CN113515568A - Graph relation network construction method, graph neural network model training method and device - Google Patents
Graph relation network construction method, graph neural network model training method and device Download PDFInfo
- Publication number
- CN113515568A CN113515568A CN202110792127.1A CN202110792127A CN113515568A CN 113515568 A CN113515568 A CN 113515568A CN 202110792127 A CN202110792127 A CN 202110792127A CN 113515568 A CN113515568 A CN 113515568A
- Authority
- CN
- China
- Prior art keywords
- event
- user
- graph
- relationship
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/26—Visual data mining; Browsing structured data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The disclosure provides a graph relation network construction method, a graph neural network model training method and a graph neural network model training device, relates to the field of big data processing, and particularly relates to the technical field of user understanding, relation understanding, graph relation network construction and graph neural network model training. The specific implementation scheme is as follows: the time sequence information among the user events is fused in the constructed graph relation network, so that the constructed graph network relation containing the time sequence information can provide better user representation and user event node representation, and a good data basis is provided for subsequent graph relation network data mining based on a graph neural network and the like.
Description
Technical Field
The disclosure relates to the technical field of big data processing, in particular to the technical fields of user understanding, relationship understanding, graph relationship network construction and graph neural network model training.
Background
The relationship understanding of the user is characterized by important dimensions of the user, and how to enable the constructed graph network relationship to contain more information by taking graph data as important data types for representing the user is one of the most important design links for graph relationship network mining.
Disclosure of Invention
The disclosure provides a graph relation network construction method, a graph neural network model training method and a graph neural network model training device.
According to a first aspect of the present disclosure, there is provided a graph relationship network construction method including timing information, including:
determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between the user and a user event;
determining an event pair relationship existing among user events of a target user, wherein the event pair relationship represents that the user events occur successively;
and adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information.
According to a second aspect of the present disclosure, there is provided a graph neural network model training method, including:
training a neural network model based on a plurality of training samples; the training sample has the graph relationship network of any one of the first aspect containing timing information.
According to a third aspect of the present disclosure, there is provided a graph relation network constructing apparatus including timing information, including:
the first determination module is used for determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between the user and a user event;
the second determining module is used for determining an event pair relationship existing among user events of the target user, wherein the event pair relationship represents that the user events occur successively;
and the first adding module is used for adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information.
According to a fourth aspect of the present disclosure, there is provided a neural network model training apparatus, including:
the training module is used for training the neural network model of the graph based on a plurality of training samples; the training sample has the graph relationship network of any one of the first aspect containing timing information.
According to a fifth aspect of the present disclosure, there is provided an electronic apparatus including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method.
According to a sixth aspect of the present disclosure, there is provided a non-transitory computer readable storage medium having stored thereon computer instructions for causing a computer to perform the above method.
According to a seventh aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the above method.
The technical scheme provided by the disclosure has the following beneficial effects: the graph network relationship constructed by the method can provide better user representation and user event node representation, and provides a good data basis for subsequent graph relationship network data mining based on a graph neural network and the like.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present disclosure, nor do they limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.
Drawings
The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:
fig. 1 is a schematic flow chart diagram of a graph relationship network construction method including timing information according to the present disclosure;
FIG. 2 is an exemplary diagram of a prior art graph relationship network provided in accordance with the present disclosure;
FIG. 3 is an exemplary diagram of a graph relationship network including timing information provided in accordance with the present disclosure;
FIG. 4 is an exemplary diagram of a graph relationship network including user relationships and user event relationships provided in accordance with the present disclosure;
FIG. 5 is a schematic flow diagram of a neural network model training method provided in accordance with the present disclosure;
fig. 6 is a schematic structural diagram of a graph relationship network construction device including timing information according to the present disclosure;
FIG. 7 is a schematic diagram of a schematic diagram neural network model training device provided by the present disclosure;
FIG. 8 is a block diagram of an electronic device used to implement an embodiment of the disclosure.
