A method, device, equipment, medium and program product for processing a hypergraph structure

By calculating the edge-travel probability and vertex-travel probability in the hypergraph structure, combining the vertex weight and the number of hyper-edge connections, the problem of low fine-grainedness of the stationary probability distribution in the existing technology is solved, and a more accurate transaction relationship between objects is achieved Extraction and advanced information extraction.

CN115114485BActive Publication Date: 2024-11-26TENCENT TECHNOLOGY (SHENZHEN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210557942.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-19
Publication Date
2024-11-26
Estimated Expiration
2042-05-19

Smart Images

  • Figure CN115114485B_ABST
    Figure CN115114485B_ABST
Patent Text Reader

Abstract

The present application provides a method, device, equipment, storage medium and program product for processing a hypergraph structure; the method includes: in a hypergraph structure composed of multiple vertices and multiple hyperedges for connecting the vertices, determining the starting wandering vertex and at least one first adjacent hyperedge connected to the starting wandering vertex; obtaining the first vertex weight of each first adjacent hyperedge corresponding to the starting wandering vertex, and determining the number of vertices connected to each first adjacent hyperedge; determining the edge wandering probability of each first adjacent hyperedge; determining the vertex wandering probability of each vertex on the first adjacent hyperedge, and based on the edge wandering probability and the vertex wandering probability, determining the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transactional relationship during the wandering of the starting wandering vertex. Through the present application, the fine-grainedness of the extracted transactional relationship between objects can be effectively improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to a method, device, equipment, storage medium and program product for processing a hypergraph structure. Background Art

[0002] With the rapid development of Internet technology, hypergraphs are ubiquitous in real-world applications. Hypergraphs can effectively express the associations between entities, such as the associations between web page information and between different accounts. However, hypergraphs often correspond to a large number of vertices and a large number of hyperedges, and there will be a large amount of metadata between hyperedges and vertices, such as location information, geographic information, etc.

[0003] In related technologies, when performing graph calculations on all vertices of a hypergraph structure, random walks are usually performed in a uniformly distributed manner. As a result, the determined stationary probability distribution cannot fully express the high-order information of the hypergraph structure, resulting in a low granularity of the transaction relationship between the determined objects. Summary of the invention

[0004] The embodiments of the present application provide a method, device, computer-readable storage medium and computer program product for processing a hypergraph structure, which can more comprehensively extract high-order information of the hypergraph structure and effectively improve the granularity of the extracted transaction relationships between objects.

[0005] The technical solution of the embodiment of the present application is implemented as follows:

[0006] The present application provides a method for processing a hypergraph structure, including:

[0007] In a hypergraph structure composed of a plurality of vertices and a plurality of hyperedges for connecting the vertices, determining a starting wandering vertex and at least one first adjacent hyperedge connected to the starting wandering vertex;

[0008] The vertices are used to represent objects, and the hyperedges connecting the vertices are used to represent associated objects that have a transactional relationship with the objects;

[0009] Respectively obtain the first vertex weight of each of the first adjacent hyperedges corresponding to the starting wandering vertex, and determine the number of vertices connected by each of the first adjacent hyperedges;

[0010] Determining an edge walk probability of each of the first adjacent hyperedges based on the first vertex weight and the number of vertices connected by the first adjacent hyperedges;

[0011] Determine the vertex walk probability of each vertex on the first adjacent hyperedge, and based on the edge walk probability and the vertex walk probability, determine the probability that during the walk of the starting walk vertex, the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship.

[0012] The present application provides a hypergraph structure processing device, including:

[0013] A first determination module is used to determine a starting wandering vertex and at least one first adjacent hyperedge connected to the starting wandering vertex in a hypergraph structure composed of a plurality of vertices and a plurality of hyperedges connecting the vertices;

[0014] A second determination module is used to respectively obtain the first vertex weight of each of the first adjacent hyperedges corresponding to the starting wandering vertex, and determine the number of vertices connected by each of the first adjacent hyperedges;

[0015] A third determination module, configured to determine an edge walk probability of each of the first adjacent hyperedges based on the first vertex weight and the number of vertices connected by the first adjacent hyperedges;

[0016] The fourth determination module is used to determine the vertex walk probability of each vertex on the first adjacent hyperedge, and based on the edge walk probability and the vertex walk probability, determine the probability that during the walk of the starting walk vertex, the objects corresponding to each vertex in the hypergraph structure and the associated objects corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship.

[0017] An embodiment of the present application provides an electronic device, including:

[0018] A memory for storing executable instructions;

[0019] The processor is used to implement the hypergraph structure processing method provided in the embodiment of the present application when executing the executable instructions stored in the memory.

[0020] An embodiment of the present application provides a computer-readable storage medium storing executable instructions for causing a processor to execute instructions to implement a method for processing a hypergraph structure provided in an embodiment of the present application.

[0021] The embodiment of the present application provides a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. The processor of the computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the above-mentioned hypergraph structure processing method of the embodiment of the present application.

[0022] The embodiments of the present application have the following beneficial effects:

[0023] By determining the first adjacent hyperedge connected to the starting wandering vertex, and based on the first vertex weight of each first adjacent hyperedge corresponding to the starting wandering vertex and the number of vertices connected by the first adjacent hyperedge, the edge wandering probability of the first adjacent hyperedge is determined, and the vertex wandering probability of each vertex on each first adjacent hyperedge is determined, so as to determine the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship in the process of wandering of the starting wandering vertex based on the edge wandering probability and the vertex wandering probability. Since the first vertex weight of the first adjacent hyperedge corresponding to the starting wandering vertex and the number of vertices connected by the first adjacent hyperedge are comprehensively considered when determining the edge wandering probability of the first adjacent hyperedge, and since the first vertex weight and the number of vertices connected by the first adjacent hyperedge can measure the importance between objects, the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship in the process of wandering of the starting wandering vertex can comprehensively extract the high-order information of the hypergraph structure, thereby effectively improving the fine-grainedness of the transaction relationship between the extracted objects. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a structural diagram of a processing system architecture of a hypergraph structure provided in an embodiment of the present application;

[0025] Figure 2 is a structural schematic diagram of a hypergraph structure processing device provided in an embodiment of the present application;

[0026] 3A to 3H It is a flowchart of a method for processing a hypergraph structure provided in an embodiment of the present application;

[0027] FIG. 4A to FIG. 4C It is a schematic diagram of the principle of the method for processing the hypergraph structure provided in the embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0029] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0030] In the following description, the terms "first\second\third" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0032] Before further describing the embodiments of the present application in detail, the nouns and terms involved in the embodiments of the present application are explained. The nouns and terms involved in the embodiments of the present application are subject to the following interpretations.

[0033] 1) Hypergraph structure: also known as hypergraph data structure, hypergraph structure includes multiple hyperedges, which connect at least two vertices of the hypergraph structure. Hypergraph data structure is also known as "hyper-relation" model, and its essence is a topological data structure model. Each fixed point of the hypergraph belongs to at least one edge, and can also belong to a closed curve. It represents basic geographic entities, complex geographic entities, and their spatial and non-spatial attributes through four basic units: category, object, attribute and relationship. The advantage of hypergraph data structure is that it can distinguish between hierarchical and non-hierarchical structures of information, and any complex information structure can be established by these two types of information.

[0034] 2) Hyperedge: It is the edge of the hypergraph structure. The hyperedge connects at least two vertices of the hypergraph structure.

[0035] 3) Vertex: It is the vertex of the hypergraph data structure and the smallest component unit of the hypergraph data structure.

[0036] 4) Leaf vertex: A vertex in a hypergraph data structure that has no child vertices.

[0037] 5) Hyper edge Degree: The number of vertices connected by the hyper edge.

[0038] 6) Node Degree: The number of hyperedges connected to any vertex.

[0039] 7) Random walk: The basic idea of ​​random walk is to traverse a graph from one or a series of vertices. At any vertex, the traverser will walk to the neighboring vertices of this vertex with probability 1-a, and randomly jump to any vertex in the graph with probability a. a is called the jump occurrence probability. After each walk, a probability distribution is obtained, which describes the probability of each vertex in the graph being visited. Use this probability distribution as the input for the next walk and iterate this process repeatedly. When certain preconditions are met, this probability distribution will tend to converge. After convergence, a stable probability distribution can be obtained.

[0040] 8) Stationary Probability Distribution: A stationary probability distribution is a probability distribution. A stationary distribution (of a Markov chain) is a probability distribution that has some invariant (or stationary) property.

[0041] During the implementation of the embodiments of the present application, the applicant discovered that the related technology has the following problems:

[0042] In the related art, for the random walk of the hypergraph structure, a hyperedge is usually selected from the hyperedges connected to the current vertex according to a uniform distribution, and then a vertex is selected from the vertices connected to the current edge according to a uniform distribution. The entire hypergraph structure is facilitated by random walks to obtain a stable probability distribution of the hypergraph structure. However, the stable probability distribution determined by the walking scheme of the related art, because the probability is determined in a uniformly distributed manner during the random walk, the final stable probability distribution does not take into account the importance between the vertices of the hypergraph structure, resulting in the determined stable probability distribution being unable to reflect the fine-grained information of the hypergraph structure, resulting in the low accuracy of the determined stable probability distribution.

