CN114880566A - User behavior analysis method, device, equipment and medium based on graph neural network - Google Patents
User behavior analysis method, device, equipment and medium based on graph neural network Download PDFInfo
- Publication number
- CN114880566A CN114880566A CN202210542103.5A CN202210542103A CN114880566A CN 114880566 A CN114880566 A CN 114880566A CN 202210542103 A CN202210542103 A CN 202210542103A CN 114880566 A CN114880566 A CN 114880566A
- Authority
- CN
- China
- Prior art keywords
- user
- recommended item
- node
- initial
- network
- 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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, and provides a user behavior analysis method, a device, equipment and a medium based on a graph neural network. The method comprises the following steps: dividing the initial bipartite graph into a training set and a verification set, and carrying out information propagation on the training set to obtain user node representation, recommended item node representation and edge labels; inputting the user node representation and the recommendation item node representation into a full-connection network to obtain a gradient training network and calculating a verification set to obtain a shielding value; if the shielding value is larger than the preset value, updating the initial bipartite graph, iteratively training a target network model for the updated bipartite graph according to the initial network model, and inputting the target bipartite graph into the target network model to obtain a result; and inputting the target bipartite graph into a target network model for analysis to obtain a behavior analysis result, and pushing recommendation items to the user to be tested according to the analysis result. The invention also relates to the technical field of block chains, and the user node and the recommendation item node can be stored in a node of a block chain.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a user behavior analysis method, a device, equipment and a medium based on a graph neural network.
Background
When the recommendation system is applied to a deep learning model, a user and a recommended item are generally used as initial network model input, training is performed by using whether the user has a historical click behavior on the recommended item or not as a label to obtain a target network model, and the target network model is applied to the prediction recommendation of the user on the recommended item.
However, the click behavior of the user has certain randomness or false touch, the click behavior is not necessarily the embodiment of the user interest, the click behavior is usually accompanied by a noise tag, and the noise tag interferes with the convergence of the model in the model learning process and influences the analysis capability of the target network model, so that the recommendation score given by the target network model cannot truly reflect the user interest.
Disclosure of Invention
In view of the above, the present invention provides a method, an apparatus, a device, and a medium for analyzing user behavior based on a graph neural network, and aims to solve the technical problems in the prior art that a noise label interferes with convergence of a prediction model in a learning process of the prediction model, and accuracy of user behavior analysis by the prediction model is not high.
In order to achieve the above object, the present invention provides a user behavior analysis method based on a graph neural network, including:
acquiring a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes;
dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set;
splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set;
if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model;
inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
Preferably, the dividing all the initial bipartite graphs into a training set and a verification set according to the connection edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set includes:
dividing all the initial bipartite graphs into a training set and a verification set according to the random proportion according to the connecting edges;
constructing a feature matrix according to the feature vectors of the user nodes and the recommended item nodes of the training set;
constructing an adjacency matrix according to the edge connection relation between the user nodes of the training set and the recommended item nodes;
and inputting the characteristic matrix and the adjacency matrix into the initial network model for information propagation to obtain user node representation and recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
Preferably, the initial network model includes more than two graph convolution network layers, and the inputting the feature matrix and the adjacency matrix into the initial network model for information propagation to obtain user node representations and recommended item node representations of the training set, and edge labels respectively corresponding to the user node representations and the recommended item node representations includes:
according to a preset information propagation formula, aggregating the eigenvectors of any node in each row of the characteristic matrix and the edge connection relation of the adjacent matrix in each graph convolution network layer to form the eigenvector of the next graph convolution network layer;
and when the information aggregation of the user node and the recommended item node in each graph convolution network layer is completed, obtaining user node representation, recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
Preferably, the preset information propagation formula includes:
wherein, X' is a feature vector aggregated in each layer of graph convolution network layer by the edge connection relationship between the feature vector of each row of any node of the feature matrix and the adjacent matrix, a is the adjacent matrix of the training set, D is the feature matrix of the training set, W is a learnable parameter, σ is an activation function, and X is the feature vector of each row of any node of the feature matrix.
Preferably, the splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a fully-connected network of the initial network model to obtain a gradient training network corresponding to the edge label includes:
splicing the user node representations and the recommended item node representations of the training set in pairs and inputting the spliced user node representations and the recommended item node representations into the fully-connected network;
and performing side prediction between the user nodes and the recommended item nodes of the training set according to the fully-connected network to obtain a gradient training network corresponding to the side labels.
