Industrial network knowledge graph construction method and device, electronic equipment, storage medium and program product

Abstract

The application provides an industry network knowledge graph construction method and device, electronic equipment, storage medium and program product. The method comprises the following steps: determining an undirected graph matrix, wherein the undirected graph matrix comprises entity information elements and undirected edge elements, and the undirected edge elements represent the undirected relationship between the entity information elements; obtaining a to-be-decomposed matrix according to the undirected edge elements of the undirected graph matrix, wherein the to-be-decomposed matrix comprises a first matrix area, and the target undirected edge elements in the undirected edge elements are arranged in the first matrix area; performing matrix decomposition processing on the decomposition nodes in the to-be-decomposed matrix to obtain sub-matrices and decomposition nodes, wherein the decomposition nodes are determined according to the target decomposition degree and the first decomposition degree of the to-be-decomposed matrix, and the target decomposition degree of the to-be-decomposed matrix is determined according to the target undirected edge elements in the to-be-decomposed matrix; and obtaining an industry network knowledge graph according to the sub-matrices and the decomposition nodes. The method of the application achieves the technical effect of improving the reliability of industrial knowledge data.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, electronic device, storage medium, and program product for constructing an industrial network knowledge graph. Background Technology

[0002] With the acceleration of economic globalization and the rapid development of information technology, the complexity and diversity of industrial chains are constantly increasing. How to optimize industrial chains, how to extract valuable data from massive amounts of data, and how to efficiently conduct data mining and analysis have become pressing issues in this field. Currently, the technology of mining knowledge graphs of internal production relations based on industrial network data to construct truly quantifiable industrial network knowledge graphs is still a technological gap. Therefore, methods for constructing industrial network knowledge graphs have become a promising direction for application.

[0003] In existing technologies, the industrial chain is commonly used to describe industrial relationship structures. The industrial chain can serve as a form of industrial knowledge data, encompassing four dimensions: value chain, enterprise chain, supply and demand chain, and spatial chain. These four dimensions form the industrial chain through a process of mutual alignment and equilibrium. Technical personnel sift through massive amounts of industrial data to extract industrial relationship information, manually analyzing the upstream and downstream relationships and the information across the four dimensions. This analysis then summarizes the findings to form an industrial chain describing these relationships. Upstream links in the industrial chain can supply products or services to downstream links, while downstream links can provide feedback to upstream links.

[0004] However, existing technologies suffer from the technical problem of low reliability of industry knowledge data. Summary of the Invention

[0005] This application provides a method, apparatus, electronic device, storage medium, and program product for constructing an industrial network knowledge graph, in order to solve the technical problem of low reliability of industrial knowledge data.

[0006] Firstly, this application provides a method for constructing an industrial network knowledge graph, applied to a construction platform, including:

[0007] Determine the undirected graph matrix, which includes entity information elements and undirected edge elements, where the undirected edge elements represent the undirected relationships between entity information elements;

[0008] Based on the undirected edge elements of the undirected graph matrix, the matrix to be decomposed is obtained, wherein the matrix to be decomposed includes a first matrix region, and the target undirected edge elements are arranged in the undirected edge elements within the first matrix region;

[0009] The decomposition nodes in the matrix to be decomposed are subjected to matrix decomposition to obtain submatrices and decomposition nodes. The decomposition nodes are determined based on the target decomposition degree and the first decomposition degree of the matrix to be decomposed. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed.

[0010] Based on the submatrix and decomposed nodes, the industrial network knowledge graph is obtained.

[0011] Optionally, the method described above determines the undirected graph matrix, including:

[0012] Acquire industrial network data, which includes entity information within the industrial network and the relationships between those entities.

[0013] Based on industry network data, a directed graph matrix is ​​constructed, which includes entity information elements and directed edge elements. Entity information elements represent entity information, and directed edge elements represent the directed relationships between entity information elements.

[0014] Determine the undirected graph matrix based on the entity information elements and directed edge elements in the directed graph matrix.

[0015] Optionally, the above method constructs a directed graph matrix based on industry network data, including:

[0016] Encode the entity information in the industrial network data to obtain the data encoding of the entity information;

[0017] Based on the data encoding of entity information, determine the first and second dimension information of the initial directed graph matrix;

[0018] The directed graph matrix is ​​obtained based on the first and second dimension information of the initial directed graph matrix and the relationship between entity information in the industrial network data.

[0019] Optionally, the method described above, based on the undirected edge elements of the undirected graph matrix, yields the matrix to be decomposed, including:

[0020] Based on the element attributes of the undirected edge elements in the undirected graph matrix, determine the target undirected edge element, as well as the first and second dimension information of the target undirected edge element. The element attributes of the undirected edge element include attributes that represent the existence of undirected relations.

[0021] Based on the first and second dimension information of the target undirected edge element, the position of the target undirected edge element is adjusted to the first matrix region of the undirected graph matrix to obtain the matrix to be decomposed. The first matrix region is determined according to the position of the target diagonal in the undirected graph matrix.

[0022] Optionally, the method described above performs matrix decomposition on the nodes in the matrix to be decomposed, obtaining submatrices and decomposition nodes, including:

[0023] Determine the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed.

[0024] Based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed, the first decomposition node is determined. The decomposition degree of the first decomposition node is between the target decomposition degree and the first decomposition degree of the matrix to be decomposed, and is closer to the first decomposition degree than the decomposition degree of other nodes to be decomposed.

[0025] Based on the first decomposition node and the matrix to be decomposed, determine the first submatrix and the submatrix to be decomposed in the matrix to be decomposed;

[0026] The process of determining the first decomposition node in the submatrix to be decomposed is repeated, taking the decomposition degree of the child node to be decomposed as the decomposition degree of the child node to be decomposed, and the target decomposition degree of the submatrix to be decomposed as the target decomposition degree of the submatrix. This process continues until the decomposition degree of the child node to be decomposed is no longer between the target decomposition degree and the first decomposition degree of the submatrix. Then, the second decomposition node and the second submatrix are determined. The target decomposition degree of the submatrix to be decomposed is determined based on the target undirected edge elements in the submatrix to be decomposed.

[0027] Based on the first and second submatrices, the submatrices are obtained;

[0028] Based on the first and second decomposition nodes, the decomposition nodes are obtained.

[0029] Optionally, the method described above determines the first decomposition node among the nodes to be decomposed based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed, including:

[0030] Compare the target decomposition degree and the first decomposition degree of the matrix to be decomposed;

[0031] If the target decomposition degree of the matrix to be decomposed is less than or equal to the first decomposition degree, then the initial first decomposition node in the nodes to be decomposed is determined.

[0032] If there are two or more initial first decomposition nodes, the first decomposition node in the initial first decomposition nodes is determined based on the initial nodes to be decomposed in the nodes to be decomposed and the target diagonal in the undirected graph matrix.

[0033] Optionally, the method described above, after comparing the target decomposition degree and the first decomposition degree of the matrix to be decomposed, further includes:

[0034] If the target decomposition degree of the matrix to be decomposed is greater than the first decomposition degree, then the first decomposition degree is adjusted to obtain the second decomposition degree, wherein the second decomposition degree is greater than the target decomposition degree of the matrix to be decomposed.

[0035] Based on the second decomposition degree and the matrix to be decomposed, matrix decomposition is performed on the matrix to be decomposed to obtain submatrices and decomposition nodes.

[0036] Optionally, the above method, based on the submatrix and decomposed nodes, yields an industry network knowledge graph, including:

[0037] Based on the decomposition nodes, determine the matrix relationships between the submatrices;

[0038] Based on the matrix relationships between the submatrices, an industry network knowledge graph is obtained.

[0039] Secondly, this application provides an industrial network knowledge graph construction device, comprising:

[0040] The first determining module is used to determine the undirected graph matrix, wherein the undirected graph matrix includes entity information elements and undirected edge elements, and the undirected edge elements represent the undirected relationship between entity information elements;

[0041] The second determining module is used to obtain the matrix to be decomposed based on the undirected edge elements of the undirected graph matrix, wherein the matrix to be decomposed includes a first matrix region, and the target undirected edge elements are arranged in the undirected edge elements within the first matrix region.

[0042] The decomposition module is used to perform matrix decomposition on the decomposition nodes in the matrix to be decomposed, and obtain submatrices and decomposition nodes. The decomposition nodes are determined based on the target decomposition degree and the first decomposition degree of the matrix to be decomposed. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed.

[0043] The third determining module is used to obtain the industry network knowledge graph based on the submatrix and decomposed nodes.