[0043] The embodiments of the present application provide a method, apparatus, device, computer-readable storage medium and computer program product for processing a hypergraph structure, which can comprehensively extract high-order information of the hypergraph structure, thereby effectively improving the granularity of the transactional relationship between the extracted objects. The following describes an exemplary application of the hypergraph structure processing device provided by the embodiments of the present application. The device provided by the embodiments of the present application can be implemented as various types of user terminals such as laptop computers, tablet computers, desktop computers, set-top boxes, mobile devices (for example, mobile phones, portable music players, personal digital assistants, dedicated messaging devices, portable gaming devices), and can also be implemented as a server.

[0044] See also Figure 1 , Figure 1 It is an architectural diagram of a hypergraph structure processing system 100 provided in an embodiment of the present application. To realize an application scenario of hypergraph structure processing, a terminal (terminal 400 is shown as an example) is connected to a server 200 via a network 300. The network 300 may be a wide area network or a local area network, or a combination of the two.

[0045] The terminal 400 is used for the user to use the client 410, which is displayed on a graphical interface 410-1 (graphic interface 410-1 is shown as an example). The terminal 400 and the server 200 are connected to each other via a wired or wireless network.

[0046] In some embodiments, the server 200 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal 400 may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, an intelligent voice interaction device, a smart home appliance, a car terminal, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected via wired or wireless communication, which is not limited in the embodiments of the present application.

[0047] In some embodiments, the server 200 determines the stationary probability distribution of the hypergraph structure based on the acquired hypergraph structure, and sends the stationary probability distribution of the hypergraph structure to the terminal 400 for display.

[0048] In some other embodiments, the terminal 400 receives the hypergraph structure sent by the server 200 , the terminal 400 determines the stationary probability distribution of the hypergraph structure, and displays the stationary probability distribution of the hypergraph structure at the terminal 400 .

[0049] In other embodiments, the embodiments of the present application can be implemented with the aid of cloud technology. Cloud technology refers to a hosting technology that unifies a series of resources such as hardware, software, and networks within a wide area network or a local area network to achieve data computing, storage, processing, and sharing.

[0050] Cloud technology is a general term for network technology, information technology, integration technology, management platform technology, and application technology based on the cloud computing business model. It can form a resource pool that can be used on demand and is flexible and convenient. Cloud computing technology will become an important support. The background services of the technical network system require a large amount of computing and storage resources.

[0051] See also Figure 2 , Figure 2 is a structural diagram of a server 200 of the hypergraph structure processing method provided in an embodiment of the present application, Figure 2The server 200 shown includes: at least one processor 210, a memory 250, and at least one network interface 220. The various components in the server 200 are coupled together via a bus system 240. It is understood that the bus system 240 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 240 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, the bus system 240 is not described in detail. Figure 2 Various buses are labeled as bus system 240 .

[0052] The processor 210 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., where the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0053] The memory 250 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disk drives, etc. The memory 250 may optionally include one or more storage devices that are physically remote from the processor 210.

[0054] The memory 250 includes a volatile memory or a non-volatile memory, and may also include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM), and the volatile memory may be a random access memory (RAM). The memory 250 described in the embodiments of the present application is intended to include any suitable type of memory.

[0055] In some embodiments, memory 250 can store data to support various operations, examples of which include programs, modules, and data structures, or a subset or superset thereof, as exemplarily described below.

[0056] The operating system 251 includes system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic businesses and process hardware-based tasks.

[0057] The network communication module 252 is used to reach other electronic devices via one or more (wired or wireless) network interfaces 220. Exemplary network interfaces 220 include: Bluetooth, Wireless LAN (WiFi), and Universal Serial Bus (USB).

[0058] In some embodiments, the hypergraph structure processing device provided in the embodiments of the present application can be implemented in software. Figure 2 The processing device 255 of the hypergraph structure stored in the memory 250 is shown, which can be software in the form of a program and a plug-in, etc., including the following software modules: a first determination module 2551, a second determination module 2552, a third determination module 2553, and a fourth determination module 2554. These modules are logical, and therefore can be arbitrarily combined or further split according to the functions implemented. The functions of each module will be described below.

[0059] The processing method of the hypergraph structure provided in the embodiment of the present application will be described in combination with the exemplary application and implementation of the server or terminal provided in the embodiment of the present application. Figure 3A , Figure 3A is a flow chart of a method for processing a hypergraph structure provided in an embodiment of the present application, which will be combined with Figure 3A Steps 101 to 105 are shown for explanation.

[0060] In step 101, in a hypergraph structure consisting of a plurality of vertices and a plurality of hyperedges for connecting the vertices, a starting wandering vertex and at least one first adjacent hyperedge connected to the starting wandering vertex are determined.

[0061] In some embodiments, the starting wandering vertex may be any vertex in the hypergraph structure, and the first adjacent hyperedge is a hyperedge connected to the starting wandering vertex.

[0062] In some embodiments, the hypergraph structure includes a plurality of hyperedges, and the hyperedges connect at least two vertices of the hypergraph structure. Figure 4A , Figure 4A Schematic diagram of the principle of the method for processing the hypergraph structure provided in the embodiment of the present application. Figure 4A In the hypergraph structure shown, the starting wandering vertex is determined to be vertex 40 , and the first adjacent hyperedges connected to the starting wandering vertex 40 are: first adjacent hyperedge 51 , first adjacent hyperedge 52 , first adjacent hyperedge 53 , and first adjacent hyperedge 54 .

[0063] As an example, see Figure 4A , the first adjacent hyperedge 51 connects the starting wandering vertex 40, vertex 48 and vertex 47; the first adjacent hyperedge 52 connects the starting wandering vertex 40, vertex 45 and vertex 46; the first adjacent hyperedge 53 connects the starting wandering vertex 40, vertex 43 and vertex 44. The first adjacent hyperedge 54 connects the starting wandering vertex 40, vertex 41 and vertex 42.

[0064] In some embodiments, vertices are used to represent objects, and hyperedges connecting vertices are used to represent associated objects that have a transactional relationship with the objects.

[0065] As an example, in the application scenario of influence ranking, the object can be the author of the paper. The hyperedge connecting the authors of the paper is used to represent multiple associated objects that have a transactional relationship with the object. Each paper represents a hyperedge. A paper corresponds to multiple authors, and each author is a vertex on the hyperedge. By determining the stationary distribution of the hypergraph structure and sorting the probability values ​​in the distribution, the author influence ranking can be obtained. For a vertex with a larger probability value, the academic influence of the author corresponding to the vertex is greater.

[0066] In the application scenario of influence ranking, in a hypergraph structure composed of multiple authors and multiple papers connecting the authors, determine the starting walk vertex (i.e., any target author) and at least one first adjacent hyperedge (target paper associated with the target author) connected to the starting walk vertex (any target author), obtain the first vertex weight of each associated paper corresponding to the target author, and determine the number of authors connected to each target paper; determine the edge walk probability of each associated paper based on the first vertex weight of each associated paper corresponding to the target author and the number of authors connected to each target paper; determine the vertex walk probability of each author on the associated paper, and determine the probability that each author in the hypergraph structure and the associated authors corresponding to the hyperedge associated with the target author have jointly published the same paper during the retrieval / search (walking) process with the target author based on the edge walk probability and the vertex walk probability. As an example, in the application scenario of Internet link analysis, the object can be an Internet link, and the hyperedge connecting the Internet links is used to represent multiple associated objects that have a transactional relationship with the object. Each browser represents a hyperedge. Each browser often clicks on multiple Internet links when browsing web pages. Each Internet link corresponds to a vertex of a hypergraph structure. By determining the stationary distribution of the hypergraph structure and sorting the probability values ​​in the distribution, the popularity ranking of the Internet links can be obtained. For a vertex with a larger probability value, the popularity of the Internet link corresponding to the vertex is greater.

[0067] As an example, in the application scenario of email, the object can be the email subject, and the hyperedge connecting the email subject is used to represent multiple associated objects that have a transactional relationship with the object. Each email account in the email sending and receiving list represents a hyperedge. Some email accounts in the email sending and receiving list may have sent emails with different email subjects multiple times. Each email subject corresponds to one or more vertices of the hypergraph structure. By determining the stationary distribution of the hypergraph structure and sorting the probability values ​​in the distribution, the sending and receiving frequency of the email subject can be obtained. For a vertex with a larger probability value, the sending and receiving frequency of the email subject corresponding to the vertex is higher.

