A method for constructing semantic network of collaborative data space of multi-value chain in manufacturing industry

Abstract

The present invention provides a method for constructing a semantic network of a collaborative data space for multiple value chains in the manufacturing industry, including: integrating the data resources of the manufacturing enterprise itself, the data resources of the upstream enterprises, and the data resources of the downstream enterprises to form a collaborative data space for multiple value chains of the enterprise; creating a domain ontology library; constructing a semantic network of the collaborative data space for multiple value chains according to the domain ontology library; performing a semantic network analysis on the semantic network to obtain a semantic network analysis result; determining whether the semantic network analysis result reaches a preset indicator; if not, repeating the above steps. Through this technical solution, a semantic network of a collaborative data space for multiple value chains in the manufacturing industry can be constructed to provide effective support for digital resource aggregation, knowledge resource discovery, knowledge services, and intelligent decision-making in smart factories.

Description

Technical Field

[0001] The present invention relates to the technical field of digital resource management, and in particular to a method for constructing a semantic network of a multi-value chain collaborative data space in a manufacturing industry. Background Art

[0002] The rise of new generation information technologies such as the Internet, big data, and artificial intelligence has provided important opportunities for my country's manufacturing enterprises to overtake others. However, in the process of integration and application of the above information technologies in manufacturing enterprises, there are problems in the aggregation and identification of digital resources caused by large data volume, low data value density, large data object granularity, and blurred digital resource boundaries, which hinder the utilization of digital resources, data asset construction, and knowledge value discovery of manufacturing enterprises. Therefore, how to effectively integrate the digital resources of manufacturing enterprises, and then discover the semantic associations between digital resources, effectively aggregate the upstream and downstream data of the manufacturing enterprise supply chain and the enterprise's own data based on semantics, and discover the key knowledge nodes and knowledge resources in the semantic network of multi-value chain collaborative data space, support the knowledge innovation and intelligent decision-making of manufacturing enterprises, which has important theoretical value and practical significance.

[0003] From the perspective of domestic and foreign research, the research on digital resource aggregation mainly focuses on metadata-based aggregation, semantic-based aggregation, big data analysis and measurement aggregation. Among them, semantic-based aggregation can deeply fit with the multi-value chain collaboration process of manufacturing enterprises, and can tap the value of data, improve enterprise competitiveness for knowledge innovation, and solve the explicit needs of bottleneck problems. The goal of semantic-based digital resource aggregation is to discover the semantic associations between digital resources, and through data integration, semantic recognition, and semantic mining, to integrate multi-source heterogeneous digital resources in multiple dimensions and levels to form a three-dimensional knowledge network that brings together domain themes, concept themes, data content, and data object entities.

[0004] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0005] Although there have been studies that apply semantic theory and technology to the field of digital resource aggregation and knowledge discovery, there are still systematic defects in the theoretical research on semantic digital resource aggregation, and there are difficulties such as the difficulty of integrating multi-value collaborative digital resources of manufacturing enterprises, low value density, poor aggregation effect, and insufficient support for knowledge innovation. Therefore, it is difficult to carry out practical applications in the field of digital resource aggregation in manufacturing enterprises, resulting in a low level of digital resource utilization in manufacturing enterprises. Therefore, how to realize a semantic network of multi-value chain collaborative data space suitable for manufacturing enterprises by expanding the theory and aggregation methods of semantic digital resources is a problem that needs to be solved. Summary of the invention

[0006] The embodiment of the present invention provides a method for constructing a semantic network of a multi-value chain collaborative data space in the manufacturing industry, which is used to solve the problem of low digital resource utilization level of manufacturing enterprises existing in the prior art.

[0007] To achieve the above-mentioned purpose, an embodiment of the present invention provides a method for constructing a semantic network of a multi-value chain collaborative data space in a manufacturing industry, comprising:

[0008] Integrate the data resources of manufacturing enterprises themselves, upstream enterprises and downstream enterprises to form a multi-value chain collaborative data space for enterprises. The data resources refer to all data of enterprises in the production value chain, manufacturing value chain, sales value chain, service value chain and management value chain.

[0009] Create a domain ontology library;

[0010] Constructing a semantic network of the multi-value chain collaborative data space according to the domain ontology library;

[0011] Performing semantic network analysis on the semantic network to obtain a semantic network analysis result;

[0012] Determine whether the semantic network analysis result reaches a preset indicator;

[0013] If not, repeat the above steps.

