High-speed train product structure tree knowledge fusion method, device and program product

Abstract

The present invention provides a method, device and program product for knowledge fusion of a high-speed train product structure tree, the method comprising: obtaining organizational analysis data of high-speed train multivariate data, constructing an ontology model, and constructing a domain knowledge graph according to the ontology model; fusing the domain knowledge graph into a stage knowledge graph, wherein the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by entity alignment based on clustering entities using multi-information fusion similarity; based on the stage domain structure tree of organizational analysis data, mapping and fusion of the ontology of the stage domain structure tree is performed to achieve fusion of knowledge graphs of each stage. The present invention can achieve fusion of knowledge graphs of each stage, make the interconnection relationship of domain knowledge of each stage clearer, and solve the defects of redundant and heterogeneous data of high-speed trains.

Description

Technical Field

[0001] The present invention relates to the technical field of train heterogeneous knowledge fusion, and in particular to a high-speed train product structure tree knowledge fusion method, device and program product. Background Art

[0002] The life cycle of a high-speed train includes multiple stages, including design, manufacturing, and operation and maintenance. With the accumulation of historical data from multiple stages, a large amount of empirical knowledge has been formed, which can provide empirical knowledge for each stage of product design, manufacturing, and operation and maintenance. However, this empirical knowledge is stored on different platforms, and the sources of knowledge in each life cycle stage are different and the formal structure is complex. Generally, knowledge graphs are used to organize and represent empirical knowledge, so as to better obtain empirical knowledge.

[0003] Knowledge graphs are called learned AI and are often used to fuse multi-source data to build large-scale knowledge bases. Knowledge graphs are essentially a semantic knowledge base, a semantic network used to represent semantic relationships between entities, and are suitable for describing the complex structural knowledge of high-speed trains.

[0004] In the existing technology, since the knowledge sources of high-speed trains are complex and exist at different levels in different stages, the knowledge graphs constructed at different stages have heterogeneous characteristics, resulting in unclear mutual correlation between knowledge at different stages, non-interoperability of knowledge, and redundant and heterogeneous knowledge. Summary of the invention

[0005] The present invention provides a method, device and program product for integrating knowledge of a high-speed train product structure tree, so as to solve the defects of redundant and heterogeneous multi-stage knowledge and heterogeneous knowledge graphs constructed in different stages in the prior art, realize the integration of knowledge graphs of each stage, and make the interconnection relationship between domain knowledge of each stage clearer.

[0006] The present invention provides a high-speed train product structure tree knowledge fusion method, comprising:

[0007] Obtaining organizational analysis data of high-speed train multivariate data, constructing an ontology model, and constructing a domain knowledge graph based on the ontology model;

[0008] The domain knowledge graph is merged into a stage knowledge graph, wherein the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning entities based on clustered entities using multi-information fusion similarity;

[0009] Based on the stage domain structure tree of the organizational analysis data, mapping and fusion of the stage domain structure tree ontology is performed to achieve fusion of knowledge graphs of each stage.

[0010] According to the knowledge fusion method of the high-speed train product structure tree provided by the present invention, the method of obtaining the organizational analysis data of the high-speed train multivariate data, constructing the ontology model, and constructing the domain knowledge graph according to the ontology model includes:

[0011] Based on the multivariate data of high-speed trains, organizational analysis data is obtained, wherein the organizational analysis data includes: data source, domain stage, domain structure tree and stage domain structure tree, and the stage domain structure tree includes: product family main structure tree, design instance structure tree and assembly instance structure tree;

[0012] Based on the organizational analysis data, obtain the context information and data types of data services and build ontology models;

[0013] The global ontology is obtained according to the domain structure tree of the stage, the local ontology is obtained based on the global ontology, and the data is obtained in combination with the ontology model to construct a domain knowledge graph.

[0014] According to the high-speed train product structure tree knowledge fusion method provided by the present invention, the data sources include structured data, semi-structured data and unstructured data.

[0015] According to the knowledge fusion method of the high-speed train product structure tree provided by the present invention, the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by the named entity recognition of the instance data, including:

[0016] Based on the pre-acquired corpus data, a named entity recognition model is trained and acquired, wherein the corpus data is entity labeled and sequence labeled and then input into the named entity recognition model;

[0017] Based on the named entity recognition model, obtaining the ontology concepts corresponding to the entities in the domain knowledge graph;

[0018] Based on the comparison result between the ontology concepts corresponding to the entities in the domain knowledge graph and the ontology concepts in the domain structure tree, the ontology concept mapping relationship of the stage knowledge graph is obtained.

[0019] According to the knowledge fusion method of the high-speed train product structure tree provided by the present invention, the entity fusion of the stage knowledge graph is realized by aligning entities based on clustering entities using multi-information fusion similarity, including:

[0020] Normalizing the attributes of the entities in the domain knowledge graph, wherein the types of the attributes include structured attributes and unstructured attributes;

[0021] Based on the ontology concept mapping relationship of the knowledge graph in the stage, cluster entities under the same ontology concept with a mapping relationship are obtained;

[0022] Based on the unit and constraint matching of the structured attributes, using the minimum edit distance, determining the similarity of the structured attributes in the clustering entities;

[0023] Based on the semantic feature vector of the unstructured attribute, the unstructured attribute similarity of the clustering entity is determined by using the cosine similarity of the semantic feature vector;

[0024] Based on the structured attribute similarity and the unstructured attribute similarity, the comprehensive similarity of entities is determined to achieve entity alignment of the knowledge graph at the stage.

