific embodiments.

[0035] like Figure 1 As shown, the embodiment of the present invention and its implementation process are as follows:

[0036] Step 1: Extract sewing equipment modular design entities and entity relationships from the sewing equipment modular design domain data to obtain a knowledge set in the sewing equipment modular design domain;

[0037] Step 1 is as follows:

[0038] Sewing equipment module design knowledge (SEMK) in the sewing equipment module design domain data includes sewing equipment module objects, sewing equipment module attributes, sewing equipment module components, and sewing equipment module processes. The specific formula is as follows:

[0039] <semk>=<O,A,C,M>

[0040] Here, O represents the sewing equipment module object (mechanism, system, etc.); A represents the sewing equipment module attributes (topological relationships, parameters, etc.); C represents the sewing equipment module components (parts, components, etc.); and M represents the sewing equipment module processes (manufacturing steps, process information, etc.). The above classification of sewing equipment modular design knowledge provides underlying support for the extraction of knowledge entities, attributes, and relationships.

[0041] Data in the field of modular design of sewing equipment is categorized into structured data, semi-structured data, and unstructured data based on data storage type. Structured data refers to data logically expressed and implemented using a two-dimensional table structure, such as data stored in system databases and spreadsheet documents. Semi-structured data is a data structure between structured and unstructured data, such as data stored in log files, XML documents, and JSON documents. Unstructured data refers to data with irregular or incomplete structures, lacking predefined definitions, and not easily represented using a two-dimensional logical table, such as data stored in text documents, e-books, and web pages. Entity extraction is performed on structured and semi-structured data by constructing regular expressions, and on unstructured data using a machine learning-based method. Entity extraction results for data in the field of modular design of sewing equipment are obtained. The entity extraction results are specifically a collection of entity words. The machine learning-based method specifically uses a hidden Markov model for entity extraction.

[0042] Next, we applied a dependency-based entity relationship extraction method to the data in the field of modular design of sewing equipment, obtaining entity relationships within the data. Specifically, through syntactic analysis, we found dependency relationships between sentences, revealing the syntactic structure. Common sentence dependency relationships include verb-object relationships, parallel relationships, subject-predicate relationships, verb-complement structures, preposition-object structures, adverbial-predicate relationships, and attributive-predicate relationships.

[0043] The entity extraction results of unstructured data and structured and semi-structured data constitute the domain dictionary of sewing equipment modular design. According to the domain dictionary of sewing equipment modular design and combined with sewing equipment modular design knowledge (SEMK), the entity words are classified into: sewing equipment module object nouns, sewing equipment module attribute nouns, sewing equipment module component nouns, sewing equipment module process actions, and sewing equipment module quantity quantifiers, thereby determining the entity relationship of the sewing equipment module.

[0044] The entity relationships of common sewing equipment modules are shown in Table 1. According to the entity relationships of common sewing equipment modules, a classification algorithm based on support vector machine is used to perform calculation and classification, and then the entity relationships of sewing equipment modules are extracted.

[0045] Table 1 Entity relationships of common sewing equipment modules

[0046]

[0047] The entity extraction results and entity relationship extraction results of the sewing equipment modular design domain data constitute the sewing equipment modular design domain knowledge set.

[0048] Step 2: Based on the domain knowledge set of modular design of sewing equipment, a graph-based method is used to perform entity linking to obtain the domain knowledge data layer of modular design of sewing equipment;

[0049] like Figure 2 As shown, step two is specifically as follows:

[0050] First, based on the domain knowledge set of modular design of sewing equipment, each target entity word and the set of candidate linked entities corresponding to each target entity word are determined. In the specific implementation, word2vec training is used to generate word vectors for the entity words in the domain knowledge set of modular design of sewing equipment. The cosine similarity between the word vector of the target entity word and the word vectors of other entity words in the set is calculated. A threshold is set, and the word vectors with a value greater than the threshold are selected as the candidate linked entities corresponding to the target entity word. For each target entity word and its corresponding candidate linked entity set, a graph-based method is used to perform entity linking on the current target entity word and its corresponding candidate linked entity set, obtaining an entity link graph G, satisfying G = (V, E), where V represents the set of graph nodes (including the target entity word item and its candidate linked entity set, with the target entity word item as the vertex), and E represents the set of relationships between word nodes in the entity link graph, and the word nodes are directed connections. Then, the comprehensive similarity between the current target entity word and each alternative link entity in the alternative link entity set is calculated according to the entity link graph, and then the alternative link entity with a comprehensive similarity greater than the comprehensive similarity threshold is selected as the target link entity of the current target entity word; finally, the knowledge data layer in the field of modular design of sewing equipment is composed of each target entity word, the corresponding target link entity and the corresponding entity relationship.