[0014] The above technical solution has the following beneficial effects:

[0015] The present invention constructs a digital resource aggregation model for the collaborative data space of multiple value chains of manufacturing enterprises, proposes a process and method for aggregating digital resources based on the knowledge structure, inherits multiple knowledge concepts and transcends a single structure, and forms effective support for intelligent decision-making of manufacturing enterprises; secondly, considering the connotation extension of the concept of multi-value chain collaboration of manufacturing enterprises, and combining the relevant conceptual elements of multi-value chain collaborative supply chain enterprises of manufacturing enterprises, a method for calculating the field relevance of the collaborative data space of multiple value chains of manufacturing enterprises is proposed; further, based on the concept degree distribution and DS evidence theory, trust function and likelihood function are introduced, and a method for discovering semantic network nodes in the collaborative data space of multiple value chains of manufacturing enterprises is proposed, forming a systematic digital resource aggregation and identification method. In summary, the method for constructing the semantic network of the collaborative data space of multiple value chains of the manufacturing industry proposed in this application provides effective support for digital resource aggregation, knowledge resource discovery, knowledge services, and intelligent decision-making of smart factories. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only 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.

[0017] Figure 1 It is a flow chart of a method for constructing a semantic network of a multi-value chain collaborative data space in a manufacturing industry according to an embodiment of the present invention;

[0018] Figure 2 It is a technical roadmap of the digital resource aggregation identification model of the multi-value chain collaborative data space of the manufacturing industry in the embodiment of the present invention;

[0019] Figure 3 It is a schematic diagram of some domain ontology in a specific embodiment;

[0020] Figure 4 It is a semantic network cloud diagram of a certain electric power equipment Co., Ltd. in a specific embodiment. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0022] like Figure 1 As shown, an embodiment of the present invention provides a method for constructing a semantic network of a multi-value chain collaborative data space in a manufacturing industry, including:

[0023] S101. Integrate the data resources of the manufacturing enterprise itself, the data resources of upstream enterprises and the data resources of downstream enterprises to form a multi-value chain collaborative data space for the enterprise. The data resources refer to all the data of the enterprise in the production value chain, manufacturing value chain, sales value chain, service value chain and management value chain;

[0024] S102, creating a domain ontology library;

[0025] S103, constructing a semantic network of the multi-value chain collaborative data space according to the domain ontology library;

[0026] S104, performing semantic network analysis on the semantic network to obtain a semantic network analysis result;

[0027] S105, determining whether the semantic network analysis result reaches a preset indicator;

[0028] S106. If not, repeat the above steps S101-S105.

[0029] Furthermore, the method for constructing the semantic network of the multi-value chain collaborative data space of the manufacturing industry also includes:

[0030] S107: Discover key nodes from the semantic network.

[0031] Furthermore, the step S103 includes:

[0032] S1031, extracting valid information from the data resources;

[0033] S1032, semantically representing the valid information based on a domain ontology library and using natural language processing technology;

[0034] S1033, calculating semantic similarity between every two pieces of valid information;

[0035] S1034. Determine the nodes of the semantic network and the relationship between the nodes according to the result of the semantic representation and the result of the semantic similarity calculation, so as to form the semantic network of the multi-value chain collaborative data space;

[0036] S1035: Eliminate inappropriate associations from the semantic network.

[0037] Furthermore, the step S1035 specifically includes:

[0038] For any two associated nodes in the semantic network, perform the following operations:

[0039] S10351. Calculate the spatial domain correlation between two current associated nodes;

[0040] S10352: If the calculated spatial domain correlation is less than a preset first threshold, it is determined that the current two associated nodes are inappropriately associated;

[0041] S10353. Disconnect the inappropriate association.

[0042] Furthermore, the step S1032 specifically includes:

[0043] S10321. Use L(G,L,G) triples to semantically represent the valid information.

[0044] Furthermore, the step S104 includes:

[0045] S1041, calculating the semantic network density of the semantic network;

[0046] S1042, calculating the semantic network reachability of the semantic network;

[0047] S1043, calculating the semantic network centrality of the semantic network;

[0048] The preset indicators include a preset value of semantic network density, a preset value of semantic network reachability and a preset value of semantic network centrality;

[0049] The step S105 is specifically as follows:

[0050] S1051. Determine whether the semantic network density, the semantic network reachability, and the semantic network centrality simultaneously reach their respective corresponding preset values.