Preferably, the performing a masking calculation on the verification set according to the gradient training network to obtain a masking value of a connection edge of the verification set includes:
calculating the network relationship between any user node and the recommended item node of the verification set according to the gradient training network to obtain a side network relationship value;
and carrying out shielding statistics on the connection edges of the verification set according to the edge network relation value to obtain the shielding value of the connection edges of the verification set.
Preferably, if the mask value is greater than a preset value, updating the initial bipartite graph includes:
a1, when the shielding value of the verification set is larger than a preset value, if a connection edge already exists between a user node and a recommended item node of the initial bipartite graph, reserving the connection edge;
a2, if no connecting edge exists between the user node and the recommended item node of the initial bipartite graph, performing edge connection between the user node and the recommended item node of the initial bipartite graph;
a3, repeating A1-A2 until the edge relations between all the user nodes and the recommended item nodes of the initial bipartite graph are updated, and obtaining an updated bipartite graph.
In order to achieve the above object, the present invention further provides a user behavior analysis device based on a graph neural network, the device comprising:
the acquisition module is used for: acquiring a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes;
the training module is used for: dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set;
the training module is further configured to: splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set;
the training module is further configured to: if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model;
an application module: inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
In order to achieve the above object, the present invention also provides an electronic device, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a program executable by the at least one processor to enable the at least one processor to perform the graph neural network-based user behavior analysis method of any one of claims 1 to 7.
To achieve the above object, the present invention further provides a computer readable medium storing a user behavior analysis program, which when executed by a processor, implements the steps of the graph neural network-based user behavior analysis method according to any one of claims 1 to 7.
The method comprises the steps of constructing an initial bipartite graph by historical click data of a user on recommended items, respectively constructing a feature matrix and an adjacent matrix according to user nodes, recommended item nodes and connecting edges of the initial bipartite graph, carrying out information propagation on the feature matrix and the adjacent matrix according to an initial network model to obtain a gradient training network, shielding connecting edges generated by random or mistaken click of the user on the recommended items in the initial bipartite graph according to the gradient training network, connecting the recommended item nodes interested by the user and the user nodes at the same time, reducing interference of noise labels in a model learning process, obtaining a target prediction model with higher accuracy in user behavior analysis, using the target prediction model for user behavior analysis, and pushing the recommended items interested by the user to the user.
Drawings
FIG. 1 is a schematic flow chart diagram illustrating a preferred embodiment of a user behavior analysis method based on graph neural network according to the present invention;
FIG. 2 is a block diagram of a user behavior analysis apparatus according to a preferred embodiment of the present invention;
FIG. 3 is a diagram of an electronic device according to a preferred embodiment of the present invention;
the implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention can acquire and process related data based on an artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result.
The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a robot technology, a biological recognition technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.
The invention provides a user behavior analysis method based on a graph neural network. Fig. 1 is a schematic method flow diagram of an embodiment of the user behavior analysis method based on the neural network of the present invention. The method may be performed by an electronic device, which may be implemented by software and/or hardware. The user behavior analysis method based on the graph neural network comprises the following steps:
step S10: the method comprises the steps of obtaining a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes.
In this embodiment, the preset database refers to a database built by a self-built enterprise or built by a third party according to information such as a user, a recommended item clicked by the user, and a click relationship between the user and the recommended item. The initial bipartite graph is a graph formed by information such as user nodes of historical users, recommendation item nodes and connection edges between the user nodes and the recommendation item nodes. The user node generates a network node according to the user information, such as the name, gender, occupation, address of standing, browsing behavior (such as website that the user likes to browse, item clicked, browsing duration, browsing times), and other attribute information. The recommended item node is a network node generated according to the attribute information of the user click action/browse action (attribute information of the clicked and browsed object, such as the name, function, consumer group, historical evaluation and the like of the article). The recommended item refers to an object clicked and browsed by the user.
Before acquiring a plurality of initial bipartite graphs from a preset database, acquiring historical behavior data of a user on recommended items from the preset database as an original data set, preprocessing the original data set to obtain the plurality of initial bipartite graphs (for example, preprocessing is data screening), and storing behavior information of the user on the recommended items in the preset database, for example, clicking behaviors of the user on the recommended items (recommended items) when the user browses a website. Some click behaviors represent that the user clicks on the recommended item due to interests or hobbies, some click behaviors may be due to random or misunderstood clicks of the user, and some recommended items may be liked by the user and are not clicked for some reasons (for example, the webpage browsing speed is too high, and no attention is paid to all non-clicks and browsing), but all click behaviors of the user on the recommended item are started from the user node on the bipartite graph, and one or more connecting edges are generated to be connected with different recommended item nodes.
