Resource pushing method and device, equipment, medium and program product

Abstract

The invention provides a resource pushing method and device, equipment, a medium and a program product, and is applied to the technical field of artificial intelligence and the technical field of financial science and technology. The method comprises the following steps: acquiring a user behavior data stream related to a user, wherein user behavior tensors corresponding to a plurality of specified moments in the user behavior data stream represent interaction behaviors executed by the user at the plurality of specified moments; according to the difference between at least two adjacent user behavior tensors, performing incremental tensor decomposition on the specified user behavior tensor to obtain time factor data and behavior feature data related to the specified moment; performing interaction behavior evolution analysis on the user based on the time factor data corresponding to the plurality of specified moments to obtain behavior evolution data representing a user interaction behavior change trend; and determining an interaction intention of the user based on the behavior evolution data and the behavior feature data, and pushing a target resource matched with the interaction intention to the user.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence technology and financial technology technology, and more specifically to a resource delivery method, apparatus, device, medium and program product. Background Technology

[0002] When pushing resources to users, users are segmented by integrating metadata (such as age and region) and behavioral data (such as transaction frequency), and push strategies are formulated for different users based on this.

[0003] In related technologies, analyzing a large number of historical interaction behaviors in a single dimension makes it difficult to accurately push resources to users, and the low analysis efficiency makes it difficult to support real-time push. Summary of the Invention

[0004] In view of the above problems, this application provides a resource push method, apparatus, device, medium and program product.

[0005] According to a first aspect of this application, a resource push method is provided, comprising: acquiring a user behavior data stream related to a user, wherein user behavior tensors corresponding to multiple specified times in the user behavior data stream represent the interactive behaviors performed by the user at the multiple specified times; performing incremental tensor decomposition on the specified user behavior tensors based on the differences between at least two adjacent user behavior tensors to obtain time factor data and behavioral feature data related to the specified times, wherein the behavioral feature data represents multiple types of behavioral attributes related to the interactive behaviors; performing interactive behavior evolution analysis on the user based on the time factor data corresponding to the multiple specified times to obtain behavioral evolution data representing the changing trend of the user's interactive behaviors; determining the user's interactive intent based on the behavioral evolution data and behavioral feature data, and pushing target resources matching the interactive intent to the user.

[0006] According to an embodiment of this application, based on the difference between at least two adjacent user behavior tensors, an incremental tensor decomposition is performed on a specified user behavior tensor to obtain time factor data and behavioral feature data related to a specified time. This includes: performing orthogonal projection processing on the (n-1)th user behavior tensor and the nth user behavior tensor from at least two adjacent user behavior tensors to obtain a difference time factor matrix, which represents the difference in time factor data between the (n-1)th time and the nth time; and processing the difference time factor matrix and the (n-1)th user behavior tensor to obtain time factor data and behavioral feature data related to the nth time.

[0007] According to an embodiment of this application, the difference time factor matrix and the (n-1)th user behavior tensor are processed to obtain time factor data and behavioral feature data related to the nth time moment, including: multiplying the difference time factor matrix and the (n-1)th user behavior tensor to obtain the nth target user behavior tensor; for the mth round of M rounds, performing incremental tensor decomposition on the nth target user behavior tensor to obtain the mth core tensor and the mth time-varying factor matrix, wherein, if the rank of the mth time-varying factor matrix satisfies a preset condition, the mth time-varying factor matrix is ​​determined as the nth time-varying factor matrix of the nth target user behavior tensor; and determining the time factor data and behavioral feature data related to the nth time moment based on the nth time-varying factor matrix.

[0008] According to an embodiment of this application, incremental tensor decomposition is performed on the nth target user behavior tensor to obtain the mth core tensor and the mth time-varying factor matrix, including: performing singular value decomposition on the core matrix corresponding to the (m-1)th core tensor to obtain the mth core tensor; and performing incremental tensor decomposition on the nth target user behavior tensor based on the mth core tensor to obtain the mth time-varying factor matrix.