[0044] Thirdly, an electronic device is provided, comprising:

[0045] Processor, and memory that is communicatively connected to the processor;

[0046] The memory stores instructions that the computer executes;

[0047] The processor executes computer execution instructions stored in memory, causing the electronic device to execute any of the industrial network knowledge graph construction methods in the first aspect.

[0048] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the industrial network knowledge graph construction method as described in any of the first aspects.

[0049] Fifthly, a computer program product, comprising a computer program / instructions that, when executed by a processor, implement the industrial network knowledge graph construction method of any one of the first aspects.

[0050] This application provides a method, apparatus, electronic device, storage medium, and program product for constructing an industrial network knowledge graph. The method first determines an undirected graph matrix, then obtains a matrix to be decomposed based on the undirected edge elements of the undirected graph matrix, then performs matrix decomposition on the decomposition nodes in the matrix to obtain sub-matrices and decomposition nodes, and finally obtains an industrial network knowledge graph based on the sub-matrices and decomposition nodes. Through matrix decomposition, the knowledge graph of production relations within an industry is mined, separating closely related industries from the industrial network to obtain the largest sub-industry group that meets the preset decomposition degree. The close internal relationships within these sub-industry groups, and the industrial network knowledge graph formed by the relationships between these sub-industry groups, can better organize, manage, and utilize industrial network data. This helps enterprises extract valuable information from massive amounts of data, providing more accurate and comprehensive support for decision-making. This method of constructing an industrial network knowledge graph, based on industrial network data, constructs an undirected graph matrix and mines the knowledge graph of production relations within the industry through matrix decomposition. It constructs an observable, realistic, and quantifiable industrial network knowledge graph as a form of industrial knowledge data, replacing manually analyzed industrial knowledge data and achieving the technical effect of improving the reliability of industrial knowledge data. Attached Figure Description

[0051] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0052] Figure 1 A flowchart illustrating the method for constructing an industrial network knowledge graph provided in this application embodiment;

[0053] Figure 2 A flowchart illustrating the method for constructing an industrial network knowledge graph provided in this application embodiment;

[0054] Figure 3 A schematic diagram of the structure of the industrial network knowledge graph construction device provided in the embodiments of this application;

[0055] Figure 4This is a hardware structure diagram of the electronic device provided in the embodiments of this application.

[0056] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0057] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0058] It should be noted that the customer information (including but not limited to customer device information, customer personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the customer or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with relevant laws, regulations and standards, and corresponding operation portals are provided for customers to choose to authorize or refuse.

[0059] First, let me explain the terms used in this application:

[0060] A knowledge graph is a semantic network used to describe relationships between entities. It's a semi-structured data representation method used to describe entities, attributes, and the relationships between them. The core idea of ​​a knowledge graph is to transform information from the real world into a graph, where nodes represent entities and edges represent relationships between them. A knowledge graph is more than just a graphical knowledge base; it also contains semantic descriptions of entities and relationships that can be understood and processed by computers.

[0061] An industry network refers to the interconnected and collaborative relationships established between different enterprises, institutions, and individuals. These relationships can include supply chains, value chains, partnerships, etc., aiming to jointly achieve the development goals of a specific industry or market. Industry networks typically involve cooperation among multiple participants in the form of resource sharing, information exchange, and technological collaboration to promote innovation, improve efficiency, and reduce costs. The establishment of industry networks can bring broader market opportunities and greater competitiveness, helping to drive the development of the entire industry.

[0062] Industry network knowledge graphs are graph structures built upon data and knowledge within an industry sector. They are used to describe and analyze key entities, relationships, and attributes within an industry ecosystem. These knowledge graphs typically consist of entities (such as companies, products, and technologies), relationships (such as supply chains, partnerships, and competition), and attributes (such as market share and technological characteristics), presenting the connections and information between these entities in a graph format. Industry network knowledge graphs can help people better understand the structure and evolution of industry ecosystems, supporting decision-making and research in areas such as industry analysis, market forecasting, and competitive intelligence. Utilizing knowledge graph technology, intelligent management, mining, and application of industry network data can be achieved.

[0063] Hash Table: A data structure that allows direct access to data stored in memory based on a key. Hash tables store data consisting of keys and values.

[0064] In existing technologies, describing industrial relationships using the industrial chain is a widespread and common approach. An industrial chain is a concept encompassing four dimensions: value chain, enterprise chain, supply and demand chain, and spatial chain. These four dimensions form the industrial chain through a process of mutual equilibrium and interconnection. Technicians sift through massive amounts of industrial data to extract information on industrial relationships. Through manual analysis of this information, they identify upstream and downstream relationships and the four dimensions to create an industrial chain describing these relationships. As a form of industrial knowledge data, the industrial chain allows upstream links to supply products or services to downstream links, and downstream links to provide feedback to upstream links. However, the subjective nature of the industrial chain makes it difficult to use for quantitative prediction of industrial production. Therefore, the reliability of existing industrial knowledge data is low.

[0065] The industrial network knowledge graph construction method, apparatus, electronic device, storage medium, and program product provided in this application simplify the complex industrial relationship structure and make it easier to understand. This requires mining the internal production relationships within industries and separating closely related industries from the industrial network. An undirected graph matrix is ​​then determined using industrial network data. The undirected edge elements of the undirected graph matrix are processed to obtain a matrix to be decomposed. The decomposition nodes of the matrix to be decomposed are then subjected to matrix decomposition processing to obtain sub-matrices and decomposition nodes that meet preset requirements. The sub-matrices represent closely related sub-industrial groups mined from the industrial network, and the decomposition nodes represent the relationship nodes between the sub-industrial groups and the global matrix. Finally, the sub-matrices and decomposition nodes form an industrial network knowledge graph. This industrial network knowledge graph construction method allows for better organization, management, and utilization of industrial network data. It helps enterprises extract valuable information from massive amounts of data, providing more accurate and comprehensive support for decision-making and achieving the technical effect of improving the reliability of industrial knowledge data.

[0066] This application provides a method for constructing an industrial network knowledge graph, which aims to solve the above-mentioned technical problems in the prior art.

[0067] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0068] Figure 1 This is a flowchart illustrating the method for constructing an industrial network knowledge graph provided in an embodiment of this application. Figure 1 As shown, the industrial network knowledge graph construction method provided in this embodiment includes:

[0069] S101. Determine the undirected graph matrix, wherein the undirected graph matrix includes entity information elements and undirected edge elements, and the undirected edge elements represent the undirected relationship between entity information elements;

[0070] An undirected graph matrix is ​​a data structure used to represent undirected graphs. In an undirected graph, the edges between vertices are undirected; that is, the edge from vertex A to vertex B is the same as the edge from vertex B to vertex A. An undirected graph matrix is ​​typically a two-dimensional matrix, where the rows and columns represent the vertices in the graph, and the elements indicate whether there is an edge connecting the vertices. If two vertices are connected by an edge, the corresponding matrix element is 1; if two vertices are not connected by an edge, the corresponding matrix element is 0.

[0071] In this embodiment, vertices in the undirected graph represent entity information in the industrial network data, and edges in the undirected graph represent the relationships between entity information in the industrial network data. The undirected graph matrix can be a matrix representing whether there is an undirected relationship between entity information, wherein the row and column information of the undirected graph matrix are determined based on the entity information elements, and the elements of the undirected graph matrix are determined based on the undirected edge elements. For example, the row information of an undirected graph matrix can be enterprise entity h1, product entity t1, and patent entity t2, and the column information can be enterprise entity h1, product entity t1, and patent entity t2. The undirected graph matrix has 9 elements, which can respectively represent the undirected relationship between enterprise entity h1 and enterprise entity h1, the undirected relationship between enterprise entity h1 and product entity t1, the undirected relationship between enterprise entity h1 and patent entity t2, the undirected relationship between product entity t1 and enterprise entity h1, the undirected relationship between product entity t1 and product entity t1, the undirected relationship between product entity t1 and patent entity t2, the undirected relationship between patent entity t2 and enterprise entity h1, the undirected relationship between patent entity t2 and product entity t1, and the undirected relationship between patent entity t2 and patent entity t2. Among them, the undirected relationship between enterprise entity h1, the undirected relationship between product entity t1, and the undirected relationship between patent entity t2 represent the relationship between entity information and itself. Therefore, the relationship between enterprise entity h1 and its elements, product entity t1 and its elements, and patent entity t2 and its elements can be set to 0. When there is no relationship between enterprise entity h1 and product entity t1, the elements of enterprise entity h1 and product entity t1 are 0. When there is a relationship between enterprise entity h1 and patent entity t2, the elements of enterprise entity h1 and patent entity t2 are 1.