[0051] The calculation formula for the comprehensive similarity between the current target entity word and each candidate link entity in the candidate link entity set is as follows:

[0052] w(v i )=α1×w1(v i )+α2×w2(v i )+α2×w3(v i )

[0053] α1+α2+α3=1

[0054]

[0055] Among them, w(v i ) represents the comprehensive similarity between the current target entity word item and the i-th candidate link entity in the candidate link entity set. α1, α2, and α3 are the important correlation coefficient, sentence structure similarity coefficient, and word node similarity coefficient, respectively. The values of α1, α2, and α3 are obtained through experiments. i ) represents the word node v corresponding to the i-th candidate link entity calculated using the PageRank algorithm in the current entity link graph i The important correlation, w2(v i ) represents the word node v corresponding to the i-th candidate link entity i The sentence structure similarity with the current target entity word item, w3(v i ) represents the word node v corresponding to the current target entity word item and the i-th candidate link entity i The word node similarity of Represents all word nodes v corresponding to the i-th candidate link entity i The set of nodes that indicate the relationship, V(v j ) represents the word node v j The total number of pointed relationships to other word nodes in the current entity link graph, N represents the total number of word nodes in the current entity link graph, ε represents the damping coefficient, which is generally 0.85. When calculating, the initial value of ε is 1 / N, and the iterative calculation reaches a steady state to obtain the important relevance of the word node; h(item) represents the lexical order annotation in the sentence where the current target entity word item is located, H(v i ) represents the word node v corresponding to the i-th candidate link entity i The order of words in the sentence is marked; Represents the word frequency vector of the current target entity word item in the corresponding alternative link entity set, Represents the word node v corresponding to the i-th candidate link entity i The term frequency vector in the corresponding set of candidate link entities, cos() represents the cosine distance calculation function.

[0056] Step 3: Based on the domain knowledge set of modular design of sewing equipment, the term frequency-inverse document rate method TF-IDF and K-means clustering algorithm are used in sequence to extract and cluster domain ontology terms to obtain the domain ontology of modular design of sewing equipment. Then, the classification and non-classification relationships between the ontology of modular design of sewing equipment are extracted using a template-based method. The domain knowledge model layer of modular design of sewing equipment is composed of the domain ontology of modular design of sewing equipment and the classification and non-classification relationships between ontologies.

[0057] Step three is as follows:

[0058] Firstly, a bottom-up approach is adopted to construct the domain ontology. The modular design domain of sewing equipment is collected. Considering the reuse of existing knowledge graph pattern layer, the domain ontology concepts, attributes, relationships and constraints are defined. The domain ontology is integrated and consolidated, and the ontology is obtained through comprehensive evaluation.

[0059] Firstly, based on the domain knowledge set of modular design of sewing equipment, the TF-IDF algorithm is used to extract ontology terms and obtain the domain ontology term set.

[0060] TF-IDF, or term frequency-inverse document frequency, is often used as the preferred method for extracting text feature information in related fields such as information retrieval and text mining. Its mathematical expression is as follows:

[0061] tf jg idf j =W tf ×W idf

[0062] Among them, W tf Indicates word frequency, indicates word node v i Frequency of occurrence, W idf Represents word node v i Inverse Document Frequency. tf jg idf j Reaction word node v i The importance in the domain is determined by setting an importance threshold to extract word nodes that meet the requirements as domain ontology terms.