[0051] Furthermore, the step S107 is performed in the following manner:

[0052] MPC(i)=M(h)-M(l)=m i (h)-m i (l)

[0053] Among them, M is the recognition framework, h and l are a pair of mutually exclusive elements, h means that fi and gi have a high degree of support for the importance of the node, l means that fi and gi have a low degree of support for the importance of the node, θ means the recognition framework, i means the node, m i (θ) represents the probability that the importance of node i cannot be determined.

[0054]

[0055]

[0056] When the value of MPC(i) is greater than a preset second threshold, the current node is determined to be a key node.

[0057] The multi-value chain collaboration of manufacturing enterprises refers to the process in which manufacturing enterprises and industrial clusters coordinate multiple basic business activities such as manufacturing, production, marketing, and services within and between enterprises and industrial clusters, as well as auxiliary support activities such as personnel, finance, and R&D, and realize the creation and value-added of information value, data value, knowledge value, and green value through subject collaboration, data collaboration, data service, and knowledge service collaborative innovation coupling. Specifically, it is the collaboration of production and business activities such as R&D, design, supply, production, marketing, and service, and also includes the integration of enterprise knowledge chain management and value chain management, and the value collaboration of data value, information value, knowledge value, and ecological value creation and value-added for enterprises, industries, and industrial clusters. In turn, it effectively improves the service capabilities of enterprises, increases the profits of knowledge value-added, and enables safer data storage and sharing, and lowers the acquisition cost. The multi-value chain collaboration data space of manufacturing enterprises is a collection of data and relationships related to the multi-value chain collaboration subjects of manufacturing enterprises, including subjects, data sets, data services, and knowledge service elements.

[0058] The present invention proposes for the first time a new idea for value chain management and data space construction for upstream and downstream enterprises in the manufacturing industry based on the concept of multi-value chain collaboration and data space in the manufacturing industry; the present invention constructs a digital resource aggregation model for the multi-value chain collaborative data space of manufacturing enterprises based on semantic theory. Existing research has never involved in the deep combination of enterprise and upstream and downstream resources. The present invention orderly organizes and mines resource relationships, mines the knowledge contained in the data space, and meets the multi-granularity and fine-grained data consumption needs of users in the value chain of manufacturing enterprises; the present invention proposes for the first time a method for calculating the relevance of the multi-value chain collaborative data space of manufacturing enterprises, and further designs a semantic network node discovery algorithm for the multi-value chain collaborative data space of manufacturing enterprises, which is used to improve the aggregation and identification accuracy of data resources in the multi-value chain collaborative data space of manufacturing enterprises, and provide support for intelligent manufacturing and smart decision-making of manufacturing enterprises from a macro perspective.

[0059] Based on semantic theory, this paper studies the digital resource aggregation and identification method of the multi-value chain collaborative data space of manufacturing enterprises, and proposes a digital resource aggregation and identification algorithm model for the multi-value chain manufacturing data space of manufacturing enterprises. First, a digital resource aggregation model of the multi-value chain collaborative data space of manufacturing enterprises is constructed, and the process and method of aggregating digital resources based on the knowledge structure are proposed, which inherits the knowledge concepts multiple times and transcends the single structure to form an effective support for intelligent decision-making of manufacturing enterprises; secondly, considering the connotation extension of the multi-value chain collaborative concept of manufacturing enterprises, and combining the relevant conceptual elements of the multi-value chain collaborative supply chain enterprises of manufacturing enterprises, a method for calculating the relevance of the multi-value chain collaborative data space of manufacturing enterprises is proposed; further, based on the concept degree distribution and DS evidence theory, trust function and likelihood function are introduced, and a method for discovering semantic network nodes in the multi-value chain collaborative data space of manufacturing enterprises is proposed, forming a systematic digital resource aggregation and identification method.

[0060] 1) Construct a digital resource aggregation model for multi-value chain collaborative data space of manufacturing enterprises:

[0061] Step 1: Integrate manufacturing enterprise data lake, cloud storage space, corporate website, WeChat public account, local application software, App program and other application data

[0062] Step 2: Build a domain ontology, perform semantic processing on digital resources, calculate domain relevance, build a semantic network, analyze semantic network indicators, and discover key nodes.