Detailed Description
Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Example one
Fig. 1 illustrates a graph relationship network construction method including timing information according to an embodiment of the present disclosure, and as shown in fig. 1, the method includes:
step S101, determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between a user and a user event;
a Graph (Graph) is an object having a broad meaning, and in the field of computers, a Graph is a common data structure, and in data science, a Graph is used to broadly describe various relational data. Generally, graphs are used to represent objects and relationships between objects, and in scenarios such as social networks, biopharmaceuticals, internet of things, etc., there is a large amount of graph data. The graph is composed of a Vertex (Vertex) indicating an object to be studied and an Edge (Edge) connecting the vertices, and the Edge indicates a specific relationship between the two objects. For the present application, graph network relationships may represent a collection of vertices and edges, where both user and user events may be vertices of a graph.
Step S102, determining an event pair relationship existing among user events of a target user, wherein the event pair relationship represents that the user events occur successively;
specifically, the target user may correspond to a plurality of user events, and the plurality of user events have a certain incidence relation, and if the target user performs the event a and then performs the event B, it may be considered that the user event a and the user event B have an event pair relation.
Step S103, adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information.
Specifically, the successive occurrence of the user events includes certain timing information, and the determined event pair relationship existing between the user events of the target user is added to the first graph relationship network, that is, the connection of the user events is established, and the variable relationship between the user events is established, so that the graph relationship network including the timing information can be established.
To better understand the invention of the present application, for example, fig. 2 shows an existing constructed graph relationship network of users, where a target user (UserA) performs a user event a (query1) and a user event B (query2), and the existing constructed graph relationship network of users only establishes an edge relationship between the target user (UserA) and the user event (query1, query2), but does not construct an edge relationship between the user events, so that the existing constructed graph relationship network of users loses time sequence information contained in the user event. According to the method, whether the user event has a user event pair relationship is determined, if the user event has the event pair relationship, namely the user event A (query1) and the user event B (query2) occur successively, it is described that certain time sequence information exists between the user event A (query1) and the user event B (query2), so that an edge relationship between the user event A (query1) and the user event B (query2) is constructed, and the time sequence information contained between the user event A (query1) and the user event B (query2) is added into the constructed graph network relationship.
According to the scheme provided by the embodiment of the disclosure, the graph network relation constructed in the prior art only contains the relation between the user and the user event, and does not contain the time sequence information between the user events. The method comprises the steps of determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between a user and a user event; determining an event pair relationship existing among user events of a target user, wherein the event pair relationship represents that the user events occur successively; and adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information. In other words, the time sequence information among the user events is fused in the constructed second graph relationship network, so that the constructed graph network relationship containing the time sequence information can provide better user representation and user event node representation, and a good data basis is provided for subsequent graph relationship network data mining based on a graph neural network and the like.
The disclosed embodiments provide a possible implementation manner, wherein the method according to claim 1, further includes:
and determining the event pair relationship existing between the user events of the target user based on the event log of the target user.
The user event log may also be referred to as a user behavior log, a user behavior trace, a traffic log, and the like. In short, it is the behavior data (access, browse, search, click, etc.) that is generated each time the user visits a website, browses an article, watches a video, etc. Specifically, the embedded point can be performed in a developed application or a webpage, so that the corresponding operation event of the user can be monitored.
Specifically, the user events having event pair relationships may be determined from the event logs of the target users.
With the embodiment of the application, the problem of how to determine whether the event pair relationship exists in the user event is solved.
The embodiment of the present disclosure provides a possible implementation manner, where determining an event pair relationship existing between user events of a target user based on an event log of the target user includes:
determining each user event and the occurrence sequence of the user events of the target user based on the event log of the target user;
and determining the event pair relationship existing between the user events of the target user based on the determined user events of the target user and the occurrence sequence of the user events.
Specifically, the user event log includes event information of the user and occurrence time information, and based on the occurrence time information of the user events, the occurrence order of the user events can be determined, and further, whether an event pair relationship exists between the user events is determined.