Step S20: and dividing all the initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set.
Specifically, step S20 includes:
dividing all the initial bipartite graphs into a training set and a verification set according to the random proportion according to the connecting edges;
constructing a feature matrix according to the feature vectors of the user nodes and the recommended item nodes of the training set;
constructing an adjacency matrix according to the edge connection relation between the user nodes of the training set and the recommended item nodes;
and inputting the characteristic matrix and the adjacency matrix into the initial network model for information propagation to obtain user node representation and recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
In one embodiment, the initial network model includes more than two graph convolution network layers, and the inputting the feature matrix and the adjacency matrix into the initial network model for information propagation to obtain user node representations and recommended item node representations of the training set, and edge labels corresponding to the user node representations and the recommended item node representations respectively includes:
according to a preset information propagation formula, aggregating the eigenvectors of any node in each row of the characteristic matrix and the edge connection relation of the adjacent matrix in each graph convolution network layer to form the eigenvector of the next graph convolution network layer;
and when the information aggregation of the user node and the recommended item node in each graph convolution network layer is completed, obtaining user node representation, recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
In one embodiment, the preset information dissemination formula includes:
wherein, X' is a feature vector aggregated in each layer of graph convolution network layer by the edge connection relationship between the feature vector of each row of any node of the feature matrix and the adjacent matrix, a is the adjacent matrix of the training set, D is the feature matrix of the training set, W is a learnable parameter, σ is an activation function, and X is the feature vector of each row of any node of the feature matrix.
For convenience of representation, the number of nodes in each initial bipartite graph of the training set is denoted by "N" below.
The feature matrix is specifically constructed by the feature vector of each node of the initial bipartite graph, and the feature vector of each node contains the attribute information of the node, that is, the feature matrix can be constructed according to the attribute information carried by each node in each initial bipartite graph. For example, assuming that the length of the node feature vector is M, the node feature matrix of the initial bipartite graph is an N × M dimensional matrix, and each row of the matrix is a feature vector of one node. If the initial bipartite graph is a bipartite graph of the user and the recommended item, and the nodes in the initial bipartite graph are user nodes and recommended item nodes, the attribute information included in each feature vector in the feature matrix is (information such as the gender, age, and hobby of the user) and (information such as the name and function of the recommended item), respectively.
The adjacency matrix is specifically constructed according to an edge connection relation between nodes in the initial bipartite graph, the edge connection relation refers to the interaction depth of the user node to the recommended item node (for example, the number of times that the user clicks the recommended item or the browsing duration), and the adjacency matrix is specifically an N × N dimensional matrix. For example, a represents an adjacency matrix, a user node i and a recommendation node j are respectively two nodes in a bipartite graph, if there is no connecting edge between the user node i and the recommendation node j, the corresponding element of the user node i and the recommendation node j in the adjacency matrix is 0, and if there is a connecting edge between the user node i and the recommendation node j, the corresponding element of the user node i and the recommendation node j in the adjacency matrix is 1.
All the initial bipartite graphs are divided into training sets and verification sets according to a random proportion (for example, 10 ten thousand of all the initial bipartite graphs are randomly distributed in the training sets and the verification sets according to a ratio of 8:2 or 7: 3), that is, the random proportion is to keep the number of the training sets larger than that of the verification sets, so as to avoid that too little training data can cause under-fitting of the initial network model.
The initial network model includes, but is not limited to, a graph neural network (GCN), which is a feature extractor for node classification, graph classification, and edge prediction of graph data. And inputting the characteristic matrix and the adjacency matrix into an initial network model, and obtaining user node representation, recommended item node representation and edge labels of the training set through processing such as characteristic extraction, attribute information aggregation and the like. The node represents a fusion feature vector including each node (user node and recommendation item node), for example, if the length of the fusion feature vector of each node is F, the dimension of the fusion feature matrix of the node is N × F, where the length F of the fusion feature vector is a hyper-parameter of an initial network model set in advance (for example, F is set to 10), and the dimension of the node generating degree vector is N × 1, that is, the node representation refers to initial modeling of a recommendation item by a user, and is used for vectorization of a full matrix. The edge label refers to a vector continuation of the recommended item by the user, for example, I represents the user, if the user I clicks the recommended item J6 times, (I, J) is a connecting edge between the user node I and the recommended item node J, if the length of the vector of the connecting edge is 6, and when the attribute information of the recommended item K is similar to that of the recommended item J, the recommended item K is a vector continuation of the edge label between the user node I and the recommended item node J.