[0025] According to the knowledge fusion method of the high-speed train product structure tree provided by the present invention, the stage domain structure tree based on the organization analysis data performs mapping fusion of the stage domain structure tree ontology to realize the fusion of the knowledge graphs of each stage, including:

[0026] Based on the coding attributes of the stage domain structure tree, a product coding structure tree is constructed to achieve ontology alignment;

[0027] Normalize the entity attributes of the domain structure tree of the stage;

[0028] Obtaining cluster entities under the ontology concept of the domain structure tree of the stage;

[0029] Based on the clustered entities, the entity comprehensive similarity is determined to obtain the mapping relationship between the entities of each domain stage under the ontology concept of the stage domain structure tree, wherein the comprehensive similarity includes structured attribute similarity and unstructured attribute similarity.

[0030] The present invention also provides a high-speed train product structure tree knowledge fusion device, comprising:

[0031] A construction module is used to obtain organizational analysis data of high-speed train multivariate data, construct an ontology model, and construct a domain knowledge graph based on the ontology model;

[0032] A first fusion module is used to fuse the domain knowledge graph into a stage knowledge graph, wherein the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning entities based on clustered entities using multi-information fusion similarity;

[0033] The second fusion module is used to perform mapping fusion of the stage domain structure tree ontology based on the stage domain structure tree of the organization analysis data, so as to realize the fusion of the knowledge graphs of each stage.

[0034] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for fusion of knowledge of a high-speed train product structure tree as described above is implemented.

[0035] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for integrating knowledge of a high-speed train product structure tree as described in any one of the above is implemented.

[0036] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned high-speed train product structure tree knowledge fusion methods.

[0037] The method, device and program product for integrating knowledge of a high-speed train product structure tree provided by the present invention organize and analyze high-speed train multivariate data, a stage domain structure tree and a domain structure tree, construct a domain knowledge graph through an ontology model based on the domain structure tree, map and align the ontology concepts output by named entity recognition based on instance data based on the domain knowledge graph, and perform entity alignment based on clustered entities using multi-information fusion similarity, so as to integrate heterogeneous domain knowledge graphs into stage knowledge graphs and realize data fusion of stages; at the same time, based on the stage domain structure tree, a product coding structure tree is constructed to realize ontology alignment of the stage domain structure tree, and mapping and fusion of instance data layers in the stage domain structure tree are realized by establishing mapping relationships between entities of each stage, so as to realize the integration of knowledge graphs of each stage of high-speed trains, so that the intercommunication and interconnection relationships between data in the field of high-speed trains are clearer, and the heterogeneous and redundant problems of multi-stage data in the field of high-speed trains are further solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 It is a flow chart of the knowledge fusion method of the high-speed train product structure tree provided by the present invention;

[0040] Figure 2 It is a schematic diagram of the construction process of the ontology model provided by the present invention;

[0041] Figure 3 This is one of the schematic diagrams of the ontology alignment process of the stage knowledge graph provided by the present invention;

[0042] Figure 4 This is the second schematic diagram of the ontology alignment process of the stage knowledge graph provided by the present invention;

[0043] Figure 5 It is a schematic diagram of the fusion process of the knowledge graphs at each stage provided by the present invention;

[0044] Figure 6 It is a schematic diagram of the structure of the knowledge fusion device of the high-speed train product structure tree provided by the present invention;

[0045] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Combine the following Figure 1-Figure 5 The knowledge fusion method of the high-speed train product structure tree of the present invention is described.

[0048] Figure 1 FIG. 1 is a flow chart of the knowledge fusion method of the high-speed train product structure tree provided by the present invention. Figure 1 As shown, the method includes:

[0049] Step 110: obtain organizational analysis data of high-speed train multivariate data, construct an ontology model, and construct a domain knowledge graph based on the ontology model.

[0050] Optionally, Figure 2 It is a schematic diagram of the construction process of the ontology model provided by the present invention, such as Figure 2 As shown, the method includes:

[0051] Based on the multivariate data of high-speed trains, the organizational analysis data is obtained, wherein the organizational analysis data includes: data source, domain stage, domain structure tree and stage domain structure tree. The stage domain structure tree includes: product family main structure tree, design instance structure tree and assembly instance structure tree.

[0052] Based on the organizational analysis data, obtain the context information and data types of data services and build ontology models;

[0053] The global ontology is obtained according to the stage domain structure tree, the local ontology is obtained based on the global ontology, and the data is obtained in combination with the ontology model to build a domain knowledge graph.

[0054] Optionally, high-speed train data has a phased characteristic, and its product life cycle can be divided into three stages: design, manufacturing, and operation and maintenance. The data in different stages have different characteristics. In the design stage, it includes but is not limited to demand data, geometric data, and attribute parameter data. The above data corresponds to a specific design instance, and the design instance is instantiated through a metamodel. There is a one-to-many relationship between the metamodel and the design instance; through the design instance, production and manufacturing will obtain multiple corresponding manufacturing instances for specific assembly; the data in the operation and maintenance stage is generated in a specific assembly instance, so there is a many-to-one relationship between the assembly instance and the design instance. Therefore, based on the analysis of the characteristics of the data source, a stage domain structure tree and a domain structure tree are constructed. For example, the fault domain structure tree and the maintenance domain structure tree in the operation and maintenance stage both belong to the domain structure tree. The division of the stage domain structure tree is shown in Table 1:

[0055] Table 1 Stage domain structure tree division and corresponding data characteristics

[0056]

[0057] Alternatively, if Figure 2 As shown, a top-down approach is adopted to construct the ontology of the domain knowledge graph. Combining the data source, domain stage and the corresponding domain structure tree, a seven-step method is used to construct the ontology model. The steps for constructing the ontology model include: obtaining context information of data services, obtaining data types, designing an ontology summary model, instantiating verification and evaluation, formalizing the ontology, defining classes and class hierarchies, and defining attributes and constraints. Among them, in the three steps of designing an ontology summary model, instantiating verification and evaluation, and formalizing the ontology, iterative optimization is performed in combination with data layer instance data to obtain the optimal ontology model.