Specifically, a certain judgment condition may also be set, such as a time judgment condition, and if the event a and the event B occur successively within a predetermined time threshold range, it is considered that the event pair relationship exists between the user event a and the user event B, and if the event B occurs after the event a occurs and exceeds the predetermined time threshold, it is considered that the event pair relationship does not exist between the user event a and the user event B. The time threshold may be set according to the event type, for example, a long video event such as watching a movie or a video, a long time threshold may be set, and a short threshold time may be set for watching a short video.
The method and the device solve the problem of determining whether the event pair relationship exists in the user event according to the user event log.
The embodiment of the present disclosure provides a possible implementation manner, wherein the method further includes:
determining the occurrence frequency of each event pair relation based on the event log of the target user;
determining the weight of each event pair relation based on the occurrence frequency of the determined event pair relation;
and adding the determined weight of each event pair relationship to the second graph relationship network to obtain a third graph relationship network.
The graph relation network can be divided into a non-weighted graph and a weighted graph, and if each edge in the graph has a real number corresponding to the edge, the graph is a weighted graph. The real numbers are referred to as weights on corresponding edges, and for example, as shown in fig. 3, w1, w2, w3, and w4 are weights of events.
Specifically, the occurrence frequency of each event pair relationship may be determined based on an event log of the target user, for example, the target user performs the operation of event a- > event B for multiple times within a certain threshold time, the occurrence frequency of event a- > event B may be statistically determined, the weight of each event pair relationship is determined according to the occurrence frequency of the determined event pair relationship, and the weight information is fused to the second graph relationship network to obtain the third graph relationship network.
Specifically, if the number of times of the event pair relationship is smaller than a certain threshold, it indicates that the event pair relationship occurs accidentally, and the probability of the event pair relationship occurring again is low, and the edge relationship of the user event corresponding to the event pair relationship whose number of times is smaller than the certain threshold may not be established.
For the example of the application, the importance of the relevant user event with the time sequence information is reflected by the occurrence frequency of the event pair relationship, so that the constructed graph network relationship fusing the time sequence information and the weight information between the events can further better reflect the representation of the user or the representation of the event node.
The embodiment of the present disclosure provides a possible implementation manner, where determining the occurrence frequency of each event pair relationship based on an event log of a target user includes:
determining a user event of the target user based on the event log of the target user;
carrying out semantic normalization processing on the user event of the target user to obtain the user event after the semantic normalization processing;
and determining the occurrence frequency of each event pair relation based on the user events after semantic normalization.
Specifically, the same user event may have different expressions, such as, for example, a qurey event, the following query events: 1. when Hua raving is born, 2. Liu De Hua year and month of birth, 3. when Liu De Hua is born, 4. Liu De Hua date of birth, the expression semantics are consistent. The synonymy expressed query can be normalized through query semantic normalization. For example, for a video viewing event, videos of the same type may be normalized, such as behaviors of viewing documentary a and documentary B are normalized to a class of events, and behaviors of viewing anarchic program a and anarchic program B are normalized to a class of events.
According to the embodiment of the application, after semantic normalization processing is carried out on the user events, the occurrence frequency of event pair relations is determined, and therefore the event pair relations determined through statistics are more accurate.
The embodiment of the present disclosure provides a possible implementation manner, wherein the method further includes:
and acquiring second graph relationship networks of a plurality of target users with association relations, and carrying out network fusion on the second graph relationship networks of the plurality of target users with association relations to obtain a fourth graph relationship network.
Exemplarily, as shown in fig. 4, if a certain relationship (such as attention, comment, like approval) exists between the UserA and the UserB, the graph relationship networks of the UserA and the UserB can be fused, so as to obtain a graph relationship network containing more information, that is, the graph relationship network contains not only the relationship between users and events, but also the relationship between users and events, and the user-event relationship, so as to further improve the capability of user representation.
The embodiment of the present disclosure provides a possible implementation manner, where a user event includes at least one of the following:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
For the present embodiment, the user event may include a watch video event; browsing article events; retrieving the event; at least one item in the advertisement click events is integrated, so that a plurality of user events are integrated, information contained in the graph relation network is further improved, and the capability of representing users is improved.