[0063] The digital resource aggregation model framework is as follows Figure 2 As shown:

[0064] 2) Aggregation and semanticization of digital resources in the collaborative data space of multiple value chains of manufacturing enterprises:

[0065] The concepts involved in the production and operation process of manufacturing enterprises, as well as the relationships between concepts, constitute the basis for the aggregation of digital resources. Based on data collection, cleaning, extraction, mining and other technologies, digital resource instances are identified.

[0066] Step 1: Use L(G,L,G) triples to describe concept relationships, where G is the concept relationship and L is the concept. The semantic relationship can be described in fine-grained terms through semantic identity, semantic similarity, semantic oppositeness, hyponymy, and part-whole semantic relationships.

[0067] Step 2: After semantic representation, the semantic similarity algorithm is further used to convert the digital resources composed of natural language fragments into structured information that can be processed by computers. The vector space model is used to calculate the feature vector of the natural language text, and the semantic similarity R (b i ,b j ), assuming that the feature vectors of two natural language segments are

[0068]

[0069] Eigenvector angle R(b i ,b j ) The inner product calculation is shown in Formula 1, the Jaccard coefficient method is shown in Formula 2, the Dice coefficient method is shown in Formula 3, and the cosine similarity method is shown in Formula 4, among which the cosine similarity algorithm is widely used.

[0070]

[0071]

[0072]

[0073]

[0074] 3) Calculate the spatial correlation of multi-value chain collaborative data of manufacturing enterprises

[0075] First, the semantic relevance and consistency algorithms are used to eliminate irrelevant concepts (i.e., inappropriate associations). The semantic relevance algorithm is shown in Formula 5.

[0076]

[0077] in Fr is the concept frequency. Considering the extension of the concept connotation in the multi-value chain collaboration process, and combining the relevant concept elements of the multi-value chain collaborative supply chain enterprises of manufacturing enterprises, this application proposes a spatial domain relevance algorithm for multi-value chain collaborative data of manufacturing enterprises based on the semantic relevance algorithm as shown in Formula 6:

[0078]

[0079] in Fr is the frequency of the concept appearing in the enterprise's single value chain data set. The calculation formula for the improved digital resource domain consistency of the multi-value chain collaborative data space of the manufacturing enterprise is shown in Formula 7:

[0080]

[0081] Then the semantic extraction concept formula is shown in Formula 8, where α, β∈(0, 1), and finally all relevant concepts of the domain ontology of the multi-value chain collaborative data space of manufacturing enterprises can be obtained.

[0082] LO I =α×LS(d i ,W)+β×LC(d i ,W) (8)

[0083] 4) Construction and structural analysis method of semantic network of multi-value chain collaborative data space of manufacturing enterprises

[0084] <1> , semantic network density

[0085] The semantic network density algorithm is shown in Formula 9, where R is the number of semantic network node connections, N is the number of concepts in the semantic network, and i is a value chain collaborative semantic network. The greater the network density, the more associations there are between concepts.

[0086] M i =R i / N i ×(N i -1) (9)

[0087] <2> , Semantic Web Accessibility

[0088] The semantic network reachability algorithm is shown in Formula 10, where U is the number of unreachable point pairs. This indicator reflects the fragility and robustness of the network. If many nodes in the network are connected through a certain concept or several concepts, it means that the semantic network is more dependent on this node and the robustness is not high.

[0089] C i =1-2U i / N i ×(N i -1) (10)

[0090] <3> , semantic network centrality

[0091] Semantic network centrality indicators include degree centrality, betweenness centrality and closeness centrality. The degree centrality indicator reflects the degree to which a concept is at the center of the semantic network, as shown in Formula 11, where n j N is the number of connections established between concept j in value chain i, i is the number of semantic network nodes in value chain i.

[0092]

[0093] Betweenness centrality measures the extent to which a concept is in the middle of other point pairs, as shown in Formula 12.

[0094] in b kl (j) = d kl (j) / d kl , j≠k≠l, k<l, d kl is the number of shortest paths between points k and 1, d kl (j) is d kl The number of shortest paths passing through concept node j in .

[0095]

[0096] The closeness centrality reflects the degree of independence of the concept in the value chain collaboration process, as shown in Formula 13, d kl is the number of shortest paths between concepts k and l in the value chain network of i.