[0058] Optionally, the stage domain structure tree of each stage of design, manufacturing, operation and maintenance is used as the global ontology, the global ontology is locally expanded to obtain the local ontology, and the knowledge is acquired by combining the above ontology model to form a domain knowledge graph.

[0059] Optionally, the data sources include structured data, semi-structured data, and unstructured data. Structured data is converted through the D2RQ tool, which maps the table name of the relational database directly to the class in RDF, maps the field to the attribute of the class, and the relationship between the classes is obtained from the table representing the relationship; semi-structured data uses template matching to obtain entity relationships, and unstructured data can apply NLP technology and deep learning models to perform named entity recognition and relationship extraction.

[0060] Step 120: fuse the domain knowledge graph into a stage knowledge graph, wherein the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning entities based on clustered entities using multi-information fusion similarity.

[0061] Figure 3 This is one of the schematic diagrams of the ontology alignment process of the stage knowledge graph provided by the present invention. Figure 4 This is the second schematic diagram of the ontology alignment process of the stage knowledge graph provided by the present invention, such as Figure 3-Figure 4 As shown in the figure, the ontology fusion method of the stage knowledge graph includes:

[0062] Based on the pre-acquired corpus data, a named entity recognition model is trained and acquired, wherein the corpus data is entity labeled and sequence labeled and then input into the named entity recognition model;

[0063] Based on the named entity recognition model, the ontology concepts corresponding to the entities in the domain knowledge graph are obtained;

[0064] Based on the comparison results between the ontology concepts corresponding to the entities in the domain knowledge graph and the ontology concepts in the domain structure tree, the ontology concept mapping relationship of the stage knowledge graph is obtained.

[0065] Alternatively, if Figure 3-Figure 4 As shown, the ontology of the stage domain structure tree corresponding to the domain knowledge graph is used as the alignment ontology of the stage knowledge graph, and the entity of the stage domain structure tree corresponding to the domain knowledge graph is used as the alignment entity of the stage knowledge graph to realize the fusion of the stage knowledge graph.

[0066] Alternatively, if Figure 3 As shown, the ontology O1 (A1, A2, ..., Am) and entity G1 (E1, E2, ..., En) of the domain knowledge graph and the ontology O2 (B1, B2, ..., Bq) of the stage domain structure tree of the corresponding stage are obtained, where Ai represents the concept in the ontology O1, Ei represents the entity set under the concept Ai in the ontology O1, and Bj represents the concept in the ontology O2 of the stage domain structure tree. The result of ontology alignment is a mapping relationship f: Ai->Bj. Ai->Bj in the mapping set indicates that the concept Ai in the ontology O1 and the concept Bj in the ontology O2 have an equal or inclusive relationship. The equal or inclusive relationship is determined as follows:

[0067] For an entity Ei under a concept Ai, for a mapping relationship f:Ai->Bj,

[0068] like and Then Ai=Bj;

[0069] like and but

[0070] Alternatively, if Figure 4 As shown, the entity set Ei of the concept Ai in the ontology O1 is input into the named entity recognition model, and the output concept of the named entity recognition model is compared with the concept Bj in the ontology O2. If the two are the same, the mapping relationship f:Ai->Bj is output, and the search continues to find whether there is an identical mapping. If not, it is determined whether there is a mapping of the concept Ai in the ontology O1. If so, the mapping is added to the existing mapping set. If not, the mapping set is created.

[0071] Optionally, the named entity recognition model adopts a BERT-BILSTM-CRF model. The steps for building and training the named entity recognition model are as follows:

[0072] 1) Obtain corpus data

[0073] The corpus data includes but is not limited to text data from web pages and domain dictionaries. The text data in web pages is obtained from web pages by crawling, such as Baidu Encyclopedia and China National Knowledge Infrastructure. The keywords are set as domain-related words, and the obtained text data is manually pre-processed and dirty data is removed to obtain the final corpus data used for training.

[0074] 2) Entity Annotation

[0075] The acquired corpus data is annotated with domain entities through the brat tool. For example, "bogie" is annotated as "system", "axle" is annotated as "part", etc., and the annotated corpus is annotated with BIO sequence through program writing. For example, "bogie" is annotated as "BII" format, and the annotated data is divided into training set, test set and validation set, and the data volume ratio is 8:1:1.

[0076] 3) Named Entity Recognition Model Construction and Training

[0077] A named entity recognition model is built based on the BERT-BILSTM-CRF model. The training set is segmented, stop words are removed, and initial word vectors are obtained. The obtained initial word vectors are input into the BERT model for fine-tuning training to obtain the output word vectors. The output word vectors are then input into the BILSTM-CRF model for sequence labeling. The CRF model is used to improve the accuracy of the sequence labeling results and output the labeling sequence with the highest probability. The training set is input into the constructed model for training, and the results are verified by the validation set until the best training results are obtained. Finally, the sequence labeling result with the highest probability is output.

[0078] Optionally, the entity fusion method of the stage knowledge graph includes:

[0079] Normalize the attributes of entities in the domain knowledge graph, where the attribute types include structured attributes and unstructured attributes;

[0080] Based on the ontology concept mapping relationship of the stage knowledge graph, cluster entities under the same ontology concept with mapping relationship are obtained;

[0081] Based on the unit and constraint matching of structured attributes, the similarity of structured attributes in clustered entities is determined using the minimum edit distance;

[0082] Based on the semantic feature vector of the unstructured attribute, the cosine similarity of the semantic feature vector is used to determine the similarity of the unstructured attribute of the clustering entity;

[0083] Based on the similarity of structured attributes and unstructured attributes, the comprehensive similarity of entities is determined to achieve entity alignment of the stage knowledge graph.