Step S30: and splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set.
In this embodiment, after the feature matrix and the adjacent matrix are input into the multilayer graph convolution layer for information exchange, the user node representation and the recommended item node representation are spliced two by two to form a mixed feature vector, and the mixed feature vector is input into the fully-connected network for side prediction to obtain a gradient training network corresponding to the side label. And inputting the user nodes and the recommended item nodes of the verification set into the gradient training network for shielding calculation, and determining whether a real network relationship exists between the user nodes and the recommended item nodes, so as to obtain a shielding value of a connecting edge of the verification set.
In an embodiment, the splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and recommended item node representation into a fully-connected network of the initial network model to obtain a gradient training network corresponding to the edge label includes:
splicing the user node representations and the recommended item node representations of the training set in pairs and inputting the spliced user node representations and the recommended item node representations into the fully-connected network;
and performing side prediction between the user nodes and the recommended item nodes of the training set according to the fully-connected network to obtain a gradient training network corresponding to the side labels.
In an embodiment, the performing a masking calculation on the verification set according to the gradient training network to obtain a masking value of a connecting edge of the verification set includes:
calculating the network relationship between any user node and the recommended item node of the verification set according to the gradient training network to obtain a side network relationship value;
and carrying out shielding statistics on the connection edges of the verification set according to the edge network relation value to obtain the shielding value of the connection edges of the verification set.
Inputting the user nodes and the recommended item nodes in the verification set into a gradient training network to obtain edge network relationship values between the nodes, comparing the edge network relationship values with preset network values to determine whether a real network relationship exists between the user nodes and the recommended item nodes, wherein the network relationship represents the closeness between the user nodes and the current recommended item, and judging the closeness of the user nodes to the connecting edges of the current recommended item through the attribute information of all clicked recommended items of the user nodes. After the comparison, the unreal connecting edges are shielded, the ratio value of the number of the shielded connecting edges to the number of the connecting edges of the verification set is counted, and the ratio value is used as the shielding value of the connecting edges of the verification set. For example, the number of connecting sides of the verification set is 10000, the number of shielded connecting sides is 3000, and 3000 is divided by 10000, so that the shielding value of the connecting sides of the verification set is 0.3.
For example, a user node i and a recommended item node j of the verification set are set, and (i, j) is a connecting edge between the user node i and the recommended item node j, a network relationship between the user node i and the recommended item node j is calculated according to the gradient training network, if a side network relationship value between the user node i and the recommended item node j is obtained and is 1, and according to a preset condition, when the side network relationship value is smaller than or equal to a preset network value (for example, the preset network value is 1), that is, the gradient training network judges that no side network relationship exists between the two nodes, that is, the gradient training network also represents that the connecting edge (i, j) is a random click or a false touch on the recommended item node j by the user node i, and that the connecting edge (i, j) between the user node i and the recommended item node j is shielded or deleted is equivalent to clearing a noise label on the initial bipartite graph.
If the edge network relationship value between the user node i and the recommended item node j is 3, and according to a preset condition, when the edge network relationship value is greater than a preset network value (for example, the preset network value is 1), namely the gradient training network judges that the edge network relationship exists between the two nodes, the edge network relationship value also represents that the user node i is likely to be interested in the recommended item node j and clicks, and the (i, j) connecting edges between the user node i and the recommended item node j are added, which is equivalent to generating click prediction on an initial bipartite graph.
Step S40: and if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model.
In this embodiment, when the mask value is greater than the preset value (for example, the preset value is 0.1), it indicates that the initial bipartite graph is still in an unstable state with more noise labels and training parameters of the initial network model, and the initial bipartite graph (all training sets and verification sets) is updated according to the mask value of the connecting edge of the verification set to obtain an updated initial bipartite graph.
And carrying out training iteration on the initial bipartite graph according to the initial network model, wherein the training iteration step comprises the following steps: dividing all initial bipartite graphs into a training set and a verification set according to a random proportion, performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and edge labels of the training set, splicing the user node representations and the recommended item node representations and inputting the spliced user node representations and the recommended item node representations into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge labels, performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set, and storing the training parameters to obtain a target network model when the shielding value is smaller than a preset value, namely, the noise-containing labels of the verification set are reduced to a smaller range, each training parameter of the initial network model tends to be stable, and the training parameters of the initial network model can be stored to obtain the target network model.