[0009] According to an embodiment of this application, user interaction behavior evolution analysis is performed based on time factor data corresponding to multiple specified times to obtain behavioral evolution data characterizing the trend of user interaction behavior changes. This includes: for the nth time among multiple specified times, determining the time factor data difference value at the nth time based on the difference between the time factor data at the (n-1)th time and the time factor data at the nth time; and determining the behavioral evolution data at the nth time based on the time factor data difference value at the nth time and the time decay coefficient.

[0010] According to embodiments of this application, determining a user's interaction intent based on behavioral evolution data and behavioral feature data, and pushing target resources matching the interaction intent to the user, includes: analyzing user information feature data, interaction information feature data, and interaction channel feature data of behavioral feature data to obtain the user's interaction behavior type; determining the user's interaction intent based on the numerical range of behavioral evolution data and the user's interaction behavior type; and pushing target resources matching the interaction intent to the user based on the user's interaction intent.

[0011] According to embodiments of this application, determining a user's interaction intent based on the numerical range of behavioral evolution data and the type of user interaction behavior includes: determining the user's interaction intent as a first interaction channel intent when the numerical range of behavioral evolution data is greater than a first preset threshold and the user interaction behavior type is an interaction channel change type; determining the user's interaction intent as a second interaction channel intent when the numerical range of behavioral evolution data is greater than the first preset threshold and the user interaction behavior type is an interaction channel association type; and determining the user's interaction intent as an interaction frequency intent when the numerical range of behavioral evolution data is less than the first preset threshold and the user interaction behavior type is an interaction frequency type.

[0012] According to an embodiment of this application, obtaining a user-related user behavior data stream includes: for any interaction channel, obtaining the user's interaction behavior performed at multiple specified times; determining the user-related user behavior data stream based on the user's interaction behavior performed at multiple specified times; wherein, in the absence of obtained interaction behavior, the user-related user behavior data stream is determined based on historical interaction behavior.

[0013] A second aspect of this application provides a resource push device, comprising: a first acquisition module, configured to acquire a user behavior data stream related to a user, wherein user behavior tensors corresponding to multiple specified times in the user behavior data stream represent the interactive behaviors performed by the user at the multiple specified times; a first decomposition module, configured to perform incremental tensor decomposition on a specified user behavior tensor based on the differences between at least two adjacent user behavior tensors, to obtain time factor data and behavioral feature data related to the specified times, wherein the behavioral feature data represents multiple types of behavioral attributes related to the interactive behaviors; a first analysis module, configured to perform interactive behavior evolution analysis on the user based on the time factor data corresponding to the multiple specified times, to obtain behavioral evolution data representing the changing trend of the user's interactive behaviors; and a first push module, configured to determine the user's interactive intent based on the behavioral evolution data and behavioral feature data, and push target resources matching the interactive intent to the user.

[0014] A third aspect of this application provides an electronic device comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the method described above.

[0015] A fourth aspect of this application also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0016] The fifth aspect of this application also provides a computer program product, including a computer program or instructions that, when executed by a processor, implement the steps of the above-described method.

[0017] According to embodiments of this application, a user behavior data stream related to the user is acquired; based on the differences between at least two adjacent user behavior tensors, an incremental tensor decomposition is performed on a specified user behavior tensor to obtain time factor data and behavioral feature data related to a specified time; based on the time factor data corresponding to multiple specified times, an interaction behavior evolution analysis is performed on the user to obtain behavioral evolution data characterizing the trend of user interaction behavior changes; based on the behavioral evolution data and behavioral feature data, the user's interaction intent is determined, and target resources matching the interaction intent are pushed to the user. Because the specified user behavior tensor is incrementally decomposed through multiple rounds of decomposition, the complexity of the decomposed user behavior tensor is reduced, resulting in low-rank behavioral evolution data. This allows for the analysis of large volumes of behavioral evolution data and behavioral feature data during the process of determining the user's interaction intent based on behavioral evolution data and behavioral feature data, thereby improving analysis efficiency. Furthermore, since the behavioral feature data is multi-dimensional data determined based on the user behavior data stream related to the user, it can enrich the dimensions of analysis and improve the accuracy and rationality of the interaction intent while improving analysis efficiency during the process of determining the user's interaction intent based on behavioral evolution data and behavioral feature data. Attached Figure Description

[0018] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0019] Figure 1 The illustration shows application scenario diagrams of resource push methods, apparatus, devices, media, and program products according to embodiments of this application;