[0072] In some embodiments, the platform can receive industry network data compiled and summarized by the relevant enterprises, generate an undirected graph matrix based on the entity information of the industry network data and the relationships between the entity information, and thus determine the undirected graph matrix.

[0073] In this embodiment of the application, the method for determining the undirected graph matrix may further include:

[0074] Acquire industrial network data, which includes entity information within the industrial network and the relationships between those entities.

[0075] Based on industry network data, a directed graph matrix is ​​constructed, which includes entity information elements and directed edge elements. Entity information elements represent entity information, and directed edge elements represent the directed relationships between entity information elements.

[0076] Determine the undirected graph matrix based on the entity information elements and directed edge elements in the directed graph matrix.

[0077] In this embodiment of the application, the method for constructing a directed graph matrix based on industrial network data may further include:

[0078] Encode the entity information in the industrial network data to obtain the data encoding of the entity information;

[0079] Based on the data encoding of entity information, determine the first and second dimension information of the initial directed graph matrix;

[0080] The directed graph matrix is ​​obtained based on the first and second dimension information of the initial directed graph matrix and the relationship between entity information in the industrial network data.

[0081] Among them, industrial network data can refer to the data formed by the mutual connections and cooperative relationships established between different enterprises, institutions and individuals.

[0082] A directed graph matrix is ​​a matrix that describes the structure of a directed graph. The rows and columns of the matrix represent the vertices in the graph, and the elements represent the directed edges between those vertices. Here, vertices refer to entity information, and the directed edges between them represent the directed relationships between these entities. Therefore, a directed graph matrix characterizes whether there are directed relationships between entity information. The row and column information of the matrix is ​​determined based on the entity information elements, and the elements are determined based on the directed edge elements.

[0083] The initial directed graph matrix is ​​a directed graph matrix without any row information, column information, or element information.

[0084] The first dimension of the initial directed graph matrix refers to the encoding information of its rows, and the second dimension refers to the encoding information of its columns. Specifically, the first dimension refers to row information, and the second dimension refers to column information. Both the row and column encoding information in the initial directed graph matrix are determined by the data encoding of the entity information.

[0085] In some embodiments, the industrial network data received by the construction platform can be represented in a triple format, where the triple format is (h, r, t), and h and t are entity information, and r represents the relationship between entity information h and entity information t. Entity information h and entity information t represent entity information in the industrial network, such as enterprise entities, product entities, patent entities, etc., and relationship r represents the relationship between entity information in the industrial network, such as procurement, production, leasing, etc. In particular, in real life, industrial data contains a large number of relationships that are difficult to name; such relationships can be simply described using r0. Therefore, industrial network data includes entity information in the industrial network and the relationships between entity information. The raw industrial network data is numerous, messy, and difficult to process; therefore, it can be summarized and organized into triple format industrial network data by contacting relevant data companies.

[0086] In this embodiment, the entity information and relationships between entities in the industrial network data can be obtained by constructing hash tables to obtain the data codes corresponding to the entity information and the data codes corresponding to the relationships between entities. The entity information in the industrial network data is encoded to obtain the corresponding data codes, which are stored in a hash table that can store entity information. Similarly, the directed edge elements between entities are encoded to obtain the data codes corresponding to the relationships between entities, which are stored in a hash table that can store the relationships between entities. The keys of the hash table storing entity information can store the entity information in the industrial network data, and the values ​​can store the corresponding data codes. The keys of the hash table storing the relationships between entities can store the relationships between entities in the industrial network data, and the values ​​can store the data codes corresponding to the relationships between entities. For example, in the construction platform, if there is a piece of industrial network data (h0, r0, t0), after adding h0, the hash table storing entity information is {h0:0}. After adding t0, the hash table is {h0:0, t0:1}. After adding r0, the hash table storing the relationship between entity information is {r0:1}. If (h0, r1, t1) is added, the hash table storing entity information is {h0:0, t0:1, t1:2}, and the hash table storing the relationship between entity information is {r0:1, r1:2}.

[0087] In this embodiment, the row and column information of the initial directed graph matrix is ​​determined based on the data encoding of the entity information. Then, the data encoding of the relationships between the entity information is placed into the corresponding element positions in the initial directed graph matrix, thus obtaining the directed graph matrix. The row and column information of the directed graph matrix correspond to the data encoding of the entity information. Based on the row and column information of the directed graph matrix, the entity information corresponding to the row and column information of the directed graph matrix can be found in the hash table storing the entity information. The element information within the directed graph matrix represents the data encoding corresponding to the relationships between the entity information, and the relationships between the entity information represented by the element information of the directed graph matrix can be found in the hash table storing the relationships between the entity information.

[0088] In this embodiment, an undirected graph matrix is ​​determined based on the entity information elements and directed edge elements in the directed graph matrix. For example, in a construction platform, the directed edge elements in the directed graph matrix are the encoded data corresponding to the directed relationships between entity information. If a directed edge element in the directed graph matrix is ​​0, it proves that there is no relationship between the two entity information corresponding to that directed edge element; the encoded values ​​of directed edge elements that are related to entity information are greater than 0. Since the elements of an undirected graph matrix are undirected edge elements, and the elements of an undirected graph matrix can only reflect whether there is a relationship between entities, the encoded values ​​of elements that are related to entity information in the undirected graph matrix can all be set to 1, and the encoded values ​​of elements that are not related to entity information can be set to 0. By reassigning the directed edge elements greater than 0 in the directed graph matrix to 1, while keeping the entity information corresponding to the row and column information of the directed graph matrix unchanged, the undirected graph matrix can be obtained.

[0089] S102. Based on the undirected edge elements of the undirected graph matrix, the matrix to be decomposed is obtained, wherein the matrix to be decomposed includes a first matrix region, and the target undirected edge elements are arranged in the undirected edge elements within the first matrix region.

[0090] The matrix to be decomposed refers to a matrix that needs to be decomposed into a simpler form or subjected to specific operations. In mathematics and linear algebra, common matrices to be decomposed include symmetric matrices, orthogonal matrices, and triangular matrices. These matrices can be decomposed using different methods to obtain simpler forms or perform specific operations. The first matrix region refers to the region to the right of the target diagonal, which can be any diagonal in the matrix. Undirected edge elements of the target refer to undirected edge elements where the undirected relationship between entity information exists.

[0091] In some embodiments, to facilitate subsequent matrix decomposition, the undirected graph matrix needs to undergo row and column transformations based on the element attributes of the undirected edge elements, so that the target undirected edge elements are primarily concentrated in the first matrix region of the matrix to be decomposed. For example, in a construction platform, based on the element attributes of the undirected edge elements in the undirected graph matrix and the row information of the target undirected edge elements, the number of target undirected edge elements in each row of the undirected graph matrix can be determined. The undirected edge elements in each row of the undirected graph matrix are then swapped from top to bottom according to the order of the number from smallest to largest, so that the target undirected edge elements are arranged at the top of the undirected graph matrix. Then, based on the column information of the undirected edge elements, the columns of the undirected graph matrix are rearranged, starting from the first row of the undirected graph matrix, selecting the target undirected edge elements from the first row. For each column of undirected edge elements, swap the positions of all undirected edge elements in that column. This ensures that the target undirected edge elements in the first row are positioned closer to the right of the target diagonal. Similarly, for the next row, without affecting the positions of the target undirected edge elements in the previous row, swap the positions of the columns of target undirected edge elements in that row, ensuring that the target undirected edge elements in the next row are positioned closer to the right of the target diagonal. Continue this process until the target undirected edge elements in the undirected graph matrix are arranged as close to the first matrix region as possible. This yields the matrix to be decomposed. While most target undirected edge elements in the matrix to be decomposed are arranged within the first matrix region, it is unavoidable that some target undirected edge elements will be located outside the first matrix region during the adjustment process. For example, in the obtained matrix to be decomposed, the last row containing the target undirected edge elements is the row with the largest number of target undirected edge elements. If this row has 8 elements, and the last row containing the target undirected edge elements is the third to last row of the matrix to be decomposed, then the third to last row has only two positions within the first matrix region. When adjusting the positions of the target undirected edge elements, at most two target undirected edge elements in this row are within the first matrix region. Therefore, the other target undirected edge elements are outside the first matrix region. In this embodiment, the method for obtaining the matrix to be decomposed based on the undirected edge elements of the undirected graph matrix may further include:

[0092] Based on the element attributes of the undirected edge elements in the undirected graph matrix, determine the target undirected edge element, as well as the first and second dimension information of the target undirected edge element. The element attributes of the undirected edge element include attributes that represent the existence of undirected relations.