[0084] For example, entity attribute normalization methods include but are not limited to: manual methods. The attributes of the entity are normalized manually, including but not limited to the attribute name, attribute value, and the unit of the attribute value. For example, the units "gram" and "g" in the numerical structured attributes are uniformly normalized to "g", and the attribute names "ambient maximum temperature" and "maximum ambient temperature" in the text-type unstructured attributes are uniformly normalized to "maximum ambient temperature".

[0085] Optionally, after ontology mapping of the stage knowledge graph, the entities are classified according to the concept labels in the ontology, the entities under the same concept labels are clustered, and the entities under the two ontology concepts with a mapping relationship are used as entities to be aligned to obtain clustered entities under the concept labels. That is, the mapping relationship (Ai, Bj) and the entity set Ei under Ai and the entity set Ej under Bj are obtained, and after entity clustering, the clustered entity sets Ei and Ej are output.

[0086] Optionally, a method for calculating entity structured attribute similarity includes:

[0087] For numerical structured attributes, units and constraints are matched. The unit and constraint matching templates are shown in Table 2.

[0088] Table 2 Unit and constraint matching template

[0089] Constraint Unit Flag No greater than mm 1 No more than ％ 1 ≤ mm 1 Not more than Lift 1 No more than mm 1 No less than Year 2 Not less than km 2 Not less than mm 2 ± g 6 ± kN 6 ± mm 6

[0090] The minimum edit distance is used to calculate the edit distance similarity of each common attribute. The calculation formula of the edit distance similarity is shown in formula (1):

[0091]

[0092] Among them, P ai represents the i-th attribute of entity a, P bi represents the i-th attribute of entity b, edit(V ai ,V bi ) represents the minimum edit distance between the i-th common attribute value of entity a and entity b, len(V ai ) and len(V bi ) represent the length of the i-th common attribute of entity a and entity b respectively.

[0093] The calculation formula of structural attribute similarity is shown in formula (2):

[0094]

[0095] Among them, w i represents the weight of the ith common attribute, and t represents the number of common attributes.

[0096] Optionally, the method for calculating the unstructured attribute similarity includes:

[0097] For the unstructured text attributes in the entity, the semantic feature vector of the unstructured text attribute is constructed by word embedding through the description text of the corresponding attribute and the domain dictionary, and the similarity of the unstructured text attribute is obtained by calculating the cosine similarity of the semantic feature vector.

[0098] 1) Get word vector

[0099] The description text of the text is used as the training corpus, and the text is preprocessed using Jieba word segmentation and LTP stop word library. The text is randomly initialized to construct a word vector, and then trained with the CBOW model and BILSTM model to obtain a comprehensive word vector.

[0100] 2) Unstructured attribute similarity calculation

[0101] Obtain the feature vector of the unstructured text attribute of each entity pair to be aligned, and use each unstructured text attribute feature vector V of entity a a and each unstructured text attribute feature vector V of entity b b Calculate the cosine similarity. The cosine similarity calculation formula is shown in formula (3):

[0102]

[0103] Among them, F ai represents the i-th unstructured text attribute of entity a, F bi represents the i-th unstructured text attribute of entity b, V ai represents the feature vector corresponding to the i-th unstructured text attribute of entity a, Vbi represents the feature vector corresponding to the i-th unstructured text attribute of entity b, |V ai | and |V bi | respectively represent the feature vector V ai and V bi The modulus length, Sim(F ai ,F bi ) represents the cosine similarity of each unstructured attribute between entity pairs. The cosine similarity threshold is set to 0.6. The attribute pair with the highest cosine similarity is selected. If the cosine similarity is greater than 0.6, it is classified as a similar attribute. The similarity is recorded. After cyclic comparison, t similar attribute pairs are obtained. The unstructured text similarity between entity a and entity b is calculated. The unstructured text similarity is shown in formula (4):

[0104]

[0105] Optionally, the structured attribute similarity and the unstructured attribute similarity are weighted normalized to obtain the entity comprehensive similarity, which is shown in formula (5):

[0106] SimE(a,b)=w1·SimZ(a,b)+w2·SimF(a,b) (5),

[0107] Among them, w1 is the similarity weight of structured attributes, and w2 is the similarity weight of unstructured attributes.

[0108] For each entity a in the domain knowledge graph concept Ai, the entity b in the mapping concept Bi in the stage domain structure tree is sorted in descending order by the entity comprehensive similarity SimE(a,b) value to generate the optimal candidate sequence, and the entity b with the highest similarity and greater than the set threshold is selected as the alignable entity.

[0109] Step 130: Based on the stage domain structure tree of the organization analysis data, mapping and fusion of the stage domain structure tree ontology are performed to achieve fusion of the knowledge graphs of each stage.

[0110] Optionally, Figure 5 It is a schematic diagram of the fusion process of the knowledge graphs at each stage provided by the present invention, such as Figure 5 As shown, the method includes:

[0111] Based on the coding attributes of the stage domain structure tree, a product coding structure tree is constructed to achieve ontology alignment;

[0112] Normalize the entity attributes of the stage domain structure tree;

[0113] Acquire the cluster entities under the ontology concept of the stage domain structure tree;

[0114] Based on the clustered entities, the comprehensive similarity of the entities is determined to obtain the mapping relationship between the entities of each domain stage under the ontology concept of the stage domain structure tree, wherein the comprehensive similarity includes the similarity of structured attributes and the similarity of unstructured attributes.

[0115] Optionally, in the stage domain structure tree, the product main family structure tree has a meta-node code as a unique identifier, the design instance structure tree has both a meta-node code and a module code, wherein the module code is a unique identifier, and the assembly instance structure tree has both a meta-node code, a module code and a manufacturing code, wherein the manufacturing code is a unique identifier, and the mapping relationship between the three is meta-node code: module code: manufacturing code = 1: N: N 2 In the process of constructing the product coding structure tree ontology, a unified method is used to describe the concepts, so the ontology concepts have been aligned.