[0097]

[0098] 5) Node discovery method of semantic network in multi-value chain collaborative data space of manufacturing enterprises

[0099] Based on the analysis of semantic network structure, the discovery of important nodes in semantic network is further carried out. Based on the concept degree distribution, the core concept nodes are screened. Based on the DS evidence theory, the trust function and likelihood function are introduced. The semantic network node discovery method of multi-value chain collaborative data space of manufacturing enterprises is proposed. The calculation is as follows:

[0100] <1> , Theoretical function definition

[0101] Assuming that there are n elements in the sample space M, the number of M subsets is 2n, 2M represents the subset, and the M subset A is mapped to [0, 1], and the value is D(A). Trust function Bel: 2 M →[0, 1], and Trust function PI: 2 M →[0, 1], then the degree of confidence that A is true Right now

[0102] Then, the final trust function is calculated by using the synthesis rule to solve the problem of two different trust functions. Assume that the evidence H1 distribution function is d1 and the focal element is C under the framework M. i , the evidence H2 distribution function is d2, the focal element is D j , then the synthesis rule is shown in formula 14.

[0103]

[0104] <2> , important node modeling

[0105] The clustering coefficient can measure the degree of clustering of semantic network concepts. For the semantic directed network of manufacturing enterprises, the clustering coefficient formula is shown in 15, where F i Represents the number of connections between neighbor concept nodes.

[0106]

[0107] Under the recognition framework M, C i , D j They are defined as concept node degree and concept node clustering coefficient respectively, and the concept node importance evaluation result is obtained based on formula 14.

[0108] <3> , Identification of important nodes (i.e. key nodes)

[0109] Based on formulas 14 and 15, a semantic network core node identification algorithm MPC for collaborative data space of multi-value chains of manufacturing enterprises is proposed. Assuming that the semantic directed network contains N nodes and M edges, the directed network E = E(G, B), where G is the concept node set, B is the edge set, G = (G1, G2, ... G N ), B={(G i , -G j)|0<i<max,0<j<max}, and G i and G j If there is a connection relationship and an inheritance relationship between edges, the sum of the degree of node i and the degree of its neighboring nodes can be expressed by formula 16. i C i The set of adjacent points, C α It is C i Adjacent point, C ij For node C i Spend.

[0110]

[0111] Define fi as shown in Formula 17, S i It is node C i The clustering coefficient, Dg j It is node C j And the sum of the neighbor node degrees.

[0112]

[0113] Define function gj as shown in formula 18

[0114]

[0115] The maximum and minimum values ​​of the local information functions dg and g of the network nodes are calculated as shown in formulas 19 to 22:

[0116] dg min =minxdg1,dg2,...,dg N} (19)

[0117] dg Max = max{dg1, dg2, ..., dg N} (20)

[0118] g min =min{g1, g2, ..., g N} (twenty one)

[0119] g Max = max{g1, g2, ..., g N} (twenty two)

[0120] Define the framework as M(h, l), where the mutually exclusive elements h and l represent the high or low support degree of fi and gi for the importance of the node. and They respectively represent the node degree and the degree to which information supports the importance of the node, and their formulas are 23 to 28, where 0<δ, γ<1.

[0121]

[0122]

[0123]

[0124]

[0125] α=dg M -dg m +2δ (27)

[0126] β=g M -g m +2γ (28)

[0127] Therefore, i and g i The basic probability distribution function is as follows:

[0128]

[0129]

[0130]

[0131]

[0132] but

[0133] M(i)=(m i (h), m i (l), m i (θ)) (33)

[0134]

[0135]

[0136] Thus, we can obtain the important node evaluation indicators of the semantic network of manufacturing enterprises:

[0137] MPC(i)=M(h)-M(l)=m i (h)-m i (l) (36)

[0138] In order to specifically demonstrate the method described in this application, this application uses the data obtained by a certain electric power equipment company as a sample for simulation. The following is an example to illustrate the effect of the technical solution of this application:

[0139] This application uses the product data of a certain electric power equipment company as the initial data source, and then integrates multi-value chain collaborative data, obtains domain concepts, calculates domain relevance and consistency, builds a semantic network, and mines key concept nodes. The integrated data sample is shown in Table 1.

[0140]

[0141] Table 1. Some data integration samples based on a company's "QEFG￡ series environmentally friendly gas switch cabinet"

[0142] This specific example constructs a domain ontology dataset of equipment manufacturing enterprises by constructing a semantic network for power equipment manufacturing, effectively analyzes the word frequency statistics, and uses the gephi tool and semantic network analysis methods to draw a semantic network cloud diagram. For the extraction process of a subclass in the design stage, the irrelevant concepts involved are eliminated using the domain relevance and consistency formulas, and the semi-automatic domain ontology construction method automatically generated by the program is used to finally construct the domain ontology dataset of equipment manufacturing enterprises. The word frequency statistics results are shown in Table 2, and the attached Figure 3 This is a schematic diagram of some domain ontologies in this specific embodiment.