[0093] Based on the first and second dimension information of the target undirected edge element, the position of the target undirected edge element is adjusted to the first matrix region of the undirected graph matrix to obtain the matrix to be decomposed. The first matrix region is determined according to the position of the target diagonal in the undirected graph matrix.

[0094] Here, the target undirected edge element refers to the undirected edge element where there is an undirected relationship between entity information. The first dimension information of the target undirected edge element refers to the row information of the target undirected edge element, and the second dimension information of the target undirected edge element refers to the column information of the target undirected edge element.

[0095] In some embodiments, the element attributes of undirected edge elements with existing undirected relationships are obtained based on the element attributes of the undirected edge elements in the undirected graph matrix, i.e., the element attributes of the target undirected edge element are obtained. Based on the element attributes of the undirected edge elements in the undirected graph matrix, the target undirected edge element, as well as the row and column information of the target undirected edge element, are determined. By adjusting the position of the target undirected edge element, the matrix to be decomposed is obtained.

[0096] In some embodiments, based on the element attributes of the undirected edge elements in the undirected graph matrix and the row information of the target undirected edge elements, the number of target undirected edge elements in each row of the undirected graph matrix can be calculated. Then, the undirected edge elements in each row of the undirected graph matrix are swapped from top to bottom according to the order of the number of target undirected edge elements in ascending order, so that the target undirected edge elements in the undirected graph matrix are arranged in the upper part of the undirected graph matrix. Specifically, the rows in the undirected graph matrix containing target undirected edge elements are compared upwards with the number of target undirected edge elements in the previous row. If the number of target undirected edge elements in the previous row is larger, the positions of the previous row and the current row are swapped. If the number of target undirected edge elements in the previous row is smaller or equal, no position is swapped. The swapping continues upwards, and each time the number of target undirected edge elements in the previous row is compared with the number of target undirected edge elements in the previous row, until all rows containing target undirected edge elements are swapped to the upper part of the undirected graph matrix, and the rows containing target undirected edge elements are arranged in descending order of the number of target undirected edge elements in the upper part of the undirected graph matrix, thus obtaining the row-swapped undirected graph matrix.

[0097] For example, in a platform construction, if the number of target undirected edge elements in the first row of an undirected graph matrix is ​​5, and the number of target undirected edge elements in the second row is 3, by comparing the counts, the first and second rows of the undirected graph matrix are swapped, resulting in the second row of the undirected graph matrix having 5 target undirected edge elements and the first row having 3. If the number of target undirected edge elements in the third row of the undirected graph matrix is ​​also 3, by comparing the counts of target undirected edge elements in the adjacent previous row, the number of target undirected edge elements in the second row is greater than the number in the third row, so the undirected graph matrices are swapped. The positions of the second and third rows of the undirected graph matrix are then considered. At this point, the number of target undirected edge elements in the second row of the undirected graph matrix is ​​3, and the number of target undirected edge elements in the third row of the undirected graph matrix is ​​5. Next, the number of target undirected edge elements in the first and second rows of the undirected graph matrix is ​​compared. At this point, the number of target undirected edge elements in the first and second rows of the undirected graph matrix is ​​3, which is equal, so there is no need to swap the rows. This process is repeated until all rows with target undirected edge elements are swapped to the top of the undirected graph matrix. The rows with target undirected edge elements are then arranged from top to bottom in the top of the undirected graph matrix according to the order of the number of target undirected edge elements from smallest to largest. This gives us the undirected graph matrix after the row swaps.

[0098] In some embodiments, after obtaining the undirected graph matrix after row swapping, the columns of the undirected graph matrix are rearranged according to the column information of the undirected edge elements in the undirected graph matrix after row swapping. Starting from the first row of the undirected graph matrix, the entire column of the target undirected edge elements in the first row is selected, and all the undirected edge elements in the column are swapped so that the target undirected edge elements in the first row are arranged to the right of the target diagonal. The entire column of the target undirected edge elements in the next row is swapped without affecting the position of the target undirected edge elements in the previous row, so that the target undirected edge elements in the next row are arranged to the right of the target diagonal. This process continues until the target undirected edge elements in the undirected graph matrix are arranged as much as possible within the first matrix area, thus obtaining the matrix to be decomposed.

[0099] For example, in a platform construction, if the target diagonal of an undirected graph matrix is ​​the main diagonal, and there are two target undirected edge elements in the first row, with their column positions being the first and tenth columns respectively, and the target diagonal is located in the first column of the first row, with the second and third columns in the first row closest to the right of the target diagonal. Therefore, swapping the undirected edge elements in the first and second columns of the undirected graph matrix, and swapping the undirected edge elements in the tenth and third columns, moves the two target undirected edge elements in the first row to the first matrix region closest to the right of the target diagonal. Similarly, if there are two target undirected edge elements in the second row, located in the fourth and sixth columns respectively, then the target diagonal is located in the second column of the second row, with the third and sixth columns in the second row closest to the right of the target diagonal. In the fourth column, since the positions of the two target undirected edge elements in the first row have already been swapped to the second and third columns, the positions of the target undirected edge elements already arranged in the previous row should not be affected. Therefore, the entire column of target undirected edge elements in the second row of the undirected graph matrix should not be swapped with the third column of the undirected graph matrix. Instead, the positions of the target undirected edge elements in the second row should be swapped to the third column, meaning the positions of the target undirected edge elements in the fourth column of the second row should be swapped to the fourth column. Therefore, the entire column of target undirected edge elements in the fourth column of the second row of the undirected graph matrix does not need to be swapped. The entire column of target undirected edge elements in the sixth column of the second row of the undirected graph matrix should be swapped with the fifth column of the undirected graph matrix. This process continues until the target undirected edge elements in the undirected graph matrix are arranged as much as possible within the first matrix region, resulting in the matrix to be decomposed.

[0100] S103. Perform matrix decomposition on the decomposition nodes in the matrix to be decomposed to obtain submatrices and decomposition nodes. The decomposition nodes are determined based on the target decomposition degree and the first decomposition degree of the matrix to be decomposed. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed.

[0101] The first decomposition degree is preset based on expert experience, or it can be directly specified by the client. The target decomposition degree refers to the minimum number of target undirected edge elements in the rows of the matrix to be decomposed. Submatrices are sub-industry clusters formed by the close relationships between entity information within them, and decomposition nodes represent the matrix relationships between submatrices.

[0102] In some embodiments, decomposition nodes are determined based on the target decomposition degree and the first decomposition degree of the matrix to be decomposed. Since the undirected edge elements of the undirected graph matrix are first processed by row and column transformations, the target undirected edge elements are concentrated in the first matrix region of the matrix to be decomposed. The first matrix region is the region to the right of the target diagonal; therefore, the selected decomposition nodes are nodes selected along the target diagonal. The number of target undirected edge elements in the row containing the decomposition node is used as the decomposition degree of the decomposition node. The decomposition degree of the decomposition node must be within the decomposition degree range between the target decomposition degree and the first decomposition degree of the matrix to be decomposed. Matrix decomposition processing is performed on the decomposition nodes in the matrix to be decomposed that meet the decomposition degree range to obtain submatrices and decomposition nodes.

[0103] In this embodiment of the application, the method for performing matrix decomposition processing on the decomposition nodes in the matrix to be decomposed to obtain submatrices and decomposition nodes may further include:

[0104] Determine the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed.

[0105] Based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed, the first decomposition node is determined. The decomposition degree of the first decomposition node is between the target decomposition degree and the first decomposition degree of the matrix to be decomposed, and is closer to the first decomposition degree than the decomposition degree of other nodes to be decomposed.

[0106] Based on the first decomposition node and the matrix to be decomposed, determine the first submatrix and the submatrix to be decomposed in the matrix to be decomposed;

[0107] The process of determining the first decomposition node in the submatrix to be decomposed is repeated, taking the decomposition degree of the child node to be decomposed as the decomposition degree of the child node to be decomposed, and the target decomposition degree of the submatrix to be decomposed as the target decomposition degree of the submatrix. This process continues until the decomposition degree of the child node to be decomposed is no longer between the target decomposition degree and the first decomposition degree of the submatrix. Then, the second decomposition node and the second submatrix are determined. The target decomposition degree of the submatrix to be decomposed is determined based on the target undirected edge elements in the submatrix to be decomposed.