[0116] Optionally, the mapping fusion of the instance data layer is expressed as the mapping relationship between the entities of the domain structure tree of each stage, the product main family structure tree M1 (A1, A2, ..., Am), the design instance structure tree M2 (B1, B2, ..., Bn), and the assembly instance structure tree M3 (C1, C2, ..., Cq), where Ai represents the entity set under the corresponding concept label in M1, Bi represents the entity set under the corresponding concept label in M2, and Ci represents the entity set under the corresponding concept label in M3. The entities under the domain structure trees of the three stages are aligned in pairs, and the alignment result is a mapping relationship f: Ai->Bi or f: Ai->Ci or f: Bi->Ci.

[0117] Optionally, the units and constraints of numerical structured attributes may be unified manually, and the attribute names of the same attributes in text-type unstructured attributes may be unified to achieve attribute standardization.

[0118] Optionally, in domain structure trees at different stages, entities under the same concept label are obtained for clustering to obtain clustered entities.

[0119] Optionally, the method for calculating the entity comprehensive similarity in step 130 is the same as the method for calculating the entity comprehensive similarity in step 120, which is not described in detail here. After the entity comprehensive similarity is calculated, a mapping relationship is established between the two entities to obtain a triple (a, mapping, b), and the mapping relationship is stored.

[0120] For example, the fault knowledge graph and the maintenance knowledge graph in the operation and maintenance stage are taken as examples to illustrate the fusion of stage knowledge graphs, and the design instance structure tree and the assembly instance structure tree are taken as examples to illustrate the fusion of stage knowledge graphs.

[0121] Step 110, both the fault ontology and the maintenance ontology belong to the operation and maintenance stage. When constructing the fault ontology model and the maintenance ontology model, the assembly instance structure tree is used as the global ontology, and the ontology model is constructed based on the seven-step method to construct the fault ontology model and the maintenance ontology model respectively. Among them, the fault and maintenance data come from the fault record text and the maintenance manual respectively, both of which are unstructured data. The BERT-BILSTM-CRF model is used for entity extraction to construct the fault knowledge graph and the maintenance knowledge graph.

[0122] Step 120: Stage knowledge graph fusion

[0123] Since both the fault knowledge graph and the maintenance knowledge graph belong to the operation and maintenance stage, the assembly instance structure tree ontology is used as the alignment ontology, and the assembly instance structure tree entity is used as the alignment entity to realize the fusion of the stage domain knowledge graph.

[0124] 1) Mapping alignment of ontology concept layer

[0125] Obtain the fault knowledge graph ontology O1 (A1, A2, ...Am), fault knowledge graph entities G1 (E1, E2, ..., En) and assembly instance structure tree ontology O2 (B1, B2, ...Bq), and input the entities under the concepts in the fault knowledge graph into the named entity recognition model. For example, input the entities under the concepts of system, subsystem, component, such as bogie, frame and other entities into the named entity recognition model, and compare the output result with the concept of ontology O2, and obtain the entity recognition results of (bogie, system), (frame, subsystem), then establish a mapping relationship f: system->system, subsystem->subsystem, among which the entities under the concept system in the fault knowledge graph all belong to the entities in the assembly instance structure tree ontology and have an equal relationship.

[0126] Table 3 Partial mapping set

[0127]

[0128] Similarly, the maintenance knowledge graph ontology, maintenance knowledge graph entities and assembly instance structure tree ontology are obtained, and the entities under the concepts of concept structure, maintenance method, maintenance strategy, etc. are input into the named entity recognition model for named entity recognition. The entity recognition results are (bogie, system), (axle, part), etc., and the mapping relationship f is established: structure->system, structure->part. Since the entities under the concept structure in the maintenance knowledge graph include the entities under the concept system, part, and subsystem in the assembly instance structure tree ontology, the concept structure and the concept system, subsystem, component, and part are included. The established partial mapping set is shown in Table 3.

[0129] 2) Instance-level entity alignment based on multi-information fusion similarity calculation

[0130] (1) Attribute Normalization

[0131] In the fault knowledge graph, the entity attributes such as manufacturing code, name, manufacturer, manufacturing date, etc. are normalized manually so that the same attributes have the same name.

[0132] (2) Entity Clustering

[0133] After ontology mapping and fusion, the mapping concept relationships (system, system), (subsystem, subsystem), (component, component), (part, part) are obtained. The entities under the corresponding concepts are obtained according to the concept labels, so as to obtain the cluster entity set. Taking the concept label "system" as an example, the entity set under the concept label "system" in the fault knowledge graph and the entity under the concept "system" in the assembly instance structure tree are calculated by similarity to obtain the aligned entity pair. The entity set under the concept label "system" in the fault knowledge graph entity is taken out as (bogie (CA358505_ZXJ_0001), bogie (CA360305_ZXJ_0004), ...), and the entity set under the concept label "system" in the assembly instance structure tree is (Xiamen 01 bogie, Fujian 01 bogie...)

[0134] (3) Structural attribute similarity calculation

[0135] The similarity calculation results of some common attributes of the bogie (CA358505_ZXJ_0001) in the fault knowledge graph entity and the Fujian 01 bogie and Xiamen 01 bogie in the assembly instance structure tree entity are shown in Table 4.

[0136] The calculation results of the overall structural attribute similarity between the entity Xiamen 01 bogie and the entity Fujian 01 bogie are shown in equations (6) and (7):

[0137] Xiamen 01 bogie:

[0138] Fujian 01 bogie:

[0139]

[0140] Table 4. Similarity calculation results of some common attributes of fault entities

[0141] Public property name Xiamen 01 bogie Fujian 01 bogie Manufacturing Code 1 0.2 Production Date 1 0.2 Axle load 1 0.8 Design speed 1 0.6

[0142] (4) Unstructured attribute similarity calculation

[0143] The similarity calculation results of some common attributes of the bogie (CA358505_ZXJ_0001) in the fault knowledge graph entity and the Fujian 01 bogie and Xiamen 01 bogie in the assembly instance structure tree entity are shown in Table 5.