[0068] As an example, in the application scenario of chemical molecular structure, the object can be a single atom of a chemical substance, and the hyperedges connecting the atoms are used to represent multiple associated objects that have a transactional relationship with the object. The molecular structure of a certain chemical substance represents a hypergraph structure, and each vertex of the hypergraph structure represents a single atom of the chemical substance. The hyperedges with a degree greater than 2 represent the central bond of the molecular structure, and the hyperedges with a degree equal to 2 represent the covalent bond. By determining the stationary distribution of the hypergraph structure and sorting the probability values ​​in the distribution, the probability of occurrence of various atoms in the chemical substance can be obtained. For a vertex with a larger probability value, the importance of the atom corresponding to the vertex is higher.

[0069] As an example, in an application scenario of remote communication, the object may be a communication participant, and the hyperedge connecting the communication participants is used to represent multiple associated objects that have a transactional relationship with the object. Each communication participant represents a vertex of the hypergraph structure, and the hyperedge of the hypergraph structure represents the communication channel between the communication participants. By determining the stationary distribution of the hypergraph structure and sorting the probability values ​​in the distribution, the importance of each communication participant in the communication scenario can be obtained. For a vertex with a larger probability value, the communication participant corresponding to the vertex is more important in the remote communication. For example, when a communication participant is a signal relay base station, the communication participant is more important in the remote communication.

[0070] In the application scenario of remote communication, in a hypergraph structure composed of multiple communication participants and multiple channels for connecting the communication participants, determine the starting wandering vertex (i.e., the communication initiator) and at least one first adjacent hyperedge (i.e., the adjacent channel) connected to the communication initiator; obtain the first vertex weight of the communication initiator corresponding to each adjacent channel respectively, and determine the number of communication initiators connected to each first adjacent hyperedge; determine the edge wandering probability of each adjacent channel based on the first vertex weight and the number of vertices connected to the adjacent channel; determine the vertex wandering probability of each communication participant on the adjacent channel, and determine the probability that, during the wandering (communication) process of the communication initiator, the object corresponding to each vertex in the hypergraph structure (communication participant) and the communication participant of the hyperedge associated with the communication initiator have a transaction relationship (establish a communication connection) based on the edge wandering probability and the vertex wandering probability.

[0071] In step 102, the first vertex weights of the starting wandering vertices corresponding to the first adjacent hyperedges are respectively obtained, and the number of vertices connected to the first adjacent hyperedges is determined.

[0072] In some embodiments, each first adjacent hyperedge corresponds to a first vertex weight.

[0073] As an example, see Figure 4A, respectively obtain the first vertex weight of the first adjacent hyperedge 51 corresponding to the starting wandering vertex 40, the first vertex weight of the first adjacent hyperedge 52 corresponding to the starting wandering vertex 40, the first vertex weight of the first adjacent hyperedge 53 corresponding to the starting wandering vertex 40, and the first vertex weight of the first adjacent hyperedge 54 corresponding to the starting wandering vertex 40.

[0074] As an example, see Figure 4A , the number of vertices connected by the first adjacent hyperedge 51 is 3, the number of vertices connected by the first adjacent hyperedge 52 is 3, the number of vertices connected by the first adjacent hyperedge 53 is 3, and the number of vertices connected by the first adjacent hyperedge 54 is 3.

[0075] In this way, by obtaining the first vertex weight of the starting walk vertex corresponding to each first adjacent hyperedge and determining the number of vertices connected by each first adjacent hyperedge, it is convenient to subsequently determine the edge walk probability based on the first vertex weight and the number of vertices connected by the first adjacent hyperedge, and then determine the stationary probability distribution of the hypergraph structure based on the edge walk probability, thereby effectively improving the accuracy of the stationary probability distribution.

[0076] In step 103, the edge walking probability of each first adjacent hyperedge is determined based on the first vertex weight and the number of vertices connected by the first adjacent hyperedge.

[0077] In some embodiments, the edge walk probability of any hyperedge of the hypergraph structure represents the probability of the hyperedge being reached during the walk of the starting walk vertex.

[0078] In some embodiments, see Figure 3B , Figure 3B A schematic flow chart of a method for processing a hypergraph structure provided in an embodiment of the present application, when the number of first adjacent hyperedges is at least two, Figure 3B The illustrated step 103 may be implemented by executing the following steps 1031 to 1033 for each first adjacent hyperedge respectively.

[0079] In step 1031, the walking weight of the first adjacent hyperedge is determined based on the first vertex weight and the number of vertices connected by the first adjacent hyperedge.

[0080] In some embodiments, the walking weight of the first neighboring hyperedge is used to determine the edge walking probability of the first neighboring hyperedge.

[0081] In some embodiments, see Figure 3C , Figure 3C A schematic diagram of a process flow of a method for processing a hypergraph structure provided in an embodiment of the present application, Figure 3C The illustrated step 1031 may be implemented by following steps 10311 to 10313 .

[0082] In step 10311, the vertex weight matrix of the first adjacent hyperedge is obtained.

[0083] In some embodiments, the vertex weight matrix of the first adjacent hyperedge can be designed in different forms in different application scenarios. For example, the designed vertex weight matrix is ​​different in the application scenario of influence ranking and in the application scenario of Internet link analysis.

[0084] In step 10312, the first vertex weight and the number of vertices connected to the first adjacent hyperedge are multiplied to obtain a multiplication result.

[0085] As an example, the expression of the multiplication result can be:

[0086] Z=w(e)δ(e) (1)

[0087] Among them, Z represents the multiplication result, w(e) represents the weight of the first vertex, and δ(e) represents the number of vertices connected by the first adjacent hyperedge.

[0088] In step 10313, based on the multiplication result and the vertex weight matrix, the walking weight of the first adjacent hyperedge is determined.

[0089] In some embodiments, the above step 10313 can be implemented in the following way: call a control function to process the number of vertices connected by the first adjacent hyperedge to obtain a control processing result; wherein the control function is used to control the influence of the first vertex weight on the walking weight; multiply the multiplication processing result, the vertex weight matrix and the control processing result to obtain the walking weight of the first adjacent hyperedge.

[0090] In some embodiments, the control function may be a monotonic real-valued function for controlling the influence of the first vertex weight on the walk weight.

[0091] As an example, the expression that controls the processing result can be:

[0092] ρ(δ(e)) (2)

[0093] Among them, ρ(.) represents the control function, δ(e) represents the number of vertices connected by the first adjacent hyperedge, and ρ(δ(e)) represents the control processing result.

[0094] As an example, the expression of the walk weight of the first adjacent hyperedge can be:

[0095] X=w(e)δ(e)ρ(δ(e))Q 1 (u,e) (3)

[0096] Among them, Q 1(u,e) represents the vertex weight matrix of the first adjacent hyperedge e corresponding to the starting wandering vertex u, δ(e) represents the number of vertices connected by the first adjacent hyperedge, ρ(δ(e)) represents the control processing result, w(e)δ(e) represents the multiplication processing result, and X represents the wandering weight of the first adjacent hyperedge.

[0097] In step 1032, the walking weights of each first adjacent hyperedge are summed to obtain the edge walking weight of the starting walking vertex.

[0098] As an example, the expression of the edge walk weight of the starting walk vertex can be:

[0099] Y=∑X=∑w(e)δ(e)ρ(δ(e))Q 1 (u,e) (4)

[0100] Among them, Y represents the edge walking weight of the starting walking vertex, X represents the walking weight of each first adjacent hyperedge, and Q 1 (u,e) represents the vertex weight matrix of the first adjacent hyperedge e corresponding to the starting wandering vertex u, δ(e) represents the number of vertices connected by the first adjacent hyperedge, ρ(δ(e)) represents the control processing result, and w(e)δ(e) represents the multiplication processing result.

[0101] In step 1033, the ratio of the walk weight to the edge walk weight is determined as the edge walk probability of the first adjacent hyperedge.

[0102] As an example, the expression of the edge walk probability of the first adjacent hyperedge can be:

[0103]

[0104] Among them, P 1 represents the edge walk probability of the first adjacent hyperedge, Y represents the edge walk weight of the starting walk vertex, and X represents the walk weight of the first adjacent hyperedge.

[0105] In this way, by determining the edge walk probability of the first adjacent hyperedge, and then determining the stationary probability distribution of the hypergraph structure according to the edge walk probability of the first adjacent hyperedge, the accuracy of the determined stationary probability distribution is guaranteed.

[0106] In step 104, the vertex walk probability of each vertex on the first adjacent hyperedge is determined.

[0107] In some embodiments, vertex walk probability is used to determine the stationary probability distribution of the hypergraph structure.

[0108] In some embodiments, the vertex walk probability of any vertex in the hypergraph structure represents the probability of the vertex being reached during the walk of the starting walk vertex.

[0109] In some embodiments, see Figure 3B , Figure 3B A schematic diagram of a process flow of a method for processing a hypergraph structure provided in an embodiment of the present application, Figure 3B The illustrated step 104 may be implemented by executing the following steps 1041 to 1042 for each first adjacent hyperedge respectively.

[0110] In step 1041, the second vertex weight of each vertex on the first adjacent hyperedge corresponding to the first adjacent hyperedge is obtained.