In an embodiment, if the mask value is greater than a preset value, the updating the initial bipartite graph includes:
a1, when the shielding value of the verification set is larger than a preset value, if a connection edge already exists between a user node and a recommended item node of the initial bipartite graph, reserving the connection edge;
a2, if no connecting edge exists between the user node and the recommended item node of the initial bipartite graph, performing edge connection between the user node and the recommended item node of the initial bipartite graph;
a3, repeating A1-A2 until the edge relations between all the user nodes and the recommended item nodes of the initial bipartite graph are updated, and obtaining an updated bipartite graph.
Step S50: inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
In this embodiment, the target network model is applied to a target bipartite graph for predicting the click behavior of the user to be tested on the recommended item, the target bipartite graph refers to a data graph of information such as a connection edge generated by clicking the recommended item by a certain user to be tested, the connection edge generated by random clicking of the user to be tested on the target bipartite graph is shielded according to the trained target network model, the recommended item node interested by the user to be tested is connected with the node of the user to be tested while the node of the recommended item is connected, a behavior analysis result of the user to be tested is obtained, and the recommended item interested by the user to be tested is pushed to the user to be tested according to the behavior analysis result.
Fig. 2 is a schematic diagram of functional modules of the user behavior analysis apparatus 100 according to the present invention.
The user behavior analysis apparatus 100 according to the present invention may be installed in an electronic device. According to the implemented functions, the user behavior analysis apparatus 100 may include an acquisition module 110, a training module 120, and an application module 130. A module according to the present invention, which may also be referred to as a unit, refers to a series of computer program segments that can be executed by a processor of an electronic device and that can perform a fixed function, and that are stored in a memory of the electronic device.
In this embodiment, the functions of the modules/units are as follows:
the obtaining module 110 is configured to: acquiring a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes;
the training module 120 is configured to: dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set;
the training module 120 is further configured to: splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set;
the training module 120 is further configured to: if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model;
the application module 130: inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
In one embodiment, the dividing all the initial bipartite graphs into a training set and a verification set according to the connection edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations, and corresponding edge labels of the training set includes:
dividing all the initial bipartite graphs into a training set and a verification set according to the random proportion according to the connecting edges;
constructing a feature matrix according to the feature vectors of the user nodes and the recommended item nodes of the training set;
constructing an adjacency matrix according to the edge connection relation between the user nodes of the training set and the recommended item nodes;
and inputting the characteristic matrix and the adjacency matrix into the initial network model for information propagation to obtain user node representation and recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
In one embodiment, the initial network model includes more than two graph convolution network layers, and the inputting the feature matrix and the adjacency matrix into the initial network model for information propagation to obtain user node representations and recommended item node representations of the training set, and edge labels corresponding to the user node representations and the recommended item node representations respectively includes:
according to a preset information propagation formula, aggregating the eigenvectors of any node in each row of the characteristic matrix and the edge connection relation of the adjacent matrix in each graph convolution network layer to form the eigenvector of the next graph convolution network layer;
and when the information aggregation of the user node and the recommended item node in each graph convolution network layer is completed, obtaining user node representation, recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
In one embodiment, the preset information dissemination formula includes:
wherein, X' is a feature vector aggregated in each layer of graph convolution network layer by the edge connection relationship between the feature vector of each row of any node of the feature matrix and the adjacent matrix, a is the adjacent matrix of the training set, D is the feature matrix of the training set, W is a learnable parameter, σ is an activation function, and X is the feature vector of each row of any node of the feature matrix.
In an embodiment, the splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and recommended item node representation into a fully-connected network of the initial network model to obtain a gradient training network corresponding to the edge label includes:
splicing the user node representations and the recommended item node representations of the training set in pairs and inputting the spliced user node representations and the recommended item node representations into the fully-connected network;
and performing side prediction between the user nodes and the recommended item nodes of the training set according to the fully-connected network to obtain a gradient training network corresponding to the side labels.
In an embodiment, the performing a masking calculation on the verification set according to the gradient training network to obtain a masking value of a connecting edge of the verification set includes:
calculating the network relationship between any user node and the recommended item node of the verification set according to the gradient training network to obtain a side network relationship value;
and carrying out shielding statistics on the connection edges of the verification set according to the edge network relation value to obtain the shielding value of the connection edges of the verification set.
In an embodiment, if the mask value is greater than a preset value, the updating the initial bipartite graph includes:
a1, when the shielding value of the verification set is larger than a preset value, if a connection edge already exists between a user node and a recommended item node of the initial bipartite graph, reserving the connection edge;
a2, if no connecting edge exists between the user node and the recommended item node of the initial bipartite graph, performing edge connection between the user node and the recommended item node of the initial bipartite graph;
a3, repeating A1-A2 until the edge relations between all the user nodes and the recommended item nodes of the initial bipartite graph are updated, and obtaining an updated bipartite graph.