[0063] Then, the K-means clustering algorithm is used to integrate and cluster the domain ontology term sets to obtain multiple sewing equipment modular design domain ontologies;

[0064] Specifically: cluster the ontology terms into k clusters, train the word vectors corresponding to the extracted domain ontology terms using the Word2Vec method, and use the Manhattan distance to calculate the distance from each word vector to the cluster center:

[0065]

[0066] in, Represents word vector v i The k-dimensional coordinates, d(v i ,v j ) represents the word vector v i ,v j The Manhattan distance between them is calculated. Based on the calculated distance, the cluster center is updated. After iterative calculation, a stable cluster center word vector is obtained. These stable cluster center words are selected as the k domain ontologies after clustering integration.

[0067] Finally, a template-based method is used to extract the classification and non-classification relationships between multiple sewing equipment modular design domain ontologies. The main classification and non-classification relationships are shown in the following table:

[0068] Table 2 Ontology classification relationship template

[0069]

[0070] Table 3 Ontology non-classification relationship template

[0071]

[0072] The knowledge model layer of the sewing equipment modular design domain is composed of the sewing equipment modular design domain ontology and the classification relationship and non-classification relationship between ontologies.

[0073] Ontology is a collection of concepts that defines the concepts and concept attributes of the knowledge graph. It is a conceptual framework, and the relationships between ontologies constitute the model layer here; entities are the integration of ontologies, instances and relationships, and can also be defined as instances of a concept in the ontology. The entities and the relationships between entities constitute the data layer here.

[0074] Step 4: Store the knowledge data layer and pattern layer in the field of modular design of sewing equipment in the graph database, and realize the visualization of the knowledge graph in the field of modular design of sewing equipment in the graph database.

[0075] The specific embodiments are as follows:

[0076] First, based on a large amount of historical sewing equipment design data and existing sewing equipment design data with different structures, entity extraction is performed in multi-dimensional heterogeneous data to achieve a more comprehensive entity extraction and data layer construction for modular design of sewing equipment.

[0077] Next, the sewing equipment modular design ontology is constructed. Based on the identified domain ontology and the reuse of existing ontologies, the ontology domain and scope are determined. Then, based on the identified classification relationships, the sewing equipment modular design ontology is classified:

[0078] ①Sewing equipment: lockstitch sewing equipment, chainstitch sewing equipment, etc.;

[0079] ②Main mechanisms: material piercing mechanism, thread picking mechanism, thread hooking mechanism and feeding mechanism, etc.;

[0080] ③Function: working mechanism, auxiliary device, etc.

[0081] Based on the identified non-classified relationships, attribute relationships and constraints are determined. Attribute types primarily include object attributes, which are relationships between entities, and data attributes, which are the data characteristics inherent in the entity itself. In the modular design of sewing equipment, the main relationships found in the object attributes of entities from non-classified relationships include adjacency, sequence, and positional topology. For example, the mechanism motion relationships are piercing, hooking, picking, and feeding. The main data attribute relationships include motion and attribute parameters. For example, the rod length of the piercing mechanism satisfies the rod length requirements of the centring crank slider mechanism.

[0082] Neo4j software is used to store the sewing equipment modular design knowledge graph according to the above rules and realize its visualization. The obtained visualization results are as follows: Figure 3 This can fully and effectively utilize a large amount of historical sewing equipment design data and the information provided in the existing sewing equipment design materials of different structures. It not only realizes the construction of the data layer and pattern layer of the modular design of sewing equipment, but also realizes its visual expression. It can provide greater convenience for sewing equipment designers and provide important support for the innovation of modular design of sewing equipment.< / semk>