[0143]

[0144]

[0145] Table 2. Word frequency statistics of "QEFG￡ environmentally friendly gas switch cabinet"

[0146] Correspondingly, the network cloud diagram is as follows Figure 4 shown.

[0147] From the perspective of the overall network structure, the network diameter is 25, the density is 0.02, the network structure is relatively concentrated, and the concept ontology link depth is large, indicating that the semantic network system constructed in this paper is clear, and the enterprise has a high degree of semantic-based value chain collaboration. From the perspective of the intermediate centrality index, the intermediate centrality of materials is high, indicating that this concept is a bridge for the multi-value chain collaboration of the enterprise. From the perspective of proximity centrality, the proximity centrality of insulating gas is low, indicating that this concept has strong independence in the multi-value chain collaboration process of the enterprise.

[0148] The present invention identifies key nodes in power equipment manufacturing. The comparison of the purity of the three methods, K-means, TCUSS and the MMPC method in this paper, proves that the constructed semantic network of the multi-value chain collaborative data space of the manufacturing enterprise is closely connected, has clear semantics and a clear system. The proposed method has good aggregation effect and high purity. The results of clustering analysis of the semantic network using K-means clustering and based on random initial cluster centers are shown in Tables 3 and 4:

[0149] category D1 D2 D3 D4 D5 Design Methodology 1021 865 452 350 56 Process Manufacturing 584 256 154 105 86 product 2014 1586 568 154 68 Material 1247 1025 564 204 85 Standard Parts 125 98 46 32 16

[0150] Table 3. Clustering results of some fields with K core of 5

[0151] The K core value is set to 10, and clustering is performed again, with a purity of 0.814. The clustering results of some fields are shown in Table 4. Comparing the data clustering results, with the increase of K core, the clustering accuracy is significantly improved.

[0152] category D6 D7 D8 D9 D10 Design Methodology 41 32 21 16 10 Process Manufacturing 75 68 61 32 5 product 62 45 35 18 8 Material 75 66 52 21 6 Standard Parts 15 11 9 5 3

[0153] Table 4. Clustering results when K core is 10

[0154] The comparison results of the aggregation algorithm purity are shown in Table 5:

[0155] Number of clusters K-means TCUSS MMPC 6 0.156 0.412 0.465 7 0.364 0.545 0.551 8 0.465 0.580 0.754 9 0.480 0.598 0.783 10 0.496 0.514 0.814

[0156] Table 5. Comparison results of the purity of aggregation algorithms

[0157] Finally, Figure 3 The indicators of the experimental data set are calculated, and the importance of the nodes is ranked from high to low. Some results are shown in Table 6. The order of importance of the concept nodes in Table 6 is from high to low: environmental protection, ring network cabinet, stainless steel, detection, gas, power distribution, switch, electrical, insulation, and automatic.

[0158] node <![CDATA[m i (h)]]> <![CDATA[m i (l)]]> MPC(i) Stainless steel 0.6626 0.1778 0.4849 Ring main unit 0.7030 0.0548 0.6482 automatic 0.1165 0.7739 -0.6574 gas 0.6182 0.3938 0.2244 switch 0.6781 0.6162 0.0618 insulation 0.0539 0.4020 -0.3481 Environmental friendly 0.9918 0.2410 0.7507 electric 0.6358 0.9203 -0.2845 Detection 0.7309 0.4567 0.2742 power distribution 0.5836 0.4652 0.1184

[0159] Table 6. Results of MPC data indicators for some nodes

[0160] The disclosed embodiments are described above to enable any person skilled in the art to implement or use the present invention. Various modifications of these embodiments are obvious to those skilled in the art, and the general principles defined herein may also be applied to other embodiments without departing from the spirit and scope of the present disclosure. Therefore, the present disclosure is not limited to the embodiments given herein, but is consistent with the broadest scope of the principles and novel features disclosed in this application.