Fig. 3 is a schematic diagram of an electronic device 1 according to a preferred embodiment of the invention.
The electronic device 1 includes but is not limited to: memory 11, processor 12, display 13, and network interface 14. The electronic device 1 is connected to a network through a network interface 14 to obtain raw data. The network may be a wireless or wired network such as an Intranet (Intranet), the Internet (Internet), a global system for mobile communications (GSM), Wideband Code Division Multiple Access (WCDMA), a 4G network, a 5G network, Bluetooth (Bluetooth), Wi-Fi, or a call network.
The memory 11 includes at least one type of readable medium including a flash memory, a hard disk, a multimedia card, a card type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the storage 11 may be an internal storage unit of the electronic device 1, such as a hard disk or a memory of the electronic device 1. In other embodiments, the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash card (FlashCard), or the like, which is equipped with the electronic device 1. Of course, the memory 11 may also comprise both an internal memory unit and an external memory device of the electronic device 1. In this embodiment, the memory 11 is generally used for storing an operating system installed in the electronic device 1 and various application software, such as program codes of the user behavior analysis program 10. Further, the memory 11 may also be used to temporarily store various types of data that have been output or are to be output.
Processor 12 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 12 is typically used for controlling the overall operation of the electronic device 1, such as performing data interaction or communication related control and processing. In this embodiment, the processor 12 is configured to run the program code stored in the memory 11 or process data, for example, run the program code of the user behavior analysis program 10.
The display 13 may be referred to as a display screen or display unit. In some embodiments, the display 13 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an organic light-emitting diode (OLED) touch panel, or the like. The display 13 is used for displaying information processed in the electronic device 1 and for displaying a visual work interface, e.g. displaying the results of data statistics.
The network interface 14 may optionally comprise a standard wired interface, a wireless interface (e.g. WI-FI interface), the network interface 14 typically being used for establishing a communication connection between the electronic device 1 and other electronic devices.
Fig. 3 only shows the electronic device 1 with the components 11-14 and the user behavior analysis program 10, but it is to be understood that not all shown components are required to be implemented, and that more or fewer components may be implemented instead.
Optionally, the electronic device 1 may further comprise a user interface, the user interface may comprise a Display (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface may further comprise a standard wired interface and a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an organic light-emitting diode (OLED) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable for displaying information processed in the electronic device 1 and for displaying a visualized user interface, among other things.
The electronic device 1 may further include a Radio Frequency (RF) circuit, a sensor, an audio circuit, and the like, which are not described in detail herein.
In the above embodiment, the processor 12, when executing the user behavior analysis program 10 stored in the memory 11, may implement the following steps:
acquiring a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes;
dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set;
splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set;
if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model;
inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
The storage device may be the memory 11 of the electronic device 1, or may be another storage device communicatively connected to the electronic device 1.
For detailed description of the above steps, please refer to the above description of fig. 2 regarding a functional block diagram of an embodiment of the user behavior analysis apparatus 100 and fig. 1 regarding a flowchart of an embodiment of a user behavior analysis method based on a graph neural network.
In addition, the embodiment of the present invention further provides a computer-readable medium, which may be non-volatile or volatile. The computer readable medium may be any one or any combination of hard disk, multimedia card, SD card, flash memory card, SMC, Read Only Memory (ROM), Erasable Programmable Read Only Memory (EPROM), portable compact disc read only memory (CD-ROM), USB memory, and the like. The computer readable medium includes a storage data area and a storage program area, the storage data area stores data created according to the use of the blockchain node, the storage program area stores a user behavior analysis program 10, and the user behavior analysis program 10 implements the following operations when executed by a processor:
acquiring a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes;
dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set;
splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set;
if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model;
inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
The embodiment of the computer readable medium of the present invention is substantially the same as the embodiment of the user behavior analysis method based on the graph neural network, and will not be described herein again.
In another embodiment, in order to further ensure the privacy and security of all the appearing data, all the data may be stored in a node of a block chain. Such as user nodes, recommendation nodes, and the like, which may be stored in block chain nodes.