[0111] As an example, see Figure 4A , for the first adjacent hyperedge 51, obtain the second vertex weights of the first adjacent hyperedge 51 corresponding to the vertices 40, 48 and 47 on the first adjacent hyperedge 51 respectively; for the first adjacent hyperedge 52, obtain the second vertex weights of the first adjacent hyperedge 52 corresponding to the vertices 40, 45 and 46 on the first adjacent hyperedge 52 respectively; for the first adjacent hyperedge 53, obtain the second vertex weights of the first adjacent hyperedge 53 corresponding to the vertices 40, 43 and 44 on the first adjacent hyperedge 53 respectively; for the first adjacent hyperedge 54, obtain the second vertex weights of the first adjacent hyperedge 54 corresponding to the vertices 40, 42 and 41 on the first adjacent hyperedge 54 respectively.

[0112] In step 1042, based on the second vertex weight, the vertex walk probability of each vertex on the first adjacent hyperedge is determined.

[0113] In this way, the vertex wandering probability of each vertex on the first adjacent hyperedge is determined by corresponding each vertex on the first adjacent hyperedge to the second vertex weight of the first adjacent hyperedge. Since the second vertex weight can measure the importance of the vertices on the first adjacent hyperedge, the vertex wandering probability is determined by the second vertex weight, so that the vertex wandering probability can more comprehensively reflect the high-order information of the hypergraph structure, and the stationary probability distribution determined according to the vertex wandering probability has a higher granularity and can comprehensively reflect the high-order information of the hypergraph structure.

[0114] In some embodiments, see Figure 3D , Figure 3D A schematic diagram of a process flow of a method for processing a hypergraph structure provided in an embodiment of the present application, Figure 3D The illustrated step 1042 may be implemented by following steps 10421 to 10422 .

[0115] In step 10421, the weights of the second vertices of each vertex on the first adjacent hyperedge corresponding to the first adjacent hyperedge are summed to obtain a weight sum.

[0116] As an example, the expression for the sum of weights can be:

[0117] F=∑Q 2 (v,e) (6)

[0118] Among them, F represents the weight and Q 2 (v,e) represents the second vertex weight of the vertex v on the first adjacent hyperedge e corresponding to the first adjacent hyperedge e.

[0119] In step 10422, the following processing is performed for each vertex on the first adjacent hyperedge: the ratio of the second vertex weight corresponding to the first adjacent hyperedge of the vertex to the sum of the weights is determined as the vertex wandering probability of the vertex.

[0120] As an example, the expression of vertex walk probability can be:

[0121]

[0122] Among them, P 2 represents the vertex wandering probability, F represents the weight and, Q 2 (v,e) represents the second vertex weight of the vertex v on the first adjacent hyperedge e corresponding to the first adjacent hyperedge e.

[0123] In this way, the vertex walk probability of each vertex on each first adjacent hyperedge can be determined, and then the stationary probability distribution of the hypergraph structure can be accurately determined based on the vertex walk probability and the edge walk probability, thereby effectively improving the accuracy of the stationary probability distribution.

[0124] In step 105, based on the edge walk probability and the vertex walk probability, the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship during the walk of the starting walk vertex is determined.

[0125] In some embodiments, the stationary probability distribution includes the probability that there is a transaction relationship between the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex during the wandering of the starting wandering vertex. The stationary probability distribution is used to characterize the probability of each vertex in the hypergraph structure being reached during the wandering of the starting wandering vertex.

[0126] In some embodiments, see Figure 3E , Figure 3E A schematic diagram of a process flow of a method for processing a hypergraph structure provided in an embodiment of the present application, Figure 3E The illustrated step 105 may be implemented by following steps 1051 to 1053 .

[0127] In step 1051, the following processing is performed for each vertex on each first adjacent hyperedge: the edge walk probability of the first adjacent hyperedge and the vertex walk probability of the vertex are multiplied to obtain the probability transfer value of the vertex.

[0128] As an example, see Figure 4A , perform the following processing for vertex 48 on the first adjacent hyperedge 51: multiply the edge walk probability of the first adjacent hyperedge 51 and the vertex walk probability of vertex 48 to obtain the probability transfer value of vertex 48; perform the following processing for vertex 47 on the first adjacent hyperedge 51: multiply the edge walk probability of the first adjacent hyperedge 51 and the vertex walk probability of vertex 47 to obtain the probability transfer value of vertex 47; perform the following processing for vertex 46 on the first adjacent hyperedge 52: multiply the edge walk probability of the first adjacent hyperedge 52 and the vertex walk probability of vertex 46 to obtain the probability transfer value of vertex 46; perform the following processing for vertex 44 on the first adjacent hyperedge 53: multiply the edge walk probability of the first adjacent hyperedge 53 and the vertex walk probability of vertex 44 to obtain the probability transfer value of vertex 44.

[0129] As an example, see Figure 4A , the expression of the probability transition value of vertex 44 can be:

[0130] P=P 1 P 2 (8)

[0131] Among them, P 1 Characterizes the edge walk probability of the first adjacent hyperedge 53, P 2 represents the vertex wandering probability of vertex 44 , and P represents the probability transition value of vertex 44 .

[0132] In step 1052, a probability transfer matrix is ​​constructed according to the probability transfer value of each vertex.

[0133] As an example, see Figure 4B , Figure 4B It is a schematic diagram of the principle of the method for processing the hypergraph structure provided in the embodiment of the present application. Figure 4B The expression of the probability transfer matrix corresponding to the hypergraph structure shown can be:

[0134] Q=(Q 1 , Q 2 , Q 3 ) (9)

[0135] Among them, Q represents Figure 4B The probability transfer matrix corresponding to the hypergraph structure shown, Q 1 Characterizes the probability transition value of vertex 60, Q 2The probability transition value characterizing vertex 61, Q 3 The probability transition value characterizing vertex 62.

[0136] In step 1053, based on the probability transition matrix, determine the stationary probability distribution of the hypergraph structure, and determine the corresponding probability in the stationary probability distribution of the hypergraph structure as the probability that there is a transaction relationship between the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex during the process of walking from the starting walk vertex.

[0137] In some embodiments, refer to Figure 3F , Figure 3F the flowchart of the method for processing the hypergraph structure provided by the embodiments of the present application, Figure 3F In step 1053 shown, determining the stationary probability distribution of the hypergraph structure based on the probability transition matrix can be implemented through the following steps 10531 to 10533.

[0138] In step 10531, determine the probability transition matrix as the first iteration probability distribution of the hypergraph structure, and based on the probability transition value of each vertex in the first iteration probability distribution, determine the second iteration probability distribution of the hypergraph structure.

[0139] As an example, the expression of the second iteration probability distribution of the hypergraph structure can be:[[]]

[0140] π 2 =Kπ 1 (10)

[0141] where, π 2 characterizes the second iteration probability distribution of the hypergraph structure, π 1 characterizes the first iteration probability distribution of the hypergraph structure, and K characterizes the transformation parameter.

[0142] In step 10532, based on the probability transition value of each vertex in the i-th iteration probability distribution of the hypergraph structure, determine the (i + 1)-th iteration probability distribution of the hypergraph structure;

[0143] where, 2 ≤ i < N, the difference between the probability transition value of any vertex in the (N - 1)-th iteration probability distribution and the probability transition value of the corresponding vertex in the N-th iteration probability distribution is less than the stationary threshold.

[0144] In some embodiments, the stability threshold can be set according to demand. When the difference between the probability transition value of any vertex in the N-1th iteration probability distribution and the probability transition value of the corresponding vertex in the Nth iteration probability distribution is less than the stability threshold, it indicates that the probability transition value of the corresponding vertex in the Nth iteration probability distribution is less different from the probability transition value of any vertex in the N-1th iteration probability distribution, that is, the probability transition value of the corresponding vertex in the Nth iteration probability distribution has tended to be stable.

[0145] As an example, the expression of the probability distribution of the i+1th iteration of the hypergraph structure can be:

[0146] π i+1 =Kπ i (11)

[0147] Among them, π i+1 Characterizes the probability distribution of the i+1th iteration of the hypergraph structure, π i represents the probability distribution of the i-th iteration of the hypergraph structure, and K represents the transformation parameter.

[0148] In step 10533, the Nth iteration probability distribution is determined as the stationary probability distribution of the hypergraph structure.

[0149] In this way, since the probability transfer values ​​of the corresponding vertices in the Nth iteration probability distribution have tended to be stable, the Nth iteration probability distribution can be determined as the stable probability distribution of the hypergraph structure, so that the probability transfer matrix can be accurately determined through multiple iterations to obtain the stable probability distribution of the hypergraph structure.

[0150] In some embodiments, see Figure 3G , Figure 3G A schematic diagram of a process flow of a method for processing a hypergraph structure provided in an embodiment of the present application, Figure 3G The illustrated step 105 may be implemented by following the steps 1054 to 1055 .

[0151] In step 1054, the association relationship between the leaf vertex of each first adjacent hyperedge and the hypergraph structure is determined.

[0152] In some embodiments, the association relationship represents whether the leaf vertex of each first adjacent hyperedge is a leaf vertex of the hypergraph structure.