[0020] Figure 2 A flowchart of a resource push method according to an embodiment of this application is shown;

[0021] Figure 3 A schematic diagram of a resource push method according to an embodiment of this application is shown;

[0022] Figure 4 A structural block diagram of a resource push device according to an embodiment of this application is shown; and

[0023] Figure 5 A block diagram of an electronic device suitable for implementing a resource push method according to an embodiment of this application is shown. Detailed Implementation

[0024] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0025] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0026] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0027] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0028] In the technical solution of this application, the user information (including but not limited to user personal information, user image information, user device information, such as location information) and data (including but not limited to data used for analysis, stored data, and displayed data) involved are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant laws, regulations, and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entry points for users to choose to authorize or refuse.

[0029] In scenarios involving automated decision-making using personal information, the methods, devices, and systems provided in this application all offer users corresponding entry points for choosing to agree to or reject the automated decision-making results. If the user chooses to reject, the process proceeds to the expert decision-making stage. Here, "automated decision-making" refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests, or economic, health, and credit status through computer programs, and then making a decision. Here, "expert decision-making" refers to the activity of making decisions by personnel who specialize in a particular field, possess specialized experience, knowledge, and skills, and have reached a certain level of professional expertise.

[0030] Related technologies integrate metadata (such as age and region) and behavioral data (such as transaction frequency) to segment users and formulate marketing strategies based on these models. However, they rely solely on historical transaction data and cannot capture real-time behavioral changes. Furthermore, they do not integrate behavioral correlations across product types, making it difficult to identify migration paths. Because they use fixed-time slices for statistics, they cannot model the continuous time-series dynamic evolution of user preferences.

[0031] Related technologies use time-series behavior modeling to process user behavior time-series data. By predicting the user's next interaction time or interaction product preference, it has certain advantages in processing continuous time series data and mining complex behavior patterns. However, single sequence modeling is difficult to integrate multi-dimensional features, and the serial structure of time-series behavior modeling leads to high training time, making it difficult to support real-time updates and output the display path of user preference migration.

[0032] Related technologies mostly use two-dimensional matrices or single time series, which cannot capture multi-dimensional time series correlations. The static features obtained cannot reflect dynamic behavioral patterns, resulting in the inability to identify cross-product line collaborative preference migration. Furthermore, batch modeling leads to poor data timeliness, making it impossible to capture sudden behavioral changes and significantly reducing marketing response rates. Traditional tensor decomposition is too complex to support real-time analysis of massive amounts of customer data.

[0033] Embodiments of this application provide a resource push method, which involves acquiring a user behavior data stream related to a user, wherein user behavior tensors corresponding to multiple specified times in the user behavior data stream represent the interactive behaviors performed by the user at multiple specified times; performing incremental tensor decomposition on the specified user behavior tensors based on the differences between at least two adjacent user behavior tensors to obtain time factor data and behavioral feature data related to the specified times, wherein the behavioral feature data represents multiple types of behavioral attributes related to the interactive behaviors; performing interactive behavior evolution analysis on the user based on the time factor data corresponding to multiple specified times to obtain behavioral evolution data representing the changing trend of user interactive behaviors; determining the user's interactive intent based on the behavioral evolution data and behavioral feature data, and pushing target resources matching the interactive intent to the user.

[0034] Figure 1 The illustration shows application scenario diagrams of resource push methods, apparatus, devices, media, and program products according to embodiments of this application.

[0035] like Figure 1 As shown, application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0036] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0037] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0038] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0039] It should be noted that the resource push method provided in this application embodiment can generally be executed by server 105. Correspondingly, the resource push device provided in this application embodiment can generally be located in server 105. The resource push method provided in this application embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the resource push device provided in this application embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105.

[0040] It should be understood that Figure 1 The number of terminal devices, networks, and servers in the system is only a limited number. Depending on implementation needs, there can be any number of terminal devices, networks, and servers.

[0041] The following will be based on Figure 1 The described scene, through Figures 2-3 The resource push method according to the embodiments of this application will be described in detail.

[0042] Figure 2 A flowchart of a resource push method according to an embodiment of this application is shown.

[0043] like Figure 2 As shown, the resource push method in this embodiment includes operations S210 to S240.