[0144] Table 5. Similarity calculation results of some common attributes of fault entities

[0145] Property Name Xiamen 01 bogie Fujian 01 bogie Bogie type 1 1 Wheel type 1 1 Wheel tread type 1 1 Axle material 0.9 0.6 Brake disc material 0.8 0.7

[0146] The similarity threshold is set to 0.6, and the attribute pairs with similarity greater than 0.6 are taken out to calculate the unstructured attribute similarity. The total structured attribute similarity calculation results of the entity Xiamen 01 bogie and the entity Fujian 01 bogie are shown in formulas (8) and (9):

[0147] Xiamen 01 bogie:

[0148] Fujian 01 bogie:

[0149] (5) Entity comprehensive calculation

[0150] The weights of structured attribute similarity and unstructured attribute similarity are normalized to obtain the comprehensive similarity of the entity. The weight of unstructured attribute similarity is set to 0.8, and the weight of structured attribute similarity is set to 0.2. The comprehensive similarity calculation results of Xiamen 01 bogie and Fujian 01 bogie in the bogie (CA358505_ZXJ_0001) structure tree entity and assembly instance structure tree entity are shown in formula (10)-(11):

[0151] Xiamen 01 bogie: SimE(a,b)=0.8×2.3+0.2×0.94=2.028 (10),

[0152] Fujian 01 bogie: SimE(a,b)=0.8×0.15+0.2×0.86=0.292 (11).

[0153] The preset is set to 0.8, and the comprehensive similarity is sorted in descending order to obtain the optimal candidate sequence: Xiamen 01 bogie. The final aligned entity pair is obtained: (bogie (CA358505_ZXJ_0001), Xiamen 01 bogie).

[0154] Step 130: Mapping and fusion of stage domain structure tree ontology

[0155] 1) Entity Clustering

[0156] Taking the concept label "system" in the design instance structure tree and the assembly instance structure tree as an example, the clustering entities Ai (CRH380A, CRH3A, ...) and Bi (Xiamen 01 bogie, Fujian 01 bogie, ...) under the corresponding concept label are obtained.

[0157] 2) Similarity calculation

[0158] (1) The similarity calculation results of some common attributes of Xiamen 01 bogie and Fujian 01 bogie in the design instance structure tree entity CRH380A and the assembly instance structure tree entity are shown in Table 6.

[0159] Table 6 Similarity calculation results of some common attributes

[0160] Public property name Xiamen 01 bogie Fujian 01 bogie Meta node encoding 1 1 Module Encoding 1 0.7 Axle load 1 1 Adapt to ambient temperature 1 0.8

[0161] The overall structural attribute similarity calculation results are shown in formula (12)-formula (13):

[0162] Xiamen 01 bogie:

[0163] Fujian 01 bogie:

[0164] (2) Unstructured attribute similarity calculation

[0165] The similarity threshold is set to 0.6, and the attribute pair with the highest similarity is selected. If the similarity is greater than 0.6, it is classified as a similar attribute. The similarity calculation results of the unstructured attributes of Xiamen 01 bogie and Fujian 01 bogie in the design instance structure tree entity CRH380A and the assembly instance structure tree entity are shown in Table 7.

[0166] Table 7. Unstructured attribute similarity calculation results

[0167] Property Name Xiamen 01 bogie Fujian 01 bogie Bogie type 1 1 Wheel type 1 1 Wheel tread type 1 1 Axle material 0.9 0.6 Brake disc material 0.8 0.7

[0168] The attribute pairs with similarity greater than 0.6 are taken out, and the unstructured attribute similarity is shown in formula (14)-formula (15):

[0169] Xiamen 01 bogie:

[0170] Fujian 01 bogie:

[0171] (3) Entity comprehensive similarity calculation

[0172] The similarity weight of the unstructured attribute is set to 0.8, and the similarity weight of the structured attribute is set to 0.2. The comprehensive similarity calculation results of the entity "CRH380A" in the design instance structure tree and the entities Xiamen 01 bogie and Fujian 01 bogie in the assembly instance structure tree are shown in (16)-(17):

[0173] Xiamen 01 bogie: SimE(a,b)=0.8×2.3+0.2×0.94=2.028 (16),

[0174] Fujian 01 bogie: SimE(a,b)=0.8×2.06+0.2×0.86=1.82 (17).

[0175] The threshold is set to 0.8, and the comprehensive similarity of the entities is sorted in descending order to obtain the optimal candidate sequence sorting: (Xiamen 01 bogie, Fujian 01 bogie), and then the final aligned entity pair is obtained: (CRH380A, Xiamen 01 bogie), and a mapping relationship is established to obtain the triple (CRH380A, mapping, Xiamen 01 bogie).

[0176] The knowledge fusion method of the high-speed train product structure tree provided by the present invention organizes and analyzes high-speed train multivariate data, a stage domain structure tree and a domain structure tree, constructs a domain knowledge graph through an ontology model based on the domain structure tree, performs mapping alignment based on the domain knowledge graph, performs entity alignment based on the named entity recognition output of the instance data, and performs entity alignment based on the cluster entity using multi-information fusion similarity, fuses the heterogeneous domain knowledge graphs into the stage knowledge graph, and realizes the data fusion of the stage; at the same time, based on the stage domain structure tree, constructs a product coding structure tree to realize the ontology alignment of the stage domain structure tree, and realizes the mapping fusion of the instance data layer in the stage domain structure tree by establishing the mapping relationship between the entities of each stage, realizes the fusion of the knowledge graphs of each stage of the high-speed train, makes the intercommunication and interconnection relationship between the high-speed train field data clearer, and further solves the heterogeneity and redundancy problems of multi-stage data in the high-speed train field.