Example two
According to a second aspect of the present disclosure, there is provided a graph neural network model training method, as shown in fig. 5, including:
step S501, training a neural network model based on a plurality of training samples; the training sample has the graph relationship network including timing information of the first embodiment.
Wherein a graph is a data structure that models a set of objects (nodes) and their relationships (edges). Due to the powerful expressive power of graph structures, the research of analyzing graphs by a machine learning method is increasingly gaining attention. Graph Neural Networks (GNNs) are a class of deep learning-based methods for processing map domain information, and due to their better performance and interpretability, GNNs have recently become a widely used method for Graph analysis.
Among them, the graph neural network can be divided into five categories, which are: graph Convolution Networks (GCNs), Graph Attention Networks (Graph Attention Networks), Graph Autoencoders (Graph Autoencoders), Graph generation Networks (Graph generating Networks), and Graph spatio-temporal Networks (Graph Spatial-temporal Networks).
For the application, based on different application scenarios, a corresponding graph neural network model can be selected in a targeted manner, the graph neural network model is trained based on sample data of the graph relation network including the timing sequence information in the first embodiment, and then corresponding application is performed.
For the application, the applied sample data contains the time sequence information of the user event, so that richer information is contained, better user or event node representation is provided for downstream applications, and the downstream applications, such as node classification, personalized recommendation and the like, are more accurate.
The embodiment of the present application provides a possible implementation manner, where a user event of a graph relationship network including timing information includes at least one of the following:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
For the present embodiment, the user event may include a watch video event; browsing article events; retrieving the event; at least one item in the advertisement click events is integrated in the sample data, so that the information contained in the graph relation network is further improved, the capability of representing the user is improved, and a good data basis is provided for subsequent downstream application.
EXAMPLE III
The embodiment of the present disclosure provides a graph neural network model training apparatus 60, as shown in fig. 6, including:
a first determining module 601, configured to determine a first graph relationship network of a target user, where the first graph relationship network includes a relationship between a user and a user event;
a second determining module 602, configured to determine an event pair relationship existing between user events of the target user, where the event pair relationship indicates that the user events occur successively;
a first adding module 603, configured to add event pair relationships existing between the user events of the determined target user to the first graph relationship network, so as to obtain a second graph relationship network including the timing information.
The embodiment of the present application provides a possible implementation manner, wherein the apparatus further includes:
and the third determining module is used for determining the event pair relationship existing between the user events of the target user based on the event log of the target user.
The embodiment of the present application provides a possible implementation manner, where the third determining module includes:
the first determining unit is used for determining each user event and the occurrence sequence of the user events of the target user based on the event log of the target user;
and the second determining unit is used for determining the event pair relationship existing between the user events of the target user based on the determined user events of the target user and the occurrence sequence of the user events.
The embodiment of the present application provides a possible implementation manner, wherein the apparatus further includes:
the fourth determining module is used for determining the occurrence frequency of each event pair relation based on the event log of the target user;
the fifth determining module is used for determining the weight of each event pair relation based on the occurrence frequency of the determined event pair relation;
and the second adding module is specifically used for adding the determined weight of each event pair relationship to the second graph relationship network to obtain a third graph relationship network.
The embodiment of the present application provides a possible implementation manner, wherein the fourth determining module is specifically configured to determine a user event of a target user based on an event log of the target user; the semantic normalization processing is carried out on the user event of the target user to obtain the user event after the semantic normalization processing; and the event pair relationship generation times are determined based on the semantic normalized user events.
The embodiment of the present application provides a possible implementation manner, wherein the apparatus further includes:
and the acquisition module is used for acquiring the second graph relationship networks of the plurality of target users with the association relationship, and performing network fusion on the second graph relationship networks of the plurality of target users with the association relationship to obtain a fourth graph relationship network.
The embodiment of the present application provides a possible implementation manner, wherein the user event includes at least one of the following:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
For the embodiment of the present application, the beneficial effects achieved by the embodiment of the present application are the same as those of the embodiment of the method described above, and are not described herein again.