[0044] In operation S210, a user behavior data stream related to the user is acquired. The user behavior tensor corresponding to multiple specified times in the user behavior data stream represents the interactive behavior performed by the user at multiple specified times.

[0045] Based on a preset time period, multiple timestamps are determined, with the timestamp at which an interaction occurs being designated as a specific moment. The preset time period can be set to 5 minutes; for example, user behavior is detected every 5 minutes, and the user's interaction is determined to occur at the 5th and 10th minutes, with the 5th and 10th minutes being designated as multiple specific moments.

[0046] Interactive behavior can be any action a user takes on a specific channel, such as the number of clicks on an interactive interface of a platform or the transaction value of a specific service. If, at a given moment, a user's interactive behavior includes both the number of clicks on an interactive interface of a platform and the transaction value of a specific service, then the transaction value of the specific service will be prioritized as the user's interactive behavior.

[0047] User-related behavior data streams can be composed of four dimensions, including a specified time, user information, interaction information, and interaction channel. For example, a user-related behavior data stream could be that at a specified time T1, a user with user information U1 performs an interaction with interaction information A1 through interaction channel C1.

[0048] In embodiments of this application, user consent or authorization can be obtained before acquiring user behavior data streams related to the user. For example, a request to acquire user information can be sent to the user before operation S210. If the user consents or authorizes the acquisition of user information, operation S210 is performed.

[0049] In the embodiments of this application, a corresponding operation entry point can be provided to the user, allowing the user to choose to agree to or reject the automated decision result. That is, before acquiring the user-related user behavior data stream, the user's instruction to agree to or reject the processing / decision can be obtained through the corresponding operation entry point. If the user agrees to the processing / decision, the user-related user behavior data stream is acquired, i.e., step S210 is executed. If the user refuses to perform the processing / decision, the expert decision-making process is initiated.

[0050] In operation S220, based on the difference between at least two adjacent user behavior tensors, incremental tensor decomposition is performed on the specified user behavior tensor to obtain time factor data and behavior feature data related to the specified time. The behavior feature data represents multiple types of behavior attributes related to the interaction behavior.

[0051] At least two adjacent user behavior tensors can be user behavior tensors at at least two adjacent specified time points. The user behavior tensors are obtained by encoding them in a four-dimensional structure. Wherein, the specified time is T, the user information is U, the interaction information is A, and the interaction channel is C.

[0052] The difference between at least two adjacent user behavior tensors can be the difference between the user behavior tensors at at least two adjacent times at a specified time dimension.

[0053] Incremental tensor decomposition decomposes the user behavior tensor into a product of a core tensor and a time-varying factor matrix, where the rank of the time-varying factor matrix depends on the size of the core tensor.

[0054] The time-varying factor matrix can be four matrices that are respectively associated with a specified time T, user information U, interaction information A, and interaction channel C.

[0055] Based on the differences between at least two adjacent user behavior tensors, incremental tensor decomposition is performed on the specified user behavior tensor. This can be achieved by updating the specified user behavior tensor using the differences, and then performing incremental tensor decomposition on the updated specified user behavior tensor to obtain time factor data and behavioral feature data related to the specified time.

[0056] Specifically, incremental tensor decomposition is performed on the updated specified user behavior tensor to obtain the core tensor and time-varying factor matrix of the updated specified user behavior tensor. This allows the determination of time factor data related to a specified time, as well as behavioral feature data related to user information, interaction information, and interaction channels in multiple types of behavioral attributes, from the time-varying factor matrix.

[0057] In operation S230, user interaction behavior evolution analysis is performed based on time factor data corresponding to multiple specified times to obtain behavioral evolution data that characterizes the trend of user interaction behavior changes.

[0058] Interactive behavior evolution analysis can be used to analyze the changing trends of interactive behaviors, determine the differences between multiple interactive behaviors, and the changing trends of the differences between multiple interactive behaviors.

[0059] By analyzing time factor data corresponding to multiple specified times, it is possible to analyze the changing trend of user interaction frequency and determine whether the user interaction frequency has suddenly increased.

[0060] In operation S240, the user's interaction intent is determined based on behavioral evolution data and behavioral feature data, and target resources matching the interaction intent are pushed to the user.