[0177] The high-speed train product structure tree knowledge fusion device provided by the present invention is described below. The high-speed train product structure tree knowledge fusion device described below and the high-speed train product structure tree knowledge fusion method described above can correspond to each other.

[0178] The present invention also provides a high-speed train product structure tree knowledge fusion device, Figure 6 is a schematic diagram of the structure of the knowledge fusion device for the high-speed train product structure tree provided by the present invention, such as Figure 6 As shown, the high-speed train product structure tree knowledge fusion device 200 includes: a construction module 201, a first fusion module 202, and a second fusion module 203, wherein:

[0179] Construction module 201, used to obtain organizational analysis data of high-speed train multivariate data, construct an ontology model, and construct a domain knowledge graph based on the ontology model;

[0180] The first fusion module 202 is used to fuse the domain knowledge graph into a stage knowledge graph, wherein the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by the named entity recognition of the instance data, and the entity fusion of the stage knowledge graph is achieved by aligning the entities based on clustered entities using multi-information fusion similarity;

[0181] The second fusion module 203 is used to perform mapping fusion of the stage domain structure tree ontology based on the stage domain structure tree of the organization analysis data, so as to realize the fusion of the knowledge graphs of each stage.

[0182] The high-speed train product structure tree knowledge fusion device provided by the present invention organizes and analyzes high-speed train multivariate data, a stage domain structure tree and a domain structure tree, and constructs a domain knowledge graph based on the domain structure tree through an ontology model, and maps and aligns the ontology concepts output by named entity recognition based on instance data based on the domain knowledge graph, and performs entity alignment based on clustering entities using multi-information fusion similarity, and fuses heterogeneous domain knowledge graphs into stage knowledge graphs to achieve stage data fusion; at the same time, based on the stage domain structure tree, a product coding structure tree is constructed to achieve ontology alignment of the stage domain structure tree, and by establishing a mapping relationship between entities of each stage, the mapping fusion of the instance data layer in the stage domain structure tree is achieved, and the fusion of the knowledge graphs of each stage of the high-speed train is achieved, so that the intercommunication and interconnection relationship between the high-speed train field data is clearer, and the heterogeneous and redundant problems of multi-stage data in the high-speed train field are further solved.

[0183] Optionally, the construction module 201 is specifically used to:

[0184] Based on the multivariate data of high-speed trains, organizational analysis data is obtained, wherein the organizational analysis data includes: data source, domain stage, domain structure tree and stage domain structure tree. The stage domain structure tree includes: product family main structure tree, design instance structure tree and assembly instance structure tree. The data sources include structured data, semi-structured data and unstructured data.

[0185] Based on the organizational analysis data, the context information, data types and ontology summary model of the data service are obtained to build the ontology model;

[0186] The global ontology is obtained according to the stage domain structure tree, the local ontology is obtained based on the global ontology, and the data is obtained in combination with the ontology model to build a domain knowledge graph.

[0187] Optionally, the first fusion module 202 is specifically configured to:

[0188] Based on the pre-acquired corpus data, a named entity recognition model is trained and acquired, wherein the corpus data is entity labeled and sequence labeled and then input into the named entity recognition model;

[0189] Based on the named entity recognition model, the ontology concepts corresponding to the entities in the domain knowledge graph are obtained;

[0190] Based on the comparison results between the ontology concepts corresponding to the entities in the domain knowledge graph and the ontology concepts in the domain structure tree, the ontology concept mapping relationship of the stage knowledge graph is obtained.

[0191] Optionally, the first fusion module 202 is specifically configured to:

[0192] Normalize the attributes of entities in the domain knowledge graph, where the attribute types include structured attributes and unstructured attributes;

[0193] Based on the ontology concept mapping relationship of the stage knowledge graph, cluster entities under the same ontology concept with mapping relationship are obtained;

[0194] Based on the unit and constraint matching of structured attributes, the similarity of structured attributes in clustered entities is determined using the minimum edit distance;

[0195] Based on the semantic feature vector of the unstructured attribute, the cosine similarity of the semantic feature vector is used to determine the similarity of the unstructured attribute of the clustering entity;

[0196] Based on the similarity of structured attributes and unstructured attributes, the comprehensive similarity of entities is determined to achieve entity alignment of the stage knowledge graph.

[0197] Optionally, the second fusion module 203 is specifically configured to:

[0198] Based on the coding attributes of the stage domain structure tree, a product coding structure tree is constructed to achieve ontology alignment;

[0199] Normalize the entity attributes of the stage domain structure tree;

[0200] Acquire the cluster entities under the ontology concept of the stage domain structure tree;

[0201] Based on the clustered entities, the comprehensive similarity of the entities is determined to obtain the mapping relationship between the entities of each domain stage under the ontology concept of the stage domain structure tree, wherein the comprehensive similarity includes the similarity of structured attributes and the similarity of unstructured attributes.

[0202] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device 300 may include: a processor 310, a communications interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communications interface 320 and the memory 330 communicate with each other through the communication bus 340. The processor 310 may call the logic instructions in the memory 330 to execute the knowledge fusion method of the high-speed train product structure tree, which includes:

[0203] Obtain organizational analysis data of high-speed train multivariate data, construct an ontology model, and build a domain knowledge graph based on the ontology model;

[0204] The domain knowledge graph is fused into a stage knowledge graph, where the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning the entities based on clustered entities using multi-information fusion similarity;

[0205] Based on the stage domain structure tree of the organizational analysis data, the mapping and fusion of the stage domain structure tree ontology is performed to achieve the fusion of the knowledge graphs of each stage.