[0108] Based on the first and second submatrices, the submatrices are obtained;

[0109] Based on the first and second decomposition nodes, the decomposition nodes are obtained.

[0110] In this embodiment of the application, the method for determining the first decomposition node among the nodes to be decomposed based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed may further include:

[0111] Compare the target decomposition degree and the first decomposition degree of the matrix to be decomposed;

[0112] If the target decomposition degree of the matrix to be decomposed is less than or equal to the first decomposition degree, then the initial first decomposition node in the nodes to be decomposed is determined.

[0113] If there are two or more initial first decomposition nodes, the first decomposition node in the initial first decomposition nodes is determined based on the initial nodes to be decomposed in the nodes to be decomposed and the target diagonal in the undirected graph matrix.

[0114] In this embodiment of the application, the method after comparing the target decomposition degree and the first decomposition degree of the matrix to be decomposed may further include:

[0115] If the target decomposition degree of the matrix to be decomposed is greater than the first decomposition degree, then the first decomposition degree is adjusted to obtain the second decomposition degree, wherein the second decomposition degree is greater than the target decomposition degree of the matrix to be decomposed.

[0116] Based on the second decomposition degree and the matrix to be decomposed, matrix decomposition is performed on the matrix to be decomposed to obtain submatrices and decomposition nodes.

[0117] In this context, the nodes to be decomposed in the matrix to be decomposed are selected along the target diagonal of the matrix. The child nodes to be decomposed in the submatrix are selected along the target diagonal of the submatrix. The decomposition degree of a node refers to the number of target undirected edge elements in the row containing the node. The target decomposition degree of the matrix to be decomposed is the minimum number of target undirected edge elements in the rows containing these elements. The target decomposition degree of the submatrix to be decomposed is the minimum number of target undirected edge elements in the rows containing these elements. The first decomposition node in the matrix to be decomposed is a node on the target diagonal of the matrix to be decomposed. The second decomposition node in the child nodes is a node on the target diagonal of the submatrix. The initial first decomposition node in the matrix to be decomposed is the first node on the target diagonal of the matrix. The second decomposition degree is given by the client.

[0118] In this embodiment, the decomposition degree is the number of target undirected edge elements in a row of the matrix or submatrix to be decomposed that contains such elements. The decomposition degree satisfies: d = B(i, U), where U is the matrix or submatrix to be decomposed, i is the row sequence in the matrix or submatrix to be decomposed, d is the decomposition degree, and B(i, U) is the number of target undirected edge elements in the i-th row of the matrix or submatrix to be decomposed.

[0119] In this embodiment, the target decomposition degree refers to the minimum number of target undirected edge elements in a row of the matrix or submatrix to be decomposed that contains target undirected edge elements. The target decomposition degree satisfies: d0 = min(B(i,U)), where d0 is the target decomposition degree and min(B(i,U)) is the minimum number of target undirected edge elements obtained by traversing each row of the matrix or submatrix to be decomposed and comparing the results.

[0120] In some embodiments, the matrix to be decomposed or its submatrix to be decomposed needs to be decomposed by calculating the target decomposition degree, the decomposition degree of the nodes to be decomposed or the sub-nodes to be decomposed, in order to obtain a submatrix that meets the requirements. Matrix decomposition involves decomposing the rows and columns of the nodes to be decomposed in the matrix to be decomposed, and the resulting submatrix is ​​a matrix that includes the target diagonal.

[0121] In this embodiment, the matrix decomposition process is as follows: First, based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed, the first decomposition node is determined. Based on the first decomposition node and the matrix to be decomposed, the first submatrix and the submatrix to be decomposed are determined. Then, matrix decomposition processing is performed on the submatrix to be decomposed to determine the second decomposition node and the second submatrix. Finally, based on the first submatrix and the second submatrix, the submatrix is ​​obtained, and based on the first decomposition node and the second decomposition node, the decomposition node is obtained.

[0122] For example, in the platform construction, suppose the matrix to be decomposed is a 20x20 matrix, with the target diagonal as the main diagonal. The target decomposition degree of the matrix to be decomposed is 2, and the first decomposition degree is 4. The nodes to be decomposed are nodes on the target diagonal. The decomposition degrees of the nodes to be decomposed in the first 10 rows are 2, 2, 2, 3, 3, 4, 4, 5, 5, 6, respectively. The node in the first row and first column is selected as the initial node to be decomposed. The decomposition degree of the node in the first row is 2, which is less than the first decomposition degree. Then, the node in the second row and second column is selected as the node to be decomposed along the target diagonal. The decomposition degree of the node in the second row is less than the first decomposition degree. This process continues along the target diagonal until the first decomposition node is found. The decomposition degree of the first decomposition node is between the target decomposition degree and the first decomposition degree of the matrix to be decomposed, and is closer to the first decomposition degree compared to the decomposition degrees of other nodes to be decomposed. Since the decomposition degree of the node in the sixth row is 4, which is equal to the first decomposition degree, the first decomposition node at this point is the node in the sixth row and sixth column.

[0123] Based on the first decomposition node in the sixth row and sixth column, the matrix to be decomposed is decomposed in the sixth row and sixth column, yielding the first submatrix and the submatrix to be decomposed. The first submatrix is ​​formed by starting with the initial node to be decomposed, extending along the target diagonal, and ending at the first decomposition node. The submatrix to be decomposed is formed by starting with the first decomposition node, extending along the target diagonal, and ending at the last node of the matrix to be decomposed. Therefore, the first submatrix starts from the node in the first row and first column, extends along the main diagonal, and ends at the node in the sixth row and sixth column. The submatrix to be decomposed starts from the node in the sixth row and sixth column, extends along the main diagonal, and ends at the node in the twentieth row and twentieth column.

[0124] The resulting submatrix is ​​now a 15x15 matrix. To perform matrix decomposition on this submatrix, determine the second decomposition node and the second submatrix. This requires recalculating the decomposition degree of the nodes in the submatrix. The nodes to be decomposed are those on the target diagonal of the submatrix. If the decomposition degrees of the nodes in the first 10 rows of the submatrix are 2, 3, 2, 3, 4, 4, 5, 5, 4, 6 respectively, and the target diagonal is the main diagonal, the target decomposition degree of the submatrix is ​​2, while the first decomposition degree remains unchanged at 4. The node in the first row and first column is selected as the initial node to be decomposed. The decomposition degree of the first row node is 2, which is less than the first decomposition degree of 4. Then, the node in the second row and second column is selected along the target diagonal. The decomposition degree of the second row node is 3, which is less than the first decomposition degree of 4. This process continues along the target diagonal until the second decomposition node is found. The decomposition degree of the second decomposition node lies between the target decomposition degree and the first decomposition degree of the submatrix to be decomposed, and is closer to the first decomposition degree than the decomposition degrees of other submatrixes to be decomposed. Since the decomposition degree 4 of the submatrix to be decomposed in the fifth row is equal to the first decomposition degree, the second decomposition node at this time is the submatrix to be decomposed in the fifth row and fifth column.

[0125] Based on the second decomposition node in the fifth row and fifth column, the submatrix to be decomposed is decomposed in the fifth row and fifth column, resulting in a second submatrix and a first submatrix to be decomposed. The second submatrix is ​​formed by starting with the initial submatrix to be decomposed, extending along the target diagonal, and ending at the second decomposition node. The first submatrix to be decomposed is formed by starting with the second decomposition node, extending along the target diagonal, and ending at the last node of the submatrix to be decomposed. Therefore, the second submatrix starts from the node in the first row and first column, extends along the main diagonal, and ends at the node in the fifth row and fifth column. The first submatrix to be decomposed starts from the node in the fifth row and fifth column, extends along the main diagonal, and ends at the node in the fifteenth row and fifteenth column. The matrix decomposition process continues on the first submatrix to be decomposed until the last submatrix to be decomposed is reached or the decomposition degree of the submatrix to be decomposed is greater than the first decomposition degree.

[0126] In this embodiment, after comparing the target decomposition degree and the first decomposition degree of the matrix to be decomposed, if the target decomposition degree of the matrix to be decomposed is greater than the first decomposition degree, the first decomposition degree is adjusted to obtain a second decomposition degree, wherein the second decomposition degree is greater than the target decomposition degree of the matrix to be decomposed. The second decomposition degree is obtained by consulting the customer. Then, matrix decomposition is performed on the matrix to be decomposed based on the second decomposition degree to obtain submatrices and decomposition nodes. For example, if the target decomposition degree of the matrix to be decomposed is 5 and the first decomposition degree is 4, and the decomposition degree of the nodes to be decomposed is greater than the first decomposition degree, matrix decomposition cannot be performed. In this case, the first decomposition degree needs to be reset, and the customer provides a second decomposition degree based on their opinion. The second decomposition degree must be greater than the target decomposition degree of 5 for matrix decomposition to be performed to obtain submatrices and decomposition nodes.