[0061] Based on behavioral evolution data and behavioral feature data, changes in user interaction behavior can be analyzed, such as whether the frequency of interaction has changed or whether the interaction channel has changed, thereby determining the user's interaction intent. User interaction intent may include changing the interaction channel or maintaining the interaction behavior.

[0062] After determining the user's interaction intent, target resources matching that intent are pushed to the user. For example, if the user's interaction intent is to change the interaction channel, then interaction offers for different channels are pushed to the user.

[0063] According to embodiments of this application, a user behavior data stream related to the user is acquired; based on the differences between at least two adjacent user behavior tensors, an incremental tensor decomposition is performed on a specified user behavior tensor to obtain time factor data and behavioral feature data related to a specified time; based on the time factor data corresponding to multiple specified times, an interaction behavior evolution analysis is performed on the user to obtain behavioral evolution data characterizing the trend of user interaction behavior changes; based on the behavioral evolution data and behavioral feature data, the user's interaction intent is determined, and target resources matching the interaction intent are pushed to the user. Because the specified user behavior tensor is incrementally decomposed through multiple rounds of decomposition, the complexity of the decomposed user behavior tensor is reduced, resulting in low-rank behavioral evolution data. This allows for the analysis of large volumes of behavioral evolution data and behavioral feature data during the process of determining the user's interaction intent based on behavioral evolution data and behavioral feature data, thereby improving analysis efficiency. Furthermore, since the behavioral feature data is multi-dimensional data determined based on the user behavior data stream related to the user, it can enrich the dimensions of analysis and improve the accuracy and rationality of the interaction intent while improving analysis efficiency during the process of determining the user's interaction intent based on behavioral evolution data and behavioral feature data.

[0064] According to an embodiment of this application, based on the difference between at least two adjacent user behavior tensors, an incremental tensor decomposition is performed on a specified user behavior tensor to obtain time factor data and behavioral feature data related to a specified time. This includes: performing orthogonal projection processing on the (n-1)th user behavior tensor and the nth user behavior tensor from at least two adjacent user behavior tensors to obtain a difference time factor matrix, which represents the difference in time factor data between the (n-1)th time and the nth time; and processing the difference time factor matrix and the (n-1)th user behavior tensor to obtain time factor data and behavioral feature data related to the nth time.

[0065] Orthogonal projection is used to project the tensor of the nth user behavior onto the factor matrix of the (n-1)th user behavior tensor, preserving the original structural information of the (n-1)th user behavior tensor while extracting the changed parts in the tensor of the nth user behavior.

[0066] For example, the factor matrix of the (n-1)th user behavior tensor has already captured most of the features. The projection of the nth user behavior tensor onto the factor matrix of the (n-1)th user behavior tensor can remove the existing information, and the remaining part can represent new changes or noise. This remaining part is used as the difference time factor matrix.

[0067] The difference time factor matrix represents the new factor of the nth user behavior tensor in the time dimension of the factor matrix of the (n-1)th user behavior tensor, that is, the difference in time factor data between the (n-1)th time and the nth time.

[0068] After determining the difference time factor matrix, the difference time factor matrix and the (n-1)th user behavior tensor are processed to obtain the updated nth user behavior tensor projection. Then, incremental tensor decomposition is performed on the nth user behavior tensor projection to obtain the time factor data and behavior feature data related to the nth time.

[0069] According to an embodiment of this application, the nth user behavior tensor is projected into the factor matrix of the (n-1)th user behavior tensor to extract the difference time factor matrix related to the time dimension. The difference time factor matrix can be expanded or updated to reflect the time changes of the user when performing interactive behavior.

[0070] According to an embodiment of this application, the difference time factor matrix and the (n-1)th user behavior tensor are processed to obtain time factor data and behavioral feature data related to the nth time moment, including: multiplying the difference time factor matrix and the (n-1)th user behavior tensor to obtain the nth target user behavior tensor; for the mth round of M rounds, performing incremental tensor decomposition on the nth target user behavior tensor to obtain the mth core tensor and the mth time-varying factor matrix, wherein, if the rank of the mth time-varying factor matrix satisfies a preset condition, the mth time-varying factor matrix is ​​determined as the nth time-varying factor matrix of the nth target user behavior tensor; and determining the time factor data and behavioral feature data related to the nth time moment based on the nth time-varying factor matrix.