It should be noted that the blockchain in the present invention is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
It should be noted that the above-mentioned numbers of the embodiments of the present invention are merely for description, and do not represent the merits of the embodiments. And the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that includes the element.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention essentially or contributing to the prior art can be embodied in the form of a software product, which is stored in a medium (such as ROM/RAM, magnetic disk, optical disk) as described above and includes instructions for enabling a terminal device (such as a mobile phone, a computer, an electronic device, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A user behavior analysis method based on a graph neural network is characterized by comprising the following steps:
acquiring a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes;
dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set;
splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set;
if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model;
inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
2. The graph neural network-based user behavior analysis method of claim 1, wherein the dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set comprises:
dividing all the initial bipartite graphs into a training set and a verification set according to the random proportion according to the connecting edges;
constructing a feature matrix according to the feature vectors of the user nodes and the recommended item nodes of the training set;
constructing an adjacency matrix according to the edge connection relation between the user nodes of the training set and the recommended item nodes;
and inputting the characteristic matrix and the adjacency matrix into the initial network model for information propagation to obtain user node representation and recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
3. The method of claim 2, wherein the initial network model comprises two or more graph convolutional network layers, and the inputting the feature matrix and the adjacency matrix into the initial network model for information propagation to obtain the user node representation and the recommended item node representation of the training set, and the user node representation and the recommended item node representation respectively correspond to edge labels comprises:
according to a preset information propagation formula, aggregating the eigenvectors of any node in each row of the characteristic matrix and the edge connection relation of the adjacent matrix in each graph convolution network layer to form the eigenvector of the next graph convolution network layer;
and when the information aggregation of the user node and the recommended item node in each graph convolution network layer is completed, obtaining user node representation, recommended item node representation of the training set, and edge labels respectively corresponding to the user node representation and the recommended item node representation.
4. The graph neural network-based user behavior analysis method of claim 3, wherein the preset information propagation formula comprises:
wherein, X' is a feature vector aggregated in each layer of graph convolution network layer by the edge connection relationship between the feature vector of each row of any node of the feature matrix and the adjacent matrix, a is the adjacent matrix of the training set, D is the feature matrix of the training set, W is a learnable parameter, σ is an activation function, and X is the feature vector of each row of any node of the feature matrix.
5. The graph neural network-based user behavior analysis method of claim 1, wherein the splicing the user node representations and the recommended item node representations and inputting the spliced user node representations and the recommended item node representations into a fully-connected network of the initial network model to obtain a gradient training network corresponding to the edge labels comprises:
splicing the user node representations and the recommended item node representations of the training set in pairs and inputting the spliced user node representations and the recommended item node representations into the fully-connected network;
and performing side prediction between the user nodes and the recommended item nodes of the training set according to the fully-connected network to obtain a gradient training network corresponding to the side labels.
6. The graph neural network-based user behavior analysis method of claim 1, wherein the performing a masking computation on the verification set according to the gradient training network to obtain a masking value of a connecting edge of the verification set comprises:
calculating the network relationship between any user node and the recommended item node of the verification set according to the gradient training network to obtain a side network relationship value;
and carrying out shielding statistics on the connection edges of the verification set according to the edge network relation value to obtain the shielding value of the connection edges of the verification set.
7. The graph neural network-based user behavior analysis method of claim 1, wherein if the masked value is greater than a preset value, updating the initial bipartite graph comprises:
a1, when the shielding value of the verification set is larger than a preset value, if a connection edge already exists between a user node and a recommended item node of the initial bipartite graph, reserving the connection edge;
a2, if no connecting edge exists between the user node and the recommended item node of the initial bipartite graph, performing edge connection between the user node and the recommended item node of the initial bipartite graph;
a3, repeating A1-A2 until the edge relations between all the user nodes and the recommended item nodes of the initial bipartite graph are updated, and obtaining an updated bipartite graph.
8. An apparatus for analyzing user behavior based on a graph neural network, the apparatus comprising:
the acquisition module is used for: acquiring a plurality of initial bipartite graphs from a preset database, wherein the initial bipartite graphs are composed of user nodes of historical users, recommended item nodes and connecting edges between the user nodes and the recommended item nodes, and the connecting edges of the initial bipartite graphs are determined according to behavior information generated by the user nodes of the initial bipartite graphs to the recommended item nodes;
the training module is used for: dividing all initial bipartite graphs into a training set and a verification set according to the connecting edges, and performing information propagation on the training set according to an initial network model to obtain user node representations, recommended item node representations and corresponding edge labels of the training set;
the training module is further configured to: splicing the user node representation and the recommended item node representation and inputting the spliced user node representation and the recommended item node representation into a full-connection network of the initial network model to obtain a gradient training network corresponding to the edge label, and performing shielding calculation on the verification set according to the gradient training network to obtain a shielding value of a connection edge of the verification set;
the training module is further configured to: if the shielding value is larger than a preset value, updating the initial bipartite graph, and performing iterative training on the updated bipartite graph according to the initial network model until the shielding value is smaller than the preset value to obtain a target network model;
an application module: inputting a target bipartite graph formed by user nodes, recommended item nodes and connecting edges between the user nodes and the recommended item nodes of the user to be tested into the target network model for analysis to obtain a behavior analysis result of the user to be tested, and pushing recommended items to the user to be tested according to the analysis result.