[0127] In this embodiment, after comparing the target decomposition degree and the first decomposition degree of the submatrix to be decomposed, if the target decomposition degree of the submatrix to be decomposed is greater than the first decomposition degree during the matrix decomposition process, the decision to continue decomposition can be made by consulting the client. If the client believes that further decomposition is possible and provides a third decomposition degree, which is greater than the target decomposition degree of the submatrix to be decomposed, the submatrix to be decomposed continues to be decomposed based on the third decomposition degree, resulting in submatrices and decomposition nodes. If the client believes that further decomposition of the submatrix to be decomposed is not meaningful and does not adjust the first decomposition degree, then the submatrix to be decomposed is taken as the last submatrix, resulting in all submatrices and decomposition nodes. For example, if the target decomposition degree of the submatrix to be decomposed is 10 and the first decomposition degree is 8, and the decomposition degree of the sub-nodes to be decomposed is greater than the first decomposition degree, matrix decomposition cannot be performed, then the first decomposition degree needs to be reset according to the client's opinion. If the client believes that the submatrix to be decomposed needs to be further decomposed, the client shall provide a second degree of decomposition. The second degree of decomposition must be greater than the target degree of decomposition of the submatrix to be decomposed (8) in order to perform matrix decomposition processing on the submatrix to be decomposed and obtain submatrices and decomposition nodes. If the client believes that there is little point in continuing to decompose the submatrix to be decomposed and there is no need to adjust the first degree of decomposition, then the submatrix to be decomposed shall be taken as the last submatrix and all submatrices and decomposition nodes shall be obtained.

[0128] In this embodiment, after comparing the target decomposition degree and the first decomposition degree of the matrix to be decomposed, if the target decomposition degree of the matrix to be decomposed is less than or equal to the first decomposition degree, then the initial first decomposition node is determined among the nodes to be decomposed. If there are two or more initial first decomposition nodes, then the first decomposition node among the initial first decomposition nodes is determined based on the initial nodes to be decomposed and the target diagonal in the undirected graph matrix. For example, in the construction platform, after traversing all nodes to be decomposed in the matrix to be decomposed, it is found that the decomposition degree of all nodes to be decomposed is less than or equal to the first decomposition degree. At this time, it is necessary to compare the decomposition degrees of all nodes to be decomposed and find the node with the largest decomposition degree as the initial first decomposition node. If there is only one node with the largest decomposition degree, then the matrix to be decomposed is decomposed according to the initial first decomposition node to obtain submatrices and decomposition nodes. If there are two or more nodes with the largest decomposition degree, that is, there are two or more initial first decomposition nodes, then the first decomposition node among the initial first decomposition nodes is determined based on the initial nodes to be decomposed and the target diagonal in the undirected graph matrix. The matrix to be decomposed is then decomposed according to the first decomposition node to obtain submatrices and decomposition nodes.

[0129] In some embodiments, when the decomposition degree of all nodes to be decomposed is less than or equal to the first decomposition degree, and there are two or more nodes to be decomposed with the largest decomposition degree, the node to be decomposed with the largest decomposition degree can be taken as the initial first decomposition node. Based on all the initial first decomposition nodes, the matrix to be decomposed can be jointly decomposed to obtain the submatrix and decomposition nodes.

[0130] S104. Based on the submatrix and decomposed nodes, obtain the industry network knowledge graph.

[0131] In some embodiments, a submatrix represents a sub-industry cluster formed by the close relationships between entity information within the submatrix. The submatrix is ​​the largest sub-industry cluster that meets the decomposition degree requirement, and the decomposition node is the relationship node between this sub-industry cluster and the global hierarchy. The relationships between the entity information corresponding to the decomposition node can be found in a hash table storing the relationships between entity information using the row and column information of the decomposition node, thus obtaining the matrix relationships between the submatrixes represented by the decomposition node. The relationships corresponding to the decomposition node characterize the matrix relationships between two adjacent submatrixes of the decomposition node. Therefore, an industry network knowledge graph is formed by the submatrixes and the matrix relationships between them.

[0132] In this embodiment of the application, the method for obtaining the industrial network knowledge graph based on the submatrix and decomposed nodes may further include:

[0133] Based on the decomposition nodes, determine the matrix relationships between the submatrices;

[0134] Based on the matrix relationships between the submatrices, an industrial network knowledge graph is obtained.

[0135] Here, a submatrix is ​​a sub-industry cluster formed by the close relationships between entity information in the submatrix, a decomposition node is a relationship node between the sub-industry cluster and the global node, and the matrix relationship between submatrixes is the relationship corresponding to the undirected edge element at the location of the decomposition node found in the hash table that stores the relationship between entity information.

[0136] In some embodiments, the generated industry network knowledge graph is presented in the form of triples, such as [(U1,r0,U2),(U2,r1,U3)], where U1, U2, and U3 are submatrices obtained from matrix decomposition, representing different sub-industry groups, and r0 and r1 are the matrix relationships between the submatrices, i.e., the relationships between the sub-industry groups. The industry network knowledge graph, composed of submatrices and the matrix relationships between them, can better organize, manage, and utilize massive amounts of data, helping enterprises extract valuable information from massive amounts of data and providing more accurate and comprehensive support for decision-making.

[0137] The industrial network knowledge graph construction method provided in this application mines the knowledge graph of internal production relations within an industry through matrix decomposition. This separates closely related industries within the industrial network, obtaining the largest sub-industry clusters that meet a preset decomposition degree. The close internal relationships within these sub-industry clusters, combined with the relationships between them, allow for better organization, management, and utilization of industrial network data. This helps enterprises extract valuable information from massive amounts of data, providing more accurate and comprehensive support for decision-making. This method constructs an undirected graph matrix based on industrial network data, processes it into matrices that are easy to decompose, and then mines the knowledge graph through matrix decomposition. This results in an observable, realistic, and quantifiable industrial network knowledge graph, serving as a form of industrial knowledge data and achieving the technical effect of improving the reliability of industrial knowledge data.

[0138] Figure 2 This is a flowchart illustrating the method for constructing an industrial network knowledge graph provided in an embodiment of this application. Figure 2 As shown, the industrial network knowledge graph construction method provided in this embodiment includes:

[0139] S201. Form an undirected graph matrix based on industry network data.

[0140] In this embodiment, a directed graph matrix is ​​constructed based on industry network data. For example, for a piece of industry network data (h0, r0, t0), two new entity nodes h0 and t0 are added to the node set, and a new directed edge r0 is added to the edge set. The node set is a hash table sorted by node name, with the node name as the key and the node sort as the value. If the initial node set is an empty hash table, after adding h0, the node set is {h0:0}, and after adding t0, the node set is {h0:0, t0:1}. The edge set is a hash table whose edge sort is incremented by 1. If the initial edge set is an empty hash table, after adding r0, the edge set is {r0:1}. Since two new elements are added to the node set, the rows and columns of the initial directed graph are each increased by 2, filled with 0s. The system assigns row and column information to entity nodes based on their corresponding numbers in the initial directed graph. Since the rows and columns of the initial directed graph are sequentially arranged starting from number 0, the number 0 for entity node h0 indicates that both the first row and first column of the initial directed graph contain the entity information h0. Similarly, the number 1 for entity node t0 indicates that both the second row and second column contain the entity information t0. The relationship r0 between entities h0 and t0 is represented by the number 1, which corresponds to the element in the first row and second column of the matrix representing the entity information. This method constructs a directed graph matrix for the industrial network. Therefore, the row and column information corresponding to the elements in the directed graph matrix reflects the entity information, and the elements in the directed graph matrix reflect the relationships between the entity information.

[0141] In some embodiments, the directed edge elements of a directed graph matrix are processed to obtain the corresponding undirected graph matrix. The row and column information of the directed graph matrix and its corresponding undirected graph matrix are the same, but their elements are different. The elements of the undirected graph matrix are symmetrical along the main diagonal. Therefore, by traversing the elements of the directed graph matrix, the values ​​of all directed edge elements with a directed edge relationship are changed to 1 for all undirected edge elements with an undirected edge relationship. Then, the undirected edge elements with an undirected edge relationship are assigned a value of 1 based on the element symmetrical to the main diagonal. This yields the undirected graph matrix corresponding to the directed graph matrix.