Example four
The embodiment of the present disclosure provides a graph neural network model training apparatus 70, as shown in fig. 7, including:
a training module 701, configured to train a neural network model based on a plurality of training samples; the training sample has a graph relationship network containing timing information of any one of the embodiments.
The embodiment of the present application provides a possible implementation manner, where a user event of a graph relationship network including timing information includes at least one of the following:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
For the embodiment of the present application, the beneficial effects achieved by the embodiment of the present application are the same as those of the embodiment of the method described above, and are not described herein again.
In the technical scheme of the disclosure, the acquisition, storage, application and the like of the personal information of the related user all accord with the regulations of related laws and regulations, and do not violate the good customs of the public order.
The present disclosure also provides an electronic device, a readable storage medium, and a computer program product according to embodiments of the present disclosure.
The electronic device includes: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method as provided by the embodiments of the present disclosure.
Compared with the graph network established in the prior art, the graph network relationship established by the electronic equipment only contains the relationship between the user and the user event, and does not contain the time sequence information between the user events. The method comprises the steps of determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between a user and a user event; determining an event pair relationship existing among user events of a target user, wherein the event pair relationship represents that the user events occur successively; and adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information. In other words, the time sequence information among the user events is fused in the constructed second graph relationship network, so that the constructed graph network relationship containing the time sequence information can provide better user representation and user event node representation, and a good data basis is provided for subsequent graph relationship network data mining based on a graph neural network and the like.
The readable storage medium is a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform a method as provided by an embodiment of the present disclosure.
The readable storage medium only contains the relation between users and user events and does not contain the time sequence information between the user events compared with the graph network relation constructed by the prior art. The method comprises the steps of determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between a user and a user event; determining an event pair relationship existing among user events of a target user, wherein the event pair relationship represents that the user events occur successively; and adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information. In other words, the time sequence information among the user events is fused in the constructed second graph relationship network, so that the constructed graph network relationship containing the time sequence information can provide better user representation and user event node representation, and a good data basis is provided for subsequent graph relationship network data mining based on a graph neural network and the like.
The computer program product comprising a computer program which, when executed by a processor, implements a method as shown in the first aspect of the disclosure.
The computer program product only contains the relation between the user and the user event and does not contain the time sequence information between the user events compared with the graph network relation constructed by the prior art. The method comprises the steps of determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between a user and a user event; determining an event pair relationship existing among user events of a target user, wherein the event pair relationship represents that the user events occur successively; and adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information. In other words, the time sequence information among the user events is fused in the constructed second graph relationship network, so that the constructed graph network relationship containing the time sequence information can provide better user representation and user event node representation, and a good data basis is provided for subsequent graph relationship network data mining based on a graph neural network and the like.
FIG. 8 illustrates a schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the disclosure described and/or claimed herein.
As shown in fig. 8, the apparatus 800 includes a computing unit 801 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM)802 or a computer program loaded from a storage unit 808 into a Random Access Memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the device 800 can also be stored. The calculation unit 801, the ROM802, and the RAM 803 are connected to each other by a bus 804. An input/output (I/O) interface 805 is also connected to bus 804.
A number of components in the device 800 are connected to the I/O interface 805, including: an input unit 806, such as a keyboard, a mouse, or the like; an output unit 807 such as various types of displays, speakers, and the like; a storage unit 808, such as a magnetic disk, optical disk, or the like; and a communication unit 809 such as a network card, modem, wireless communication transceiver, etc. The communication unit 809 allows the device 800 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server with a combined blockchain.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present disclosure is not limited herein.
The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the scope of protection of the present disclosure.
Claims (20)
1. A graph relation network construction method containing time sequence information comprises the following steps:
determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between the user and a user event;
determining an event pair relationship existing between user events of a target user, wherein the event pair relationship represents that the user events occur successively;
and adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information.
2. The method of claim 1, wherein the method further comprises:
and determining the event pair relationship existing between the user events of the target user based on the event log of the target user.
3. The method of claim 2, wherein determining an event pair relationship existing between user events of the target user based on the event log of the target user comprises:
determining each user event and the occurrence sequence of the user events of the target user based on the event log of the target user;
and determining the event pair relationship existing between the user events of the target user based on the determined user events of the target user and the occurrence sequence of the user events.