[0071] Multiplying the difference time factor matrix and the (n-1)th user behavior tensor can be done by multiplying the transpose of the difference time factor matrix and the factor matrix of the (n-1)th user behavior tensor, folding the resulting matrix into a new tensor, and obtaining the nth target user behavior tensor.

[0072] After obtaining the nth target user behavior tensor, incremental tensor decomposition is performed on it. The decomposition process involves M rounds, where each round, based on the decomposition result of the previous round, further decomposes the nth target user behavior tensor to determine the nth time-varying factor matrix of the nth target user behavior tensor.

[0073] The preset condition can be that if the rank of the m-th time-varying factor matrix is ​​the smallest among the rank of the time-varying factor matrices in each of the M rounds, then it is determined as the n-th time-varying factor matrix of the n-th target user behavior tensor; or it can be that if the rank of the m-th time-varying factor matrix satisfies a preset threshold, then the iteration stops and the m-th time-varying factor matrix is ​​determined as the n-th time-varying factor matrix of the n-th target user behavior tensor.

[0074] The nth time-varying factor matrix of the nth target user behavior tensor includes time factor data and behavioral feature data related to the nth time. After obtaining the nth time-varying factor matrix, time factor data and behavioral feature data can be extracted from the nth time-varying factor matrix.

[0075] According to an embodiment of this application, by performing incremental tensor decomposition on the nth target user behavior tensor in multiple rounds, the rank of the time-varying factor matrix obtained in each round gradually decreases, reducing the complexity of the time-varying factor matrix. This reduces the complexity of analyzing large amounts of time factor data and behavioral feature data in subsequent operations, thereby improving analysis efficiency.

[0076] According to an embodiment of this application, incremental tensor decomposition is performed on the nth target user behavior tensor to obtain the mth core tensor and the mth time-varying factor matrix, including: performing singular value decomposition on the core matrix corresponding to the (m-1)th core tensor to obtain the mth core tensor; and performing incremental tensor decomposition on the nth target user behavior tensor based on the mth core tensor to obtain the mth time-varying factor matrix.

[0077] In the process of incremental tensor decomposition of the nth target user behavior tensor, the core tensor of each round is determined based on the core tensor of the previous round.

[0078] In this process, the core tensor of the previous round is expanded into a core matrix, the singular values ​​of the core matrix are calculated, and the singular values ​​are used to perform singular value decomposition on the core matrix to construct a new core matrix. This new core matrix is ​​then folded to obtain a new core tensor, which is then used as the core tensor of the current round.

[0079] The decomposition result obtained by incremental tensor decomposition of the target user behavior tensor is represented as follows: ,in, For the target user behavior tensor, For the core tensor, Given the time-varying factor matrix, based on the core tensor of this round, the time-varying factor matrix of this round can be determined.

[0080] For example, for the first round, incremental tensor decomposition is performed on the tensor of the nth target user's behavior to obtain the first core tensor and the first time-varying factor matrix. , Let n be the target user behavior tensor. For the first core tensor, This is the first time-varying factor matrix.

[0081] In the second round, singular value decomposition is performed on the core matrix corresponding to the first core tensor to obtain the second core tensor. Based on the second core tensor Incremental tensor decomposition is performed on the nth target user behavior tensor to obtain the second time-varying factor matrix. .

[0082] According to the embodiments of this application, by using singular values ​​to perform singular value decomposition on the core matrix of the core tensor of the previous round, a new core matrix is ​​constructed to obtain a new core tensor, thereby avoiding the loss of tensor information caused by operations such as random updating of the core tensor, and maintaining sparsity and stability in the decomposition process.

[0083] According to an embodiment of this application, user interaction behavior evolution analysis is performed based on time factor data corresponding to multiple specified times to obtain behavioral evolution data characterizing the trend of user interaction behavior changes. This includes: for the nth time among multiple specified times, determining the time factor data difference value at the nth time based on the difference between the time factor data at the (n-1)th time and the time factor data at the nth time; and determining the behavioral evolution data at the nth time based on the time factor data difference value at the nth time and the time decay coefficient.