[0206] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0207] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the high-speed train product structure tree knowledge fusion method provided by the above methods, the method includes:

[0208] Obtain organizational analysis data of high-speed train multivariate data, construct an ontology model, and build a domain knowledge graph based on the ontology model;

[0209] The domain knowledge graph is fused into a stage knowledge graph, where the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning the entities based on clustered entities using multi-information fusion similarity;

[0210] Based on the stage domain structure tree of the organizational analysis data, the mapping and fusion of the stage domain structure tree ontology is performed to achieve the fusion of the knowledge graphs of each stage.

[0211] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for integrating knowledge of a high-speed train product structure tree provided by the above methods is implemented, and the method comprises:

[0212] Obtain organizational analysis data of high-speed train multivariate data, construct an ontology model, and build a domain knowledge graph based on the ontology model;

[0213] The domain knowledge graph is fused into a stage knowledge graph, where the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning the entities based on clustered entities using multi-information fusion similarity;

[0214] Based on the stage domain structure tree of the organizational analysis data, the mapping and fusion of the stage domain structure tree ontology is performed to achieve the fusion of the knowledge graphs of each stage.

[0215] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0216] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0217] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A knowledge fusion method for a high-speed train product structure tree, characterized in that: include: Obtaining organizational analysis data of high-speed train multivariate data, constructing an ontology model, and constructing a domain knowledge graph based on the ontology model; The domain knowledge graph is merged into a stage knowledge graph, wherein the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning entities based on clustered entities using multi-information fusion similarity; Based on the stage domain structure tree of the organizational analysis data, mapping and fusion of the stage domain structure tree ontology are performed to achieve fusion of knowledge graphs of each stage; The stage domain structure tree based on the organization analysis data is used to map and fuse the ontology of the stage domain structure tree to achieve the fusion of knowledge graphs of each stage, including: Based on the coding attributes of the stage domain structure tree, a product coding structure tree is constructed to achieve ontology alignment; Normalize the entity attributes of the domain structure tree of the stage; Obtaining cluster entities under the ontology concept of the domain structure tree of the stage; Based on the clustered entities, the entity comprehensive similarity is determined to obtain the mapping relationship between the entities of each domain stage under the ontology concept of the stage domain structure tree, wherein the comprehensive similarity includes structured attribute similarity and unstructured attribute similarity.

2. The high-speed train product structure tree knowledge fusion method according to claim 1 is characterized in that: The obtaining of organizational analysis data of high-speed train multivariate data, constructing an ontology model, and constructing a domain knowledge graph according to the ontology model includes: Based on the multivariate data of high-speed trains, organizational analysis data is obtained, wherein the organizational analysis data includes: data source, domain stage, domain structure tree and stage domain structure tree, and the stage domain structure tree includes: product family main structure tree, design instance structure tree and assembly instance structure tree; Based on the organizational analysis data, obtain the context information and data types of data services and build ontology models; The global ontology is obtained according to the domain structure tree of the stage, the local ontology is obtained based on the global ontology, and the data is obtained in combination with the ontology model to construct a domain knowledge graph.

3. The high-speed train product structure tree knowledge fusion method according to claim 2 is characterized in that: The data sources include structured data, semi-structured data and unstructured data.

4. The high-speed train product structure tree knowledge fusion method according to claim 2 or 3 is characterized in that: The ontology fusion of the knowledge graph in this stage is achieved by mapping and aligning the ontology concepts output by named entity recognition based on instance data, including: Based on the pre-acquired corpus data, a named entity recognition model is trained and acquired, wherein the corpus data is entity labeled and sequence labeled and then input into the named entity recognition model; Based on the named entity recognition model, obtaining the ontology concepts corresponding to the entities in the domain knowledge graph; Based on the comparison result between the ontology concepts corresponding to the entities in the domain knowledge graph and the ontology concepts in the stage domain structure tree, the ontology concept mapping relationship of the stage knowledge graph is obtained.

5. The high-speed train product structure tree knowledge fusion method according to claim 4 is characterized in that: The entity fusion of the knowledge graph in the above stage is achieved by aligning entities based on clustered entities using multi-information fusion similarity, including: Normalizing the attributes of the entities in the domain knowledge graph, wherein the types of the attributes include structured attributes and unstructured attributes; Based on the ontology concept mapping relationship of the knowledge graph in the stage, cluster entities under the same ontology concept with a mapping relationship are obtained; Based on the unit and constraint matching of the structured attributes, using the minimum edit distance, determining the similarity of the structured attributes in the clustering entities; Based on the semantic feature vector of the unstructured attribute, the unstructured attribute similarity of the clustering entity is determined by using the cosine similarity of the semantic feature vector; Based on the structured attribute similarity and the unstructured attribute similarity, the comprehensive similarity of entities is determined to achieve entity alignment of the knowledge graph at the stage.

6. A device for implementing the high-speed train product structure tree knowledge fusion method as claimed in any one of claims 1 to 5, characterized in that: include: A construction module is used to obtain organizational analysis data of high-speed train multivariate data, construct an ontology model, and construct a domain knowledge graph based on the ontology model; A first fusion module is used to fuse the domain knowledge graph into a stage knowledge graph, wherein the ontology fusion of the stage knowledge graph is achieved by mapping and aligning the ontology concepts output by named entity recognition of instance data, and the entity fusion of the stage knowledge graph is achieved by aligning entities based on clustered entities using multi-information fusion similarity; The second fusion module is used to perform mapping fusion of the stage domain structure tree ontology based on the stage domain structure tree of the organization analysis data, so as to realize the fusion of the knowledge graphs of each stage.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it implements the high-speed train product structure tree knowledge fusion method as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the high-speed train product structure tree knowledge fusion method as described in any one of claims 1 to 5 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the high-speed train product structure tree knowledge fusion method as described in any one of claims 1 to 5 is implemented.

Images