[0153] In step 1055, when the leaf vertices of each first adjacent hyperedge represented by the association relationship are all leaf vertices of the hypergraph structure, based on the edge walk probability and the vertex walk probability, the probability that a transaction relationship exists between the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex during the walk of the starting walk vertex is determined.

[0154] As an example, see Figure 4A , the leaf vertices 48 and 47 of the first adjacent hyperedge 51 are both leaf vertices of the hypergraph structure; the leaf vertices 45 and 46 of the first adjacent hyperedge 52 are both leaf vertices of the hypergraph structure; the leaf vertices 43 and 44 of the first adjacent hyperedge 53 are both leaf vertices of the hypergraph structure; the leaf vertices 41 and 42 of the first adjacent hyperedge 54 are both leaf vertices of the hypergraph structure. That is, the leaf vertices of each first adjacent hyperedge are leaf vertices of the hypergraph structure. Then, based on the edge walk probability and the vertex walk probability, the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship during the walk of the starting walk vertex can be determined.

[0155] In this way, when the association relationship represents that the leaf vertices of each first adjacent hyperedge are all leaf vertices of the hypergraph structure, the stable probability distribution of the hypergraph structure can be determined directly based on the edge walk probability and the vertex walk probability. Therefore, by determining the association relationship, the stable probability distribution of the hypergraph structure can be fully and comprehensively determined.

[0156] In some embodiments, see Figure 3H , Figure 3H A schematic diagram of a process flow of a method for processing a hypergraph structure provided in an embodiment of the present application, Figure 3H The illustrated step 105 may be implemented by following steps 1056 to 1058 .

[0157] In step 1056, when there is a second adjacent hyperedge in at least one first adjacent hyperedge, wherein the leaf vertex of the second adjacent hyperedge is not a leaf vertex of the hypergraph structure, the following processing is performed on the leaf vertex of the second adjacent hyperedge: determining at least one third adjacent hyperedge connected to the leaf vertex of the second adjacent hyperedge.

[0158] As an example, see Figure 4C , Figure 4C It is a schematic diagram of the principle of the method for processing the hypergraph structure provided by the embodiment of the present application. When there is a second adjacent hyperedge 81 among the first adjacent hyperedge 81, the first adjacent hyperedge 82, the first adjacent hyperedge 83 and the first adjacent hyperedge 84, wherein the leaf vertex 77 of the second adjacent hyperedge 81 is not a leaf vertex of the hypergraph structure, the following processing is performed on the leaf vertex 77 of the second adjacent hyperedge 81: determine the third adjacent hyperedge 60 connected to the leaf vertex 77 of the second adjacent hyperedge.

[0159] As an example, see Figure 4C, when there is a second adjacent hyperedge 82 among the first adjacent hyperedge 81, the first adjacent hyperedge 82, the first adjacent hyperedge 83 and the first adjacent hyperedge 84, wherein the leaf vertex 75 of the second adjacent hyperedge 82 is not a leaf vertex of the hypergraph structure, the following processing is performed on the leaf vertex 75 of the second adjacent hyperedge 82: determine the third adjacent hyperedge 85 connected to the leaf vertex 75 of the second adjacent hyperedge.

[0160] As an example, see Figure 4C In the first adjacent hyperedge 83, the leaf vertex 73 and the leaf vertex 74 of the first adjacent hyperedge 83 are both leaf vertices of the hypergraph structure, so the first adjacent hyperedge 83 does not belong to the second adjacent hyperedge.

[0161] As an example, see Figure 4C In the first adjacent hyperedge 84, the leaf vertex 72 and the leaf vertex 71 of the first adjacent hyperedge 84 are both leaf vertices of the hypergraph structure, so the first adjacent hyperedge 84 does not belong to the second adjacent hyperedge.

[0162] In step 1057, the edge walk probability of each third adjacent hyperedge and the vertex walk probability of each vertex on the third adjacent hyperedge are determined.

[0163] As an example, see Figure 4C , determine the edge walk probability of the third adjacent hyperedge 60 and the vertex walk probability of vertices 61, vertices 62 and vertices 63 on the third adjacent hyperedge 60.

[0164] As an example, see Figure 4C , determine the edge walk probability of the third adjacent hyperedge 85 and the vertex walk probability of vertices 64 and 65 on the third adjacent hyperedge 85.

[0165] In some embodiments, determining the edge walking probability of each third adjacent hyperedge in the above step 1057 can be achieved in the following manner: respectively obtaining the first vertex weight of the starting walking vertex corresponding to each third adjacent hyperedge, and determining the number of vertices connected to each third adjacent hyperedge; based on the first vertex weight and the number of vertices connected to the third adjacent hyperedge, determining the edge walking probability of each third adjacent hyperedge.

[0166] In step 1058, when the leaf vertices of each third adjacent hyperedge are leaf vertices of the hypergraph structure, based on the edge walk probability of the first adjacent hyperedge, the vertex walk probability of each vertex on the first adjacent hyperedge, the edge walk probability of the third adjacent hyperedge, and the vertex walk probability of each vertex on the third adjacent hyperedge, the probability that there is a transaction relationship between the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walking vertex during the walking of the starting walking vertex is determined.

[0167] As an example, see Figure 4C , when the leaf vertex 62 and the leaf vertex 63 of the third adjacent hyperedge 60 are both leaf vertices of the hypergraph structure, and the leaf vertex 64 and the leaf vertex 65 of the third adjacent hyperedge 85 are both leaf vertices of the hypergraph structure, based on the edge walk probability of the first adjacent hyperedges 81 to 84, the vertex walk probability of each vertex on the first adjacent hyperedges 81 to 84, the edge walk probability of the third adjacent hyperedges 60 and 85, and the vertex walk probability of each vertex (vertex 62, vertex 63, vertex 64 and vertex 65) on the third adjacent hyperedges 60 and 85, determine the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship during the walk of the starting walk vertex in the hypergraph structure.

[0168] In some embodiments, the above step 1058 can be implemented in the following manner: for each vertex on each first adjacent hyperedge, the following processing is performed respectively: the edge walk probability of the first adjacent hyperedge and the vertex walk probability of the vertex are multiplied to obtain the probability transfer value of the vertex. For each vertex on each third adjacent hyperedge, the following processing is performed respectively: the edge walk probability of the third adjacent hyperedge and the vertex walk probability of the vertex are multiplied to obtain the probability transfer value of the vertex. According to the probability transfer value of each vertex, a probability transfer matrix is ​​constructed. Based on the probability transfer matrix, the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship during the walk of the starting walk vertex of the hypergraph structure is determined.

[0169] As an example, see Figure 4C , the expression of the probability transition value of vertex 62 can be:

[0170] P=P 4 P 5 (12)

[0171] Among them, P 4 Characterizes the edge walk probability of the third adjacent hyperedge 60, P 5 represents the vertex wandering probability of vertex 62, and P represents the probability transition value of vertex 62.

[0172] In this way, by determining the first adjacent hyperedge connected to the starting wandering vertex, and based on the first vertex weight of each first adjacent hyperedge corresponding to the starting wandering vertex and the number of vertices connected by the first adjacent hyperedge, the edge wandering probability of the first adjacent hyperedge is determined, and the vertex wandering probability of each vertex on each first adjacent hyperedge is determined, so as to determine the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship in the process of wandering of the starting wandering vertex based on the edge wandering probability and the vertex wandering probability. Since the first vertex weight of the first adjacent hyperedge corresponding to the starting wandering vertex and the number of vertices connected by the first adjacent hyperedge are comprehensively considered when determining the edge wandering probability of the first adjacent hyperedge, and since the first vertex weight and the number of vertices connected by the first adjacent hyperedge can measure the importance between objects, the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship in the process of wandering of the starting wandering vertex can comprehensively extract the high-order information of the hypergraph structure, thereby effectively improving the fine-grainedness of the transaction relationship between objects.

[0173] In the application scenario of Internet link analysis, by determining the stationary probability distribution of the hypergraph structure, the popularity of Internet links can be accurately determined. When determining the edge walk probability of the first adjacent hyperedge, the first vertex weight of the first adjacent hyperedge corresponding to the starting walk vertex (the first visited Internet link) and the number of vertices connected by the first adjacent hyperedge are comprehensively considered. Since the first vertex weight and the number of vertices connected by the first adjacent hyperedge can measure the importance of each Internet link, the determined stationary probability distribution can comprehensively extract the high-order information of the Internet links, thereby effectively improving the granularity of the stationary probability distribution and more accurately reflecting the popularity of Internet links.

[0174] In the application scenario of influence ranking, the academic influence of the author can be accurately determined by determining the stationary probability distribution of the hypergraph structure. When determining the edge walk probability of the first adjacent hyperedge, the first vertex weight of the first adjacent hyperedge corresponding to the starting walk vertex (one of the authors of the paper) and the number of vertices connected by the first adjacent hyperedge are comprehensively considered. Since the first vertex weight and the number of vertices connected by the first adjacent hyperedge can measure the academic influence of the paper author, the determined stationary probability distribution can comprehensively extract the high-order information of the paper author, thereby effectively improving the granularity of the stationary probability distribution and more accurately reflecting the academic influence of the paper author.