[0084] For time n among multiple specified times, the difference between the time factor data at time n-1 and time factor data at time n is determined based on the difference between the time factor data at time n-1 and time factor data at time n, thus obtaining the time factor data difference value at time n.

[0085] After determining the time factor data difference value at time n, the norm of the absolute value of the time factor data difference value at time n is taken. Furthermore, after obtaining the norm of the absolute value of the time factor data difference value at time n, it is divided by the time decay coefficient to obtain the initial behavioral evolution data at time n.

[0086] After determining the initial behavioral evolution data at time n, the Sigmoid function is used to map the initial behavioral evolution data at time n to the range of 0-1, thus obtaining the behavioral evolution data at time n.

[0087] As shown in formula (1).

[0088] (1).

[0089] in, For the behavioral evolution data at time n, For the time factor data at time n, For the time factor data at time n-1, The time decay coefficient, This is the Sigmoid function.

[0090] According to the embodiments of this application, by comparing the time factor data at time n-1 and time factor data at time n, it is possible to analyze whether the difference between the time factor data at time n-1 and time factor data at time n is too large or too small, and accurately analyze the behavioral evolution data that characterizes the trend of user interaction behavior changes.

[0091] According to embodiments of this application, determining a user's interaction intent based on behavioral evolution data and behavioral feature data, and pushing target resources matching the interaction intent to the user, includes: analyzing user information feature data, interaction information feature data, and interaction channel feature data of behavioral feature data to obtain the user's interaction behavior type; determining the user's interaction intent based on the numerical range of behavioral evolution data and the user's interaction behavior type; and pushing target resources matching the interaction intent to the user based on the user's interaction intent.

[0092] The time-varying factor matrix can be four matrices that are respectively associated with a specified time T, user information U, interaction information A, and interaction channel C. Among them, the behavioral feature data includes three matrices that are respectively associated with user information U, interaction information A, and interaction channel C. Specifically, the user information feature data of the behavioral feature data is the user information matrix U associated with user information in the time-varying factor matrix, the interaction information feature data is the interaction information matrix A associated with interaction information in the time-varying factor matrix, and the interaction channel feature data is the interaction channel matrix C associated with interaction channel in the time-varying factor matrix.

[0093] Analyzing user information feature data, interaction information feature data, and interaction channel feature data based on behavioral feature data can involve coupling these three types of data to obtain... The coupling result is used to quantify the changes in the interaction behavior of user i with interaction information j, and the changes in the interaction channel k. Then, based on the changes in interaction behavior and the changes in the interaction channel, the type of user interaction behavior is determined.

[0094] After determining the type of user interaction behavior, the user's interaction intent is determined based on the numerical range of the behavior evolution data and the type of user interaction behavior, so as to determine the user's interaction preference for multiple interactions.

[0095] After determining the user's interaction preferences across multiple interactions, the system can push target resources that match the user's interaction intent to the user.

[0096] In addition, time-series trajectory diagrams can be drawn from behavioral feature data to determine the path of user behavior changes, thereby analyzing the user's interaction intent.

[0097] According to the embodiments of this application, by analyzing user information feature data, interaction information feature data and interaction channel feature data of behavioral feature data, it is possible to analyze user behavior preferences based on the sum of multiple dimensions, accurately identify user interaction behavior types, and further, based on the numerical range of behavioral evolution data and user interaction behavior types, accurately analyze user interaction intentions, not limited to single data, and improve the rationality of user interaction intentions.

[0098] According to embodiments of this application, determining a user's interaction intent based on the numerical range of behavioral evolution data and the type of user interaction behavior includes: determining the user's interaction intent as a first interaction channel intent when the numerical range of behavioral evolution data is greater than a first preset threshold and the user interaction behavior type is an interaction channel change type; determining the user's interaction intent as a second interaction channel intent when the numerical range of behavioral evolution data is greater than the first preset threshold and the user interaction behavior type is an interaction channel association type; and determining the user's interaction intent as an interaction frequency intent when the numerical range of behavioral evolution data is less than a second preset threshold and the user interaction behavior type is an interaction frequency type.

[0099] The first preset threshold can be 0.8, which can be set according to