Claims

1. A method for constructing a knowledge graph in the field of modular design of sewing equipment, characterized in that: The following steps are involved: Step 1: Extract sewing equipment modular design entities and entity relationships from the sewing equipment modular design domain data to obtain a knowledge set in the sewing equipment modular design domain; Step 2: Based on the domain knowledge set of modular design of sewing equipment, a graph-based method is used to perform entity linking to obtain the domain knowledge data layer of modular design of sewing equipment; Step 3: Based on the domain knowledge set of sewing equipment modular design, the word frequency-inverse document rate method and K-means clustering algorithm are used to extract and cluster domain ontology terms to obtain the sewing equipment modular design domain ontology. Then, the template-based method is used to extract the classification and non-classification relationships between the ontology of sewing equipment modular design domain ontology. The sewing equipment modular design domain ontology and the classification and non-classification relationships between ontologies constitute the knowledge model layer of sewing equipment modular design domain. Step 4: Store the domain knowledge data layer and pattern layer of sewing equipment modular design in the graph database, and visualize the domain knowledge graph of sewing equipment modular design in the graph database; The step 2 is specifically as follows: First, based on the domain knowledge set of modular design of sewing equipment, each target entity word and the corresponding alternative link entity set of each target entity word are determined. For each target entity word and its corresponding alternative link entity set, a graph-based method is used to perform entity linking on the current target entity word and its corresponding alternative link entity set to obtain an entity link graph. Then, based on the entity link graph, the comprehensive similarity between the current target entity word and each alternative link entity in the alternative link entity set is calculated. Then, the alternative link entity with a comprehensive similarity greater than a comprehensive similarity threshold is selected as the target link entity of the current target entity word. Finally, the domain knowledge data layer of modular design of sewing equipment is composed of each target entity word, the corresponding target link entity, and the corresponding entity relationship. The calculation formula for the comprehensive similarity between the current target entity word and each candidate link entity in the candidate link entity set is as follows: w(v i )=α1×w1(v i )+α2×w2(v i )+α2×w3(v i ) α1+α2+α3＝1 Among them, w(v i ) represents the comprehensive similarity between the current target entity word item and the i-th candidate link entity in the candidate link entity set, α1, α2 and α3 are the important correlation coefficient, sentence structure similarity coefficient and word node similarity coefficient respectively; w1(v i ) represents the word node v corresponding to the i-th candidate link entity in the current entity link graph i The important correlation, w2(v i ) represents the word node v corresponding to the i-th candidate link entity i The sentence structure similarity with the current target entity word item, w3(v i ) represents the word node v corresponding to the current target entity word item and the i-th candidate link entity i The word node similarity of Represents all word nodes v corresponding to the i-th candidate link entity i The set of nodes that indicate the relationship, V(v j ) represents the word node v j The total number of pointed relationships to other word nodes in the current entity link graph, N represents the total number of word nodes in the current entity link graph, ε represents the damping coefficient; h(item) represents the lexical order annotation in the sentence where the current target entity word item is located, H(v i ) represents the word node v corresponding to the i-th candidate link entity i The order of words in the sentence is marked; Represents the word frequency vector of the current target entity word item in the corresponding alternative link entity set, Represents the word node v corresponding to the i-th candidate link entity i The term frequency vector in the corresponding set of candidate link entities, cos() represents the cosine distance calculation function.

2. The method for constructing a knowledge graph in the field of modular design of sewing equipment according to claim 1, characterized in that: The step 1 is specifically as follows: The data in the field of modular design of sewing equipment is divided into structured data, semi-structured data, and unstructured data according to the data storage type. Entities are extracted from structured data and semi-structured data by constructing regular expressions, and entities are extracted from unstructured data by using a machine learning-based method, thereby obtaining entity extraction results for the data in the field of modular design of sewing equipment; Then, the entity relationship extraction method based on dependency relationship is used to extract entity relationship of the sewing equipment modular design domain data, and the entity relationship of the sewing equipment modular design domain data is obtained; The entity extraction results and entity relationship extraction results of the sewing equipment modular design domain data constitute the sewing equipment modular design domain knowledge set.

3. The method for constructing a knowledge graph in the field of modular design of sewing equipment according to claim 1, characterized in that: The step three is specifically as follows: Firstly, based on the domain knowledge set of modular design of sewing equipment, the word frequency-inverse document rate method is used to extract ontology terms and obtain the domain ontology term set; Then, the K-means clustering algorithm is used to integrate and cluster the domain ontology term sets to obtain multiple sewing equipment modular design domain ontologies; Finally, a template-based method is used to extract the classification relationships and non-classification relationships between multiple sewing equipment modular design domain ontologies. The sewing equipment modular design domain knowledge model layer is composed of the sewing equipment modular design domain ontology and the classification relationships and non-classification relationships between ontologies.

4. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

5. A storage medium according to claim 4, characterized in that: The computer program is an instruction corresponding to the method described in any one of claims 1 to 3.

Images