[0175] The following is an explanation of an exemplary application of the embodiments of the present application in an actual application scenario of a hypergraph structure.

[0176] In the application scenario of influence ranking, each paper represents a hyperedge. One paper corresponds to multiple authors, and each author is a vertex on the hyperedge. By determining the stationary distribution of the hypergraph structure and sorting the probability values ​​in the distribution, the author influence ranking can be obtained. For a vertex with a larger probability value, the academic influence of the author corresponding to the vertex is greater.

[0177] In the application scenario of Internet link analysis, each browser represents a hyperedge. Each browser often clicks on multiple Internet links when browsing web pages. Each Internet link corresponds to a vertex of the hypergraph structure. By determining the stationary distribution of the hypergraph structure and sorting the probability values ​​in the distribution, the popularity ranking of Internet links can be obtained. For a vertex with a larger probability value, the popularity of the Internet link corresponding to the vertex is greater.

[0178] In some embodiments, the random walk of the hypergraph structure can be used for representation learning of the hypergraph, and then for a wide range of downstream tasks. For example, using the Restart Random Walk, the stationary probability distribution of the Restart Random Walk can be used to rank the author importance of the citation network, or to merge different rankings. The stationary probability distribution determined by the random walk of the hypergraph structure can also be used to construct a hypergraph neural network.

[0179] The method for processing the hypergraph structure provided in the embodiment of the present application provides a theoretical framework for random walks in a hypergraph. Since the edges of the hypergraph structure can connect multiple vertices, the hypergraph structure from a starting walk vertex to a terminating walk vertex can be divided into two steps: 1. From the starting walk vertex to the hyperedge; 2. From the hyperedge to the terminating walk vertex. In the above two-step walk process, the first vertex weight of the starting walk vertex corresponding to the adjacent hyperedge and the second vertex weight of the vertex corresponding to the first adjacent hyperedge are taken into account, and the settings of the first vertex weight and the second vertex weight are not the same, and the degree of the hyperedge is also taken into account. Therefore, the determined stationary probability distribution can comprehensively extract the high-order information of the hypergraph structure, thereby effectively improving the fine-grainedness of the stationary probability distribution.

[0180] In some embodiments, by effectively integrating the hypergraph information, a unified two-step random walk framework is obtained, which is described in detail below.

[0181] First, see Figure 4B , the first step of the random walk (see Figure 4B Label 1), select edge e from the adjacent hyperedges of the starting vertex u, and the edge walking probability of edge e is:

[0182]

[0183] Among them, p1 represents the edge walking probability of edge e; w(e) represents the first vertex weight of edge e; δ(e) represents the number of vertices connected by edge e, that is, the degree of edge e; ρ(.) represents the control function; Q 1 (u,e) represents the vertex weight matrix of edge e corresponding to the starting vertex u; d(u) represents the weight of the starting vertex u.

[0184] Then, see Figure 4B , the second step of the random walk (see Figure 4B Label 2), take the connected vertex v from edge e (see Figure 4B , vertex 62 and vertex 61). The vertex walk probability of vertex v is:

[0185]

[0186] Among them, p 2 Characterizes the vertex wandering probability of vertex v; Q 2 (v,e) represents the vertex weight of vertex v on edge e corresponding to edge e; δ(e) represents the number of vertices connected by edge e, that is, the degree of edge e.

[0187] In some embodiments, the expression of the probability transfer matrix between vertices of the hypergraph structure can be:

[0188]

[0189] Among them, P(u,v) represents the probability transfer matrix, p 1 Characterizes the edge walking probability of edge e; p 2 Characterizes the vertex wandering probability of vertex v; Q 2 (v,e) represents the vertex weight of vertex v on edge e corresponding to edge e; δ(e) represents the number of vertices connected by edge e, that is, the degree of edge e; ρ(.) represents the control function; Q 1 (u,e) represents the vertex weight matrix of edge e corresponding to the starting vertex u; d(u) represents the weight of the starting vertex u.

[0190] In some embodiments, under the framework of the above two-step random walk, only Q 1 (u,e) and Q 2 (v,e) satisfies certain conditions, and an explicit expression of the stationary probability distribution can be obtained. The specific conditions include: (Condition 1) Q 1 (u,e)=Q 2 (v,e); (Condition 2)Q 1 (u,e) and Q 2 (v,e) is the vertex weight independent of the edge, i.e., Q 1 (u,e)=q 1 (e),Q2 (u,e)=q 2 (e).

[0191] Assume that the stationary probability distributions corresponding to the above conditions 1 and 2 are π 1 and π 2 , whose explicit expressions are:

[0192]

[0193]

[0194] Among them, π 1 (v) represents the stationary probability distribution corresponding to condition 1, π 2 (v) Characterizes the stationary probability distribution corresponding to condition 2.

[0195] That is, the above two conditions have the same expression, so the above formula (17) can be used as a unified expression of the above conditions 1 and 2. Among them, π 1 (v) and π 2 The correctness of (v) can be verified by the verification formula:

[0196] π i (v)=∑ u π i (u)P(u,v) (18)

[0197] Among them, π i (v) represents the i-th stationary probability distribution, π i-1 (u) represents the i-1th stationary probability distribution, and P(u,v) represents the probability transfer matrix.

[0198] In this way, the hypergraph structure processing method provided in the embodiment of the present application can obtain the hypergraph structure random walk solution in the related technology by setting the above-mentioned specific conditions, and has strong compatibility and versatility.

[0199] It is understandable that in the embodiments of the present application, related data such as hypergraph structures are involved. When the embodiments of the present application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0200] The following is a description of an exemplary structure of a hypergraph structure processing device 255 provided in an embodiment of the present application implemented as a software module. In some embodiments, Figure 2As shown, the software modules in the processing device 255 of the hypergraph structure stored in the memory 240 may include: a first determination module 2551, which is used to determine the starting wandering vertex and at least one first adjacent hyperedge connected to the starting wandering vertex in a hypergraph structure composed of multiple vertices and multiple hyperedges for connecting the vertices; wherein the vertex is used to represent the object, and the hyperedge connecting the vertex is used to represent the associated object having a transaction relationship with the object; a second determination module 2552, which is used to respectively obtain the first vertex weight of each first adjacent hyperedge corresponding to the starting wandering vertex, and determine the number of vertices connected to each first adjacent hyperedge; a third determination module 2553, which is used to determine the edge wandering probability of each first adjacent hyperedge based on the first vertex weight and the number of vertices connected to the first adjacent hyperedge; a fourth determination module 2554, which is used to determine the vertex wandering probability of each vertex on the first adjacent hyperedge, and based on the edge wandering probability and the vertex wandering probability, determine the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship during the wandering of the starting wandering vertex.

[0201] In some embodiments, the third determination module 2553 is further used to perform the following processing for each first adjacent hyperedge: determine the walking weight of the first adjacent hyperedge based on the first vertex weight and the number of vertices connected to the first adjacent hyperedge; sum the walking weights of each first adjacent hyperedge to obtain the edge walking weight of the starting walking vertex; determine the ratio of the walking weight to the edge walking weight as the edge walking probability of the first adjacent hyperedge.

[0202] In some embodiments, the third determination module 2553 is also used to obtain the vertex weight matrix of the first adjacent hyperedge; multiply the first vertex weight and the number of vertices connected by the first adjacent hyperedge to obtain a multiplication result; and determine the walking weight of the first adjacent hyperedge based on the multiplication result and the vertex weight matrix.

[0203] In some embodiments, the third determination module 2553 is also used to call a control function to process the number of vertices connected by the first adjacent hyperedge to obtain a control processing result; wherein the control function is used to control the influence of the first vertex weight on the walking weight; the multiplication processing result, the vertex weight matrix and the control processing result are multiplied to obtain the walking weight of the first adjacent hyperedge.

[0204] In some embodiments, the fourth determination module 2554 is also used to perform the following processing for each first adjacent hyperedge: obtain the second vertex weight of the first adjacent hyperedge corresponding to each vertex on the first adjacent hyperedge; based on the second vertex weight, determine the vertex wandering probability of each vertex on the first adjacent hyperedge.

[0205] In some embodiments, the above-mentioned fourth determination module 2554 is further configured to sum the second vertex weights corresponding to each vertex on the first adjacent hyperedge respectively to obtain a weight sum; for each vertex on the first adjacent hyperedge, perform the following processing respectively: determine the ratio of the second vertex weight corresponding to the vertex on the first adjacent hyperedge to the weight sum as the vertex walk probability of the vertex.