[0142] S202. Based on the undirected graph matrix, rearrange the undirected graph matrix to obtain the matrix to be decomposed.

[0143] In this embodiment, the row information of the undirected edge elements in the undirected graph matrix is ​​the entity information corresponding to the row number; the column information of the undirected edge elements is the entity information corresponding to the column number. In some embodiments, based on the row information of the undirected edge elements, the number of target undirected edge elements in each row of the undirected graph matrix is ​​determined. The target undirected edge elements are the undirected edge elements in the undirected graph matrix where the undirected edge relationship between the entity information exists. The row positions of the data in each row of the undirected graph matrix are rearranged from top to bottom according to the order of the number from smallest to largest. Based on the column information of the undirected edge elements, the columns of the undirected graph matrix are rearranged. Starting from the first row of the matrix, the columns of the target undirected edge elements in the first row are arranged to the right of the target diagonal. The columns of the target undirected edge elements in the next row are arranged to the right of the target diagonal without affecting the position of the target undirected edge elements in the previous row, until all the target undirected edge elements in the undirected graph matrix are arranged as much as possible within the first matrix area, thus obtaining the matrix to be decomposed.

[0144] S203. Determine the initial nodes and target decomposition degree of the matrix to be decomposed.

[0145] In this embodiment, the initial node is the node corresponding to the element at the target position on the target diagonal of the matrix to be decomposed. The target decomposition degree refers to the degree to which the number of target undirected edge elements in a row of the matrix meets the quantity requirement. Meeting the quantity requirement means meeting the minimum quantity requirement; that is, the target decomposition degree is the number of target undirected edge elements in the row with the fewest target undirected edge elements in the matrix. The target decomposition degree satisfies: d0 = min(B(i,U)), where d0 is the target decomposition degree, U is any matrix, i is the row sequence in matrix U, and B(i,U) is the number of target undirected edge elements in a row of the matrix.

[0146] S204. Based on the initial nodes and target decomposition degree of the matrix to be decomposed, perform matrix decomposition processing on the matrix to be decomposed to obtain submatrices and decomposition nodes.

[0147] In this embodiment, based on the initial node and target decomposition degree of the matrix to be decomposed, if the lowest decomposition degree of the matrix to be decomposed is less than or equal to the first decomposition degree, then the next node is determined based on the initial node of the matrix to be decomposed, wherein the next node is the next node on the target diagonal of the matrix to be decomposed that is adjacent to the initial node; the next node is taken as the first node; if the decomposition degree of the first node is greater than the first decomposition degree, then the second node is determined based on the first node, until the target node is found, wherein the second node is the next node on the target diagonal of the matrix to be decomposed that is adjacent to the first node, and the target node is the node on the target diagonal of the matrix to be decomposed that has a decomposition degree less than or equal to the first decomposition degree; based on the target node, the matrix to be decomposed is decomposed to obtain the target matrix and decomposition nodes, wherein the decomposition should preferentially select the target node whose decomposition degree is closer to the first decomposition degree for decomposition, and the target matrix is ​​the matrix obtained in the decomposition process; based on the target matrix, the process returns to determine the initial node and target decomposition degree of the target matrix, until the target decomposition degree of the target matrix is ​​greater than the first decomposition degree, and the target matrix is ​​taken as a submatrix to obtain the submatrix and decomposition nodes.

[0148] S205. Based on the submatrix and decomposed nodes, obtain the industry network knowledge graph.

[0149] In this embodiment, the relationship between two adjacent sub-matrices of a decomposition node is obtained based on the decomposition node; an industrial network knowledge graph is obtained based on the sub-matrices of the industrial network and the relationship between two adjacent sub-matrices of the decomposition node. The sub-matrices of the industrial network represent the largest sub-industrial clusters that meet the preset decomposition degree. The internal relationships of the sub-industrial clusters are close. The industrial network knowledge graph, which is composed of the sub-industrial clusters and the relationships between them, can better organize, manage, and utilize massive amounts of data. It can help enterprises extract valuable information from massive amounts of data and provide more accurate and comprehensive support for decision-making.

[0150] The industrial network knowledge graph construction method provided in this application simplifies the complex industrial relationship structure, making it easier to understand. Therefore, it constructs an undirected graph matrix based on industrial network data, processes this matrix into a decomposable matrix, and then performs knowledge graph mining through matrix decomposition to obtain the largest sub-industrial cluster that meets the preset decomposition degree. An observable, realistic, and quantifiable industrial network knowledge graph is constructed from the relationships between these sub-industrial clusters. This allows for better organization, management, and utilization of industrial network data, helping enterprises extract valuable information from massive amounts of data and providing more accurate and comprehensive support for decision-making. This achieves the technical effect of improving the reliability of industrial knowledge data.

[0151] Figure 3This is a schematic diagram of the structure of an industrial network knowledge graph construction device provided in an embodiment of this application. The device in this embodiment can be in the form of software and / or hardware. For example... Figure 3 As shown in the embodiment of this application, an industrial network knowledge graph construction device 300 includes: a first determining module 301, a second determining module 302, a decomposition module 303, and a third determining module 304.

[0152] The first determining module 301 is used to determine an undirected graph matrix, wherein the undirected graph matrix includes entity information elements and undirected edge elements, and the undirected edge elements represent the undirected relationship between entity information elements;

[0153] The second determining module 302 is used to obtain the matrix to be decomposed based on the undirected edge elements of the undirected graph matrix, wherein the matrix to be decomposed includes a first matrix region, and the target undirected edge elements are arranged in the undirected edge elements within the first matrix region.

[0154] Decomposition module 303 is used to perform matrix decomposition processing on the decomposition nodes in the matrix to be decomposed, to obtain submatrices and decomposition nodes. The decomposition nodes are determined based on the target decomposition degree and the first decomposition degree of the matrix to be decomposed. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed.

[0155] The third determining module 304 is used to obtain the industrial network knowledge graph based on the submatrix and decomposed nodes.

[0156] Optionally, the first determining module 301 can also be used for:

[0157] Acquire industrial network data, which includes entity information within the industrial network and the relationships between those entities.

[0158] Based on industry network data, a directed graph matrix is ​​constructed, which includes entity information elements and directed edge elements. Entity information elements represent entity information, and directed edge elements represent the directed relationships between entity information elements.

[0159] Determine the undirected graph matrix based on the entity information elements and directed edge elements in the directed graph matrix.

[0160] Optionally, the first determining module 301 can also be used for:

[0161] Encode the entity information in the industrial network data to obtain the data encoding of the entity information;

[0162] Based on the data encoding of entity information, determine the first and second dimension information of the initial directed graph matrix;

[0163] The directed graph matrix is ​​obtained based on the first and second dimension information of the initial directed graph matrix and the relationship between entity information in the industrial network data.

[0164] Optionally, the first determining module 302 can also be used for:

[0165] Based on the element attributes of the undirected edge elements in the undirected graph matrix, determine the target undirected edge element, as well as the first and second dimension information of the target undirected edge element. The element attributes of the undirected edge element include attributes that represent the existence of undirected relations.

[0166] Based on the first and second dimension information of the target undirected edge element, the position of the target undirected edge element is adjusted to the first matrix region of the undirected graph matrix to obtain the matrix to be decomposed. The first matrix region is determined according to the position of the target diagonal in the undirected graph matrix.

[0167] Optionally, the first determining module 303 can also be used for:

[0168] Determine the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed.

[0169] Based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed, the first decomposition node is determined. The decomposition degree of the first decomposition node is between the target decomposition degree and the first decomposition degree of the matrix to be decomposed, and is closer to the first decomposition degree than the decomposition degree of other nodes to be decomposed.

[0170] Based on the first decomposition node and the matrix to be decomposed, determine the first submatrix and the submatrix to be decomposed in the matrix to be decomposed;

[0171] The process of determining the first decomposition node in the submatrix to be decomposed is repeated, taking the decomposition degree of the child node to be decomposed as the decomposition degree of the child node to be decomposed, and the target decomposition degree of the submatrix to be decomposed as the target decomposition degree of the submatrix. This process continues until the decomposition degree of the child node to be decomposed is no longer between the target decomposition degree and the first decomposition degree of the submatrix. Then, the second decomposition node and the second submatrix are determined. The target decomposition degree of the submatrix to be decomposed is determined based on the target undirected edge elements in the submatrix to be decomposed.