[0161] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for constructing a semantic network of a multi-value chain collaborative data space in the manufacturing industry, characterized in that: include: Integrate the data resources of manufacturing enterprises themselves, upstream enterprises and downstream enterprises to form a multi-value chain collaborative data space for enterprises. The data resources refer to all data of enterprises in the production value chain, manufacturing value chain, sales value chain, service value chain and management value chain. The data resources are the manufacturing enterprise data lake, cloud storage space, enterprise website, WeChat public account, local application software or App program. Create a domain ontology library; Constructing a semantic network of the multi-value chain collaborative data space according to the domain ontology library; Performing semantic network analysis on the semantic network to obtain a semantic network analysis result; Determine whether the semantic network analysis result reaches a preset indicator; If not, repeat the above steps; discovering key nodes from the semantic network; The step of constructing the semantic network of the multi-value chain collaborative data space according to the domain ontology library specifically includes: Extracting valid information from the data resource; Based on the domain ontology library, the L(G,L,G) triple is used to semantically represent the effective information, where L is the concept involved in the production and operation process of the manufacturing enterprise, G is the relationship between the concepts, and the semantic relationship is finely described by the same semantics, similar semantics, opposite semantics, upper and lower semantics, or part-whole semantics relationship; After the semantic representation, a semantic similarity calculation is performed between each two of the valid information, specifically including: converting the data resources composed of the natural language text into structured information that can be processed by a computer, using a vector space model to calculate the natural language text feature vector, and calculating the semantic similarity through the inner product of the feature vector, or the Jaccard coefficient, or the cosine similarity, or the Dice coefficient; According to the result of semantic representation and the result of semantic similarity calculation, the nodes of the semantic network and the relationship between the nodes are determined, so as to form the semantic network of the multi-value chain collaborative data space; Afterwards, removing inappropriate associations from the semantic network of the multi-value chain collaborative data space; The step of removing inappropriate associations from the semantic network of the multi-value chain collaborative data space specifically includes: for any two associated nodes in the semantic network of the multi-value chain collaborative data space, performing the following operations: calculating the spatial domain correlation between the two current associated nodes; if the calculated spatial domain correlation is less than a first threshold, determining that the current two associated nodes are inappropriately associated; and disconnecting the inappropriate association; The calculation formula of the spatial domain correlation is: in, Fr is the frequency of the concept in the enterprise’s single value chain data set; According to the calculation result of the spatial domain relevance, the digital resource domain consistency of the multi-value chain collaborative data space is calculated. The calculation formula of the digital resource domain consistency of the multi-value chain collaborative data space of the manufacturing enterprise is: Then, all relevant concepts of the domain ontology are obtained according to the semantic extraction concept formula, and the semantic extraction concept formula is: THE I =α×LS(d i ,W)+β×LC(d i ,W), Among them, α, β∈(0,1); The performing semantic network analysis on the semantic network to obtain a semantic network analysis result includes: The semantic network density of the semantic network is calculated, and the semantic network density algorithm is: M i =R i / N i ×(N i -1); Among them, R is the number of semantic network node connections, N is the number of concepts in the semantic network, and i is a value chain collaborative semantic network. The greater the network density, the more associations there are between concepts. The semantic network reachability of the semantic network is calculated, and the semantic network reachability algorithm is: C i =1-2U i / N i ×(N i -1); Where U is the number of unreachable network node pairs; Calculating the semantic network centrality of the semantic network; The semantic network centrality index includes degree centrality, betweenness centrality and closeness centrality. The calculation formula of degree centrality is: where n j N is the number of connections established between concept j in value chain i, i is the number of semantic network nodes in value chain i; The calculation formula of the betweenness centrality is: in b kl (j) = d kl (j) / d kl , j≠k≠l, k <l,d kl is the number of shortest paths between points k and l, d kl (j) is d kl The number of shortest paths passing through concept node j in ; The calculation formula of the closeness center is: where d kl is the number of shortest paths between concepts k and l in the value chain network of i; The preset indicators include a preset value of semantic network density, a preset value of semantic network reachability, and a preset value of semantic network centrality; judging whether the semantic network density, the semantic network reachability, and the semantic network centrality reach their respective corresponding preset values ​​at the same time; The key nodes are found from the semantic network by using the following formula: MPC(i)=M(h)-M(l)=m i (h)-m i (l), Among them, M is the key node identification framework, h and l are a pair of mutually exclusive elements, h means that fi and gi have a high degree of support for the importance of the node, l means that fi and gi have a low degree of support for the importance of the node, θ means the node importance probability identification framework, i means the node, m i (θ) represents the probability that the importance of node i cannot be determined. When the value of MPC(i) is greater than the second threshold, the current node is determined to be a key node.

Images