[0206] In some embodiments, the above-mentioned fourth determination module 2554 is further configured to perform the following processing for each vertex on each first adjacent hyperedge respectively: multiply the edge walk probability of the first adjacent hyperedge and the vertex walk probability of the vertex to obtain a probability transition value of the vertex; construct a probability transition matrix according to the probability transition value of each vertex; determine a stationary probability distribution of the hypergraph structure based on the probability transition matrix; determine the probability corresponding in the stationary probability distribution of the hypergraph structure as the probability that there is a transaction relationship between the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex during the process of walking at the starting walk vertex.

[0207] In some embodiments, the above-mentioned fourth determination module 2554 is further configured to determine the probability distribution of the first iteration of the hypergraph structure as the probability transition matrix, and determine the probability distribution of the second iteration of the hypergraph structure based on the probability transition value of each vertex in the probability distribution of the first iteration; determine the probability distribution of the (i + 1)-th iteration of the hypergraph structure based on the probability transition value of each vertex in the probability distribution of the i-th iteration of the hypergraph structure, where 2 ≤ i < N; wherein, the difference between the probability transition value of any vertex in the probability distribution of the (N - 1)-th iteration and the probability transition value of the corresponding vertex in the probability distribution of the N-th iteration is less than a stationary threshold; determine the probability distribution of the N-th iteration as the stationary probability distribution of the hypergraph structure.

[0208] In some embodiments, the above-mentioned fourth determination module 2554 is further configured to determine the association relationship between the leaf vertex of each first adjacent hyperedge and the hypergraph structure; when the association relationship indicates that the leaf vertex of each first adjacent hyperedge is a leaf vertex of the hypergraph structure, based on the edge walk probability and the vertex walk probability, determine the probability corresponding in the stationary probability distribution of the hypergraph structure as the probability that there is a transaction relationship between the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex during the process of walking at the starting walk vertex.

[0209] In some embodiments, the fourth determination module 2554 is further used to, when there is a second adjacent hyperedge in at least one first adjacent hyperedge, wherein the leaf vertex of the second adjacent hyperedge is not a leaf vertex of the hypergraph structure, perform the following processing on the leaf vertex of the second adjacent hyperedge: determine at least one third adjacent hyperedge connected to the leaf vertex of the second adjacent hyperedge; determine the edge walk probability of each third adjacent hyperedge, and the vertex walk probability of each vertex on the third adjacent hyperedge; when the leaf vertex of each third adjacent hyperedge is a leaf vertex of the hypergraph structure, determine the corresponding probability in the stationary probability distribution of the hypergraph structure as the probability that, during the walking of the starting walking vertex, the objects corresponding to each vertex in the hypergraph structure and the associated objects corresponding to the hyperedge associated with the starting walking vertex have a transaction relationship based on the edge walk probability of the first adjacent hyperedge, the vertex walk probability of each vertex on the first adjacent hyperedge, the edge walk probability of the third adjacent hyperedge, and the vertex walk probability of each vertex on the third adjacent hyperedge.

[0210] In some embodiments, the fourth determination module 2554 is further used to respectively obtain the first vertex weight of the starting walking vertex corresponding to each third adjacent hyperedge, and determine the number of vertices connected by each third adjacent hyperedge; based on the first vertex weight and the number of vertices connected by the third adjacent hyperedge, determine the edge walking probability of each third adjacent hyperedge.

[0211] The embodiment of the present application provides a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. The processor of the computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the above-mentioned hypergraph structure processing method of the embodiment of the present application.

[0212] The present application embodiment provides a computer-readable storage medium storing executable instructions, wherein the executable instructions are stored. When the executable instructions are executed by a processor, the processor will execute the hypergraph structure processing method provided by the present application embodiment, for example, Figure 3A The method for processing the hypergraph structure is shown.

[0213] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface storage, optical disk, or CD-ROM; or it may be various devices including one or any combination of the above memories.

[0214] In some embodiments, executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0215] As an example, executable instructions may, but need not, correspond to a file in a file system, may be stored as part of a file that stores other programs or data, such as in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions).

[0216] By way of example, executable instructions may be deployed to be executed on one computing device, or on multiple computing devices located at one site, or on multiple computing devices distributed across multiple sites and interconnected by a communication network.

[0217] In summary, the embodiments of the present application have the following beneficial effects:

[0218] (1) By determining the first adjacent hyperedge connected to the starting wandering vertex, and based on the first vertex weight of each first adjacent hyperedge corresponding to the starting wandering vertex and the number of vertices connected by the first adjacent hyperedge, the edge walking probability of the first adjacent hyperedge is determined, and the vertex walking probability of each vertex on each first adjacent hyperedge is determined, thereby determining the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship during the wandering process of the starting wandering vertex based on the edge walking probability and the vertex walking probability. Since the first vertex weight of the first adjacent hyperedge corresponding to the starting wandering vertex and the number of vertices connected by the first adjacent hyperedge are comprehensively considered when determining the edge walking probability of the first adjacent hyperedge, and since the first vertex weight and the number of vertices connected by the first adjacent hyperedge can measure the importance between objects, the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship during the wandering process of the starting wandering vertex can comprehensively extract the high-order information of the hypergraph structure, thereby effectively improving the fine-grainedness of the extracted transaction relationship between objects.

[0219] (2) By obtaining the first vertex weight of each first adjacent hyperedge corresponding to the starting walk vertex and determining the number of vertices connected by each first adjacent hyperedge, it is convenient to subsequently determine the edge walk probability based on the first vertex weight and the number of vertices connected by the first adjacent hyperedge, and then determine the stationary probability distribution of the hypergraph structure based on the edge walk probability, thereby effectively improving the accuracy of the stationary probability distribution.

[0220] (3) By determining the edge walk probability of the first adjacent hyperedge, and then determining the stationary probability distribution of the hypergraph structure according to the edge walk probability of the first adjacent hyperedge, the accuracy of the determined stationary probability distribution is ensured.

[0221] (4) The vertex walk probability of each vertex on each first adjacent hyperedge can be determined, and then the stationary probability distribution of the hypergraph structure can be accurately determined based on the vertex walk probability and the edge walk probability, thereby effectively improving the accuracy of the stationary probability distribution.

[0222] (5) Since the probability transfer values ​​of the corresponding vertices in the Nth iteration probability distribution have tended to be stable, the Nth iteration probability distribution can be determined as the stable probability distribution of the hypergraph structure, thereby accurately determining the stable probability distribution of the hypergraph structure through multiple iterations of the probability transfer matrix.

[0223] (6) When the association relationship represents that the leaf vertices of each first adjacent hyperedge are all leaf vertices of the hypergraph structure, the stationary probability distribution of the hypergraph structure can be determined directly based on the edge walk probability and the vertex walk probability. Thus, by determining the association relationship, the stationary probability distribution of the hypergraph structure can be completely and comprehensively determined.

[0224] (7) The vertex wandering probability of each vertex on the first adjacent hyperedge is determined by corresponding each vertex on the first adjacent hyperedge to the second vertex weight of the first adjacent hyperedge. Since the second vertex weight can measure the importance of the vertices on the first adjacent hyperedge, the vertex wandering probability is determined by the second vertex weight, so that the vertex wandering probability can more comprehensively reflect the high-order information of the hypergraph structure, and the stationary probability distribution determined by the vertex wandering probability has a higher granularity and can comprehensively reflect the high-order information of the hypergraph structure.

[0225] (8) In the application scenario of Internet link analysis, by determining the stationary probability distribution of the hypergraph structure, the popularity of Internet links can be accurately determined. When determining the edge walk probability of the first adjacent hyperedge, the first vertex weight of the first adjacent hyperedge corresponding to the starting walk vertex (the first visited Internet link) and the number of vertices connected by the first adjacent hyperedge are comprehensively considered. Since the first vertex weight and the number of vertices connected by the first adjacent hyperedge can measure the importance of each Internet link, the determined stationary probability distribution can comprehensively extract the high-order information of the Internet links, thereby effectively improving the granularity of the stationary probability distribution and more accurately reflecting the popularity of Internet links.

[0226] (9) In the application scenario of influence ranking, the academic influence of the author can be accurately determined by determining the stationary probability distribution of the hypergraph structure. When determining the edge walk probability of the first adjacent hyperedge, the first vertex weight of the first adjacent hyperedge corresponding to the starting walk vertex (one of the authors of the paper) and the number of vertices connected by the first adjacent hyperedge are taken into consideration. Since the first vertex weight and the number of vertices connected by the first adjacent hyperedge can measure the academic influence of the paper author, the determined stationary probability distribution can comprehensively extract the high-order information of the paper author, thereby effectively improving the granularity of the stationary probability distribution and more accurately reflecting the academic influence of the paper author.

[0227] The above is only an embodiment of the present application and is not intended to limit the protection scope of the present application. Any modifications, equivalent substitutions and improvements made within the spirit and scope of the present application are included in the protection scope of the present application.