[0172] Based on the first and second submatrices, the submatrices are obtained;

[0173] Based on the first and second decomposition nodes, the decomposition nodes are obtained.

[0174] Optionally, the first determining module 303 can also be used for:

[0175] Compare the target decomposition degree and the first decomposition degree of the matrix to be decomposed;

[0176] If the target decomposition degree of the matrix to be decomposed is less than or equal to the first decomposition degree, then the initial first decomposition node in the nodes to be decomposed is determined.

[0177] If there are two or more initial first decomposition nodes, the first decomposition node in the initial first decomposition nodes is determined based on the initial nodes to be decomposed in the nodes to be decomposed and the target diagonal in the undirected graph matrix.

[0178] Optionally, the first determining module 303 can also be used for:

[0179] If the target decomposition degree of the matrix to be decomposed is greater than the first decomposition degree, then the first decomposition degree is adjusted to obtain the second decomposition degree, wherein the second decomposition degree is greater than the target decomposition degree of the matrix to be decomposed.

[0180] Based on the second decomposition degree and the matrix to be decomposed, matrix decomposition is performed on the matrix to be decomposed to obtain submatrices and decomposition nodes.

[0181] Optionally, the first determining module 304 can also be used for:

[0182] Based on the decomposition nodes, determine the matrix relationships between the submatrices;

[0183] Based on the matrix relationships between the submatrices, an industrial network knowledge graph is obtained.

[0184] Figure 4 This is a hardware structure diagram of the electronic device provided in an embodiment of this application. For example... Figure 4 As shown, the electronic device 400 provided in this embodiment includes:

[0185] The device 400 includes at least one processor 401 and a memory 402. Optionally, the device 400 also includes a communication component 403. The processor 401, memory 402, and communication component 403 are connected via a bus 404.

[0186] In a specific implementation, at least one processor 401 executes computer execution instructions stored in memory 402, causing at least one processor 401 to perform the above-described method.

[0187] The specific implementation process of processor 401 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0188] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0189] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0190] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0191] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement a method for constructing an industrial network knowledge graph.

[0192] This application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements a method for constructing an industrial network knowledge graph.

[0193] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0194] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0195] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0196] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0197] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0198] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0199] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0200] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for constructing an industrial network knowledge graph, characterized in that, Applied to building platforms, including: Determine an undirected graph matrix, wherein the undirected graph matrix includes entity information elements and undirected edge elements, and the undirected edge elements represent the undirected relationships between the entity information elements; Based on the undirected edge elements of the undirected graph matrix, a matrix to be decomposed is obtained, wherein the matrix to be decomposed includes a first matrix region, and the target undirected edge elements among the undirected edge elements are arranged in the first matrix region; Matrix decomposition is performed on the decomposition nodes in the matrix to be decomposed to obtain submatrices and decomposition nodes. The decomposition nodes are determined based on the target decomposition degree of the matrix to be decomposed and a preset first decomposition degree. The target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed. Based on the submatrix and decomposed nodes, an industrial network knowledge graph is obtained.

2. The method according to claim 1, characterized in that, The determination of the undirected graph matrix includes: Acquire industrial network data of the industrial network, wherein the industrial network data includes entity information in the industrial network and the relationships between the entity information; Based on the industry network data, a directed graph matrix is ​​constructed, wherein the directed graph matrix includes entity information elements and directed edge elements, the entity information elements represent the entity information, and the directed edge elements represent the directed relationship between the entity information elements; The undirected graph matrix is ​​determined based on the entity information elements and directed edge elements in the directed graph matrix.

3. The method according to claim 2, characterized in that, The step of constructing a directed graph matrix based on the industry network data includes: The entity information in the industrial network data is encoded to obtain the data encoding of the entity information; Based on the data encoding of the entity information, determine the first and second dimension information of the initial directed graph matrix; Based on the first and second dimension information of the initial directed graph matrix and the relationship between entity information in the industrial network data, a directed graph matrix is ​​obtained.

4. The method according to claim 1, characterized in that, The step of obtaining the matrix to be decomposed based on the undirected edge elements of the undirected graph matrix includes: Based on the element attributes of the undirected edge elements in the undirected graph matrix, the target undirected edge element, the first dimension information and the second dimension information of the target undirected edge element are determined. The element attributes of the undirected edge element include attributes that characterize the existence of the undirected relationship. Based on the first and second dimension information of the target undirected edge element, the position of the target undirected edge element is adjusted to the first matrix region of the undirected graph matrix to obtain the matrix to be decomposed, wherein the first matrix region is determined according to the position of the target diagonal in the undirected graph matrix.

5. The method according to claim 1, characterized in that, The step of performing matrix decomposition on the decomposition nodes in the matrix to be decomposed to obtain submatrices and decomposition nodes includes: Determine the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed, wherein the target decomposition degree of the matrix to be decomposed is determined based on the target undirected edge elements in the matrix to be decomposed. Based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed, a first decomposition node is determined among the nodes to be decomposed. The decomposition degree of the first decomposition node is between the target decomposition degree of the matrix to be decomposed and the first decomposition degree, and is closer to the first decomposition degree than the decomposition degree of other nodes to be decomposed among the nodes to be decomposed. Based on the first decomposition node and the matrix to be decomposed, determine the first submatrix and the submatrix to be decomposed in the matrix to be decomposed; The process of determining the first decomposition node based on the decomposition node in the matrix to be decomposed, the decomposition degree of the decomposition node, and the target decomposition degree of the submatrix to be decomposed are repeated until the decomposition degree of the decomposition node is no longer between the target decomposition degree and the first decomposition degree. Then, the second decomposition node and the second submatrix are determined. The target decomposition degree of the submatrix to be decomposed is determined based on the target undirected edge elements in the submatrix to be decomposed. The submatrix is ​​obtained based on the first submatrix and the second submatrix; The decomposition node is obtained based on the first decomposition node and the second decomposition node.

6. The method according to claim 5, characterized in that, The step of determining the first decomposition node among the nodes to be decomposed based on the nodes to be decomposed in the matrix to be decomposed, the decomposition degree of the nodes to be decomposed, and the target decomposition degree of the matrix to be decomposed includes: Compare the target decomposition degree of the matrix to be decomposed with the first decomposition degree; If the target decomposition degree of the matrix to be decomposed is less than or equal to the first decomposition degree, then the initial first decomposition node in the nodes to be decomposed is determined; If there are two or more initial first decomposition nodes, then the first decomposition node in the initial first decomposition nodes is determined based on the initial nodes to be decomposed in the nodes to be decomposed and the target diagonal in the undirected graph matrix.

7. The method according to claim 6, characterized in that, After comparing the target decomposition degree of the matrix to be decomposed with the first decomposition degree, the method further includes: If the target decomposition degree of the matrix to be decomposed is greater than the first decomposition degree, then the first decomposition degree is adjusted to obtain a second decomposition degree, wherein the second decomposition degree is greater than the target decomposition degree of the matrix to be decomposed. Based on the second decomposition degree and the matrix to be decomposed, matrix decomposition is performed on the matrix to be decomposed to obtain submatrices and decomposition nodes.

8. The method according to claim 1, characterized in that, The process of obtaining the industry network knowledge graph based on the submatrix and decomposed nodes includes: Based on the decomposition nodes, determine the matrix relationships between the submatrices; Based on the matrix relationships between the sub-matrices, an industrial network knowledge graph is obtained.

9. An industrial network knowledge graph construction device, characterized in that, include: The first determining module is used to determine an undirected graph matrix, wherein the undirected graph matrix includes entity information elements and undirected edge elements, and the undirected edge elements represent the undirected relationship between the entity information elements; The second determining module is used to obtain a matrix to be decomposed based on the undirected edge elements of the undirected graph matrix, wherein the matrix to be decomposed includes a first matrix region, and the target undirected edge elements among the undirected edge elements are arranged in the first matrix region. The decomposition module is used to perform matrix decomposition processing on the decomposition nodes in the matrix to be decomposed to obtain submatrices and decomposition nodes. The decomposition nodes are determined according to the target decomposition degree of the matrix to be decomposed and a preset first decomposition degree. The target decomposition degree of the matrix to be decomposed is determined according to the target undirected edge elements in the matrix to be decomposed. The third determining module is used to obtain the industrial network knowledge graph based on the submatrix and decomposed nodes.

10. An electronic device, characterized in that, include: Processor; memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the industrial network knowledge graph construction method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the industrial network knowledge graph construction method as described in any one of claims 1 to 8.

12. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the industrial network knowledge graph construction method according to any one of claims 1 to 8.

Images