9. An electronic device, characterized in that the electronic device comprises:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a program executable by the at least one processor to enable the at least one processor to perform the graph neural network-based user behavior analysis method of any one of claims 1 to 7.
10. A computer-readable medium, in which a user behavior analysis program is stored, and the user behavior analysis program, when executed by a processor, implements the steps of the graph neural network-based user behavior analysis method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210542103.5A CN114880566A (en) | 2022-05-17 | 2022-05-17 | User behavior analysis method, device, equipment and medium based on graph neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210542103.5A CN114880566A (en) | 2022-05-17 | 2022-05-17 | User behavior analysis method, device, equipment and medium based on graph neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114880566A true CN114880566A (en) | 2022-08-09 |
Family
ID=82676175
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210542103.5A Pending CN114880566A (en) | 2022-05-17 | 2022-05-17 | User behavior analysis method, device, equipment and medium based on graph neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114880566A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115270005A (en) * | 2022-09-30 | 2022-11-01 | 腾讯科技（深圳）有限公司 | Information recommendation method, device, equipment and storage medium |
| CN117057929A (en) * | 2023-10-11 | 2023-11-14 | 中邮消费金融有限公司 | Abnormal user behavior detection method, device, equipment and storage medium |
-
2022
- 2022-05-17 CN CN202210542103.5A patent/CN114880566A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115270005A (en) * | 2022-09-30 | 2022-11-01 | 腾讯科技（深圳）有限公司 | Information recommendation method, device, equipment and storage medium |
| CN117057929A (en) * | 2023-10-11 | 2023-11-14 | 中邮消费金融有限公司 | Abnormal user behavior detection method, device, equipment and storage medium |
| CN117057929B (en) * | 2023-10-11 | 2024-01-26 | 中邮消费金融有限公司 | Abnormal user behavior detection method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110866181B (en) | Resource recommendation method, device and storage medium | |
| CN110598845B (en) | Data processing method, data processing device, computer equipment and storage medium | |
| CN112085565B (en) | Deep learning-based information recommendation method, device, equipment and storage medium | |
| CN114880566A (en) | User behavior analysis method, device, equipment and medium based on graph neural network | |
| CN114372573B (en) | User portrait information recognition method and device, computer equipment and storage medium | |
| CN111723292B (en) | Recommendation method, system, electronic equipment and storage medium based on graph neural network | |
| CN112328909B (en) | Information recommendation method and device, computer equipment and medium | |
| US10909145B2 (en) | Techniques for determining whether to associate new user information with an existing user | |
| CN111738780A (en) | Method and system for recommending object | |
| CN111275205A (en) | Virtual sample generation method, terminal device and storage medium | |
| CN111177568A (en) | Object pushing method based on multi-source data, electronic device and storage medium | |
| CN113343091A (en) | Industrial and enterprise oriented science and technology service recommendation calculation method, medium and program | |
| CN114565196B (en) | Multi-event trend prejudging method, device, equipment and medium based on government affair hotline | |
| CN111198967A (en) | User grouping method and device based on relational graph and electronic equipment | |
| CN111625567A (en) | Data model matching method, device, computer system and readable storage medium | |
| CN113434762A (en) | Association pushing method, device and equipment based on user information and storage medium | |
| CN111489196B (en) | Prediction method and device based on deep learning network, electronic equipment and medium | |
| CN113569162A (en) | Data processing method, device, equipment and storage medium | |
| CN111475720A (en) | Recommendation method, recommendation device, server and storage medium | |
| CN115204971B (en) | Product recommendation method, device, electronic equipment and computer readable storage medium | |
| CN116127188A (en) | Target feedback value determining method and device, electronic equipment and storage medium | |
| CN115392361A (en) | Intelligent sorting method and device, computer equipment and storage medium | |
| CN115952438A (en) | Social platform user attribute prediction method and system, mobile device and storage medium | |
| CN108304407B (en) | Method and system for sequencing objects | |
| CN115186188A (en) | Product recommendation method, device and equipment based on behavior analysis and storage medium |
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 |