Claims

1. A method for processing a hypergraph structure, characterized in that: The method comprises: In a hypergraph structure composed of a plurality of vertices and a plurality of hyperedges for connecting the vertices, determining a starting wandering vertex and at least one first adjacent hyperedge connected to the starting wandering vertex; The vertices are used to represent objects, the hyperedges connecting the vertices are used to represent the communication channels between papers or communication participants, and the hyperedges are used to connect associated objects that have transactional relationships with the objects; Wherein, when the hyperedge represents the paper, the object includes the paper author, and the transaction relationship represents that the associated object is the associated author who published the paper with the paper author; when the hyperedge represents the communication channel between the communication participants, the object includes the communication participant, and the transaction relationship represents that the associated object is the associated communication participant that establishes a communication connection with the communication participant; Respectively obtain the first vertex weight of each of the first adjacent hyperedges corresponding to the starting wandering vertex, and determine the number of vertices connected by each of the first adjacent hyperedges; Determining an edge walk probability of each of the first adjacent hyperedges based on the first vertex weight and the number of vertices connected by the first adjacent hyperedges; Determine the vertex walk probability of each vertex on the first adjacent hyperedge, and Based on the edge walking probability and the vertex walking probability, the probability that the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walking vertex have a transaction relationship during the walking of the starting walking vertex is determined.

2. The method according to claim 1, characterized in that: The step of determining the edge walking probability of each of the first adjacent hyperedges based on the first vertex weight and the number of vertices connected by the first adjacent hyperedges includes: The following processing is performed for each of the first adjacent hyperedges: Determining a walk weight of the first adjacent hyperedge based on the first vertex weight and the number of vertices connected by the first adjacent hyperedge; Sum the walking weights of each of the first adjacent hyperedges to obtain the walking weight of the starting walking vertex; A ratio of the walk weight to the edge walk weight is determined as the edge walk probability of the first adjacent hyperedge.

3. The method according to claim 2, characterized in that The determining the walking weight of the first adjacent hyperedge based on the first vertex weight and the number of vertices connected by the first adjacent hyperedge includes: Obtaining a vertex weight matrix of the first adjacent hyperedge; Multiplying the first vertex weight and the number of vertices connected by the first adjacent hyperedge to obtain a multiplication result; Based on the multiplication result and the vertex weight matrix, a walking weight of the first adjacent hyperedge is determined.

4. The method according to claim 3, characterized in that The determining the walking weight of the first adjacent hyperedge based on the multiplication result and the vertex weight matrix includes: Calling a control function to process the number of vertices connected by the first adjacent hyperedge to obtain a control processing result; Wherein, the control function is used to control the influence of the first vertex weight on the walk weight; The multiplication result, the vertex weight matrix and the control processing result are multiplied to obtain the walk weight of the first adjacent hyperedge.

5. The method according to claim 1, characterized in that The step of determining the vertex walk probability of each vertex on the first adjacent hyperedge comprises: The following processing is performed for each of the first adjacent hyperedges: Obtain a second vertex weight of the first adjacent hyperedge corresponding to each vertex on the first adjacent hyperedge; Based on the second vertex weight, a vertex walk probability of each of the vertices on the first adjacent hyperedge is determined.

6. The method according to claim 5, characterized in that The step of determining the vertex walk probability of each vertex on the first adjacent hyperedge based on the second vertex weight includes: Sum the weights of the second vertices of the first adjacent hyperedge corresponding to each vertex on the first adjacent hyperedge to obtain a weight sum; The following processing is performed for each of the vertices on the first adjacent hyperedge: a ratio of the second vertex weight corresponding to the first adjacent hyperedge of the vertex to the sum of the weights is determined as the vertex wandering probability of the vertex.

7. The method according to claim 1, characterized in that The determining, based on the edge walk probability and the vertex walk probability, the probability that, during the process of the starting walk vertex walking, the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship includes: For each of the vertices on each of the first adjacent hyperedges, the following processing is performed respectively: the edge walk probability of the first adjacent hyperedge and the vertex walk probability of the vertex are multiplied to obtain a probability transfer value of the vertex; Constructing a probability transfer matrix according to the probability transfer value of each vertex; Based on the probability transfer matrix, determining a stationary probability distribution of the hypergraph structure; The corresponding probability in the stationary probability distribution of the hypergraph structure is determined as the probability that, during the process of the starting wandering vertex wandering, the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting wandering vertex have a transaction relationship.

8. The method according to claim 7, characterized in that The step of determining the stationary probability distribution of the hypergraph structure based on the probability transfer matrix includes: Determine the probability transfer matrix as the first iteration probability distribution of the hypergraph structure, and determine the second iteration probability distribution of the hypergraph structure based on the probability transfer value of each vertex in the first iteration probability distribution; Based on the probability transfer value of each vertex in the i-th iteration probability distribution of the hypergraph structure, determine the i+1-th iteration probability distribution of the hypergraph structure, where 2≤i <N; Wherein, the difference between the probability transition value of any vertex in the N-1th iteration probability distribution and the probability transition value of the corresponding vertex in the Nth iteration probability distribution is less than the stable threshold; The Nth iteration probability distribution is determined as the stationary probability distribution of the hypergraph structure.

9. The method according to claim 1, characterized in that: The determining, based on the edge walk probability and the vertex walk probability, the probability that, during the process of the starting walk vertex walking, the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship includes: Determine the association relationship between the leaf vertex of each of the first adjacent hyperedges and the hypergraph structure; When the association relationship represents that the leaf vertices of each of the first adjacent hyperedges are leaf vertices of the hypergraph structure, based on the edge walk probability and the vertex walk probability, the probability that a transaction relationship exists between the objects corresponding to each vertex in the hypergraph structure and the associated objects corresponding to the hyperedge associated with the starting walk vertex during the walk of the starting walk vertex is determined.

10. The method according to claim 1, characterized in that The determining, based on the edge walk probability and the vertex walk probability, the probability that, during the process of the starting walk vertex walking, the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship includes: When there is a second adjacent hyperedge in the at least one first adjacent hyperedge, wherein the leaf vertex of the second adjacent hyperedge is not a leaf vertex of the hypergraph structure, the following processing is performed on the leaf vertex of the second adjacent hyperedge: determining at least one third adjacent hyperedge connected to the leaf vertex of the second adjacent hyperedge; Determine an edge walk probability of each of the third adjacent hyperedges and a vertex walk probability of each vertex on the third adjacent hyperedge; When each leaf vertex of the third adjacent hyperedge is a leaf vertex of the hypergraph structure, based on the edge walk probability of the first adjacent hyperedge, the vertex walk probability of each vertex on the first adjacent hyperedge, the edge walk probability of the third adjacent hyperedge, and the vertex walk probability of each vertex on the third adjacent hyperedge, the probability that a transaction relationship exists between the object corresponding to each vertex in the hypergraph structure and the associated object corresponding to the hyperedge associated with the starting walking vertex during the walking of the starting walking vertex is determined.

11. The method according to claim 10, characterized in that The step of determining the edge walk probability of each of the third adjacent hyperedges comprises: Respectively obtain the first vertex weight of each of the third adjacent hyperedges corresponding to the starting wandering vertex, and determine the number of vertices connected by each of the third adjacent hyperedges; Based on the first vertex weight and the number of vertices connected by the third adjacent hyperedges, an edge walking probability of each of the third adjacent hyperedges is determined.

12. A hypergraph structure processing device, characterized in that: The device comprises: A first determination module is used to determine a starting wandering vertex and at least one first adjacent hyperedge connected to the starting wandering vertex in a hypergraph structure composed of multiple vertices and multiple hyperedges for connecting the vertices; wherein the vertex is used to represent an object, the hyperedge connecting the vertex is used to represent a communication channel between papers or communication participants, and the hyperedge is used to connect an associated object that has a transactional relationship with the object; wherein, when the hyperedge represents the paper, the object includes the paper author, and the transactional relationship represents that the associated object is an associated author who published the paper with the paper author; when the hyperedge represents the communication channel between the communication participants, the object includes the communication participant, and the transactional relationship represents that the associated object is an associated communication participant that establishes a communication connection with the communication participant; A second determination module is used to respectively obtain the first vertex weight of each of the first adjacent hyperedges corresponding to the starting wandering vertex, and determine the number of vertices connected by each of the first adjacent hyperedges; A third determination module, configured to determine an edge walking probability of each of the first adjacent hyperedges based on the first vertex weight and the number of vertices connected by the first adjacent hyperedges; The fourth determination module is used to determine the vertex walk probability of each vertex on the first adjacent hyperedge, and based on the edge walk probability and the vertex walk probability, determine the probability that during the walk of the starting walk vertex, the objects corresponding to each vertex in the hypergraph structure and the associated objects corresponding to the hyperedge associated with the starting walk vertex have a transaction relationship.

13. An electronic device, characterized in that: The electronic device comprises: A memory for storing executable instructions; A processor, configured to implement the method for processing a hypergraph structure according to any one of claims 1 to 11 when executing executable instructions or computer programs stored in the memory.

14. A computer-readable storage medium storing executable instructions or a computer program, characterized in that: When the executable instructions are executed by a processor, the method for processing the hypergraph structure described in any one of claims 1 to 11 is implemented.

15. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method for processing a hypergraph structure according to any one of claims 1 to 11 is implemented.