4. The method of claim 3, wherein the method further comprises:
determining the occurrence frequency of each event pair relation based on the event log of the target user;
determining the weight of each event pair relation based on the occurrence frequency of the determined event pair relation;
and adding the determined weight of each event pair relationship to the second graph relationship network to obtain a third graph relationship network.
5. The method of claim 4, wherein the determining a number of occurrences of each event pair relationship based on the event log of the target user previously comprises:
determining a user event of the target user based on the event log of the target user;
carrying out semantic normalization processing on the user event of the target user to obtain the user event after the semantic normalization processing;
and determining the occurrence frequency of each event pair relation based on the user events after semantic normalization.
6. The method of claim 1, wherein the method further comprises:
and acquiring second graph relationship networks of a plurality of target users with incidence relations, and carrying out network fusion on the second graph relationship networks of the plurality of target users with the incidence relations to obtain a fourth graph relationship network.
7. The method of any of claims 1-6, wherein the user event comprises at least one of:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
8. A graph neural network model training method comprises the following steps:
training a neural network model based on a plurality of training samples; the training sample has the graph relationship network of any of claims 1-7 containing timing information.
9. The method of claim 8, wherein the user events of the graph relationship network containing timing information comprise at least one of:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
10. A graph relationship network construction apparatus containing timing information, comprising:
the first determination module is used for determining a first graph relation network of a target user, wherein the first graph relation network comprises the relation between the user and a user event;
the second determining module is used for determining an event pair relationship existing among user events of the target user, wherein the event pair relationship represents that the user events occur successively;
and the first adding module is used for adding the determined event pair relationship existing among the user events of the target user to the first graph relationship network to obtain a second graph relationship network containing the time sequence information.
11. The apparatus of claim 10, wherein the apparatus further comprises:
and the third determining module is used for determining the event pair relationship existing between the user events of the target user based on the event log of the target user.
12. The apparatus of claim 11, wherein the third determining means comprises:
the first determining unit is used for determining each user event and the occurrence sequence of the user events of the target user based on the event log of the target user;
and the second determining unit is used for determining the event pair relationship existing between the user events of the target user based on the determined user events of the target user and the occurrence sequence of the user events.
13. The apparatus of claim 12, wherein the apparatus further comprises:
the fourth determining module is used for determining the occurrence frequency of each event pair relation based on the event log of the target user;
the fifth determining module is used for determining the weight of each event pair relation based on the occurrence frequency of the determined event pair relation;
and the second adding module is specifically used for adding the determined weight of each event pair relationship to the second graph relationship network to obtain a third graph relationship network.
14. The apparatus according to claim 13, wherein the fourth determining module is specifically configured to determine the user event of the target user based on the event log of the target user; the semantic normalization processing is carried out on the user event of the target user to obtain the user event after the semantic normalization processing; and the event pair relationship generation times are determined based on the semantic normalized user events.
15. The apparatus of any of claims 1-6, wherein the user event comprises at least one of:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
16. A graph neural network model training device, comprising:
the training module is used for training the neural network model of the graph based on a plurality of training samples; the training sample has the graph relationship network of any of claims 1-7 containing timing information.
17. The apparatus of claim 16, wherein the user events of the graph relationship network containing timing information comprise at least one of:
watching a video event; browsing article events; retrieving the event; an advertisement click event.
18. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.
19. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-9.
20. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110792127.1A CN113515568A (en) | 2021-07-13 | 2021-07-13 | Graph relation network construction method, graph neural network model training method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110792127.1A CN113515568A (en) | 2021-07-13 | 2021-07-13 | Graph relation network construction method, graph neural network model training method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113515568A true CN113515568A (en) | 2021-10-19 |
Family
ID=78066999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110792127.1A Pending CN113515568A (en) | 2021-07-13 | 2021-07-13 | Graph relation network construction method, graph neural network model training method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113515568A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570013A (en) * | 2015-10-09 | 2017-04-19 | 北京国双科技有限公司 | Method and device for processing page access data |
| CN111177188A (en) * | 2019-12-30 | 2020-05-19 | 浙江邦盛科技有限公司 | Rapid massive time sequence data processing method based on aggregation edge and time sequence aggregation edge |
| CN111914569A (en) * | 2020-08-10 | 2020-11-10 | 哈尔滨安天科技集团股份有限公司 | Prediction method and device based on fusion map, electronic equipment and storage medium |
| CN112348269A (en) * | 2020-11-11 | 2021-02-09 | 北京航空航天大学 | Time series prediction modeling method of fusion graph structure |
| CN112347316A (en) * | 2020-10-21 | 2021-02-09 | 上海淇玥信息技术有限公司 | GraphSAGE-based bad preference behavior detection method and device and electronic equipment |
-
2021
- 2021-07-13 CN CN202110792127.1A patent/CN113515568A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570013A (en) * | 2015-10-09 | 2017-04-19 | 北京国双科技有限公司 | Method and device for processing page access data |
| CN111177188A (en) * | 2019-12-30 | 2020-05-19 | 浙江邦盛科技有限公司 | Rapid massive time sequence data processing method based on aggregation edge and time sequence aggregation edge |
| CN111914569A (en) * | 2020-08-10 | 2020-11-10 | 哈尔滨安天科技集团股份有限公司 | Prediction method and device based on fusion map, electronic equipment and storage medium |
| CN112347316A (en) * | 2020-10-21 | 2021-02-09 | 上海淇玥信息技术有限公司 | GraphSAGE-based bad preference behavior detection method and device and electronic equipment |
| CN112348269A (en) * | 2020-11-11 | 2021-02-09 | 北京航空航天大学 | Time series prediction modeling method of fusion graph structure |
Non-Patent Citations (1)
| Title |
|---|
| 段永朝: "《互联网思想十讲 北大讲义》", 31 October 2014 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114265979B (en) | Method for determining fusion parameters, information recommendation method and model training method | |
| CN112771564A (en) | Artificial intelligence engine that generates semantic directions for web sites to map identities for automated entity seeking | |
| CN110866040A (en) | User portrait generation method, device and system | |
| CN114036398A (en) | Content recommendation and ranking model training method, device, equipment and storage medium | |
| CN112560461A (en) | News clue generation method and device, electronic equipment and storage medium | |
| CN115099239A (en) | Resource identification method, device, equipment and storage medium | |
| CN114064925A (en) | Knowledge graph construction method, data query method, device, equipment and medium | |
| CN113722593B (en) | Event data processing method, device, electronic equipment and medium | |
| US20230041339A1 (en) | Method, device, and computer program product for user behavior prediction | |
| CN115563310A (en) | Method, device, equipment and medium for determining key service node | |
| CN114862479A (en) | Information pushing method and device, electronic equipment and medium | |
| CN113515568A (en) | Graph relation network construction method, graph neural network model training method and device | |
| CN114048315A (en) | Method and device for determining document tag, electronic equipment and storage medium | |
| CN114021642A (en) | Data processing method and device, electronic equipment and storage medium | |
| CN113961797A (en) | Resource recommendation method and device, electronic equipment and readable storage medium | |
| CN112948584A (en) | Short text classification method, device, equipment and storage medium | |
| CN113239296B (en) | Method, device, equipment and medium for displaying small program | |
| CN114297489B (en) | Content recommendation method and device, electronic equipment and storage medium | |
| CN113347079B (en) | Mail identification method, mail identification device, electronic equipment and readable storage medium | |
| CN113971216B (en) | Data processing method and device, electronic equipment and memory | |
| CN113190749B (en) | Method and device for determining object attribute, electronic equipment and medium | |
| US20230085599A1 (en) | Method and device for training tag recommendation model, and method and device for obtaining tag | |
| CN113934931A (en) | Information recommendation method, device, equipment, storage medium and program product | |
| CN115730580A (en) | Power supply service public opinion analysis method and device based on topic web crawler | |
| CN114329205A (en) | Method and device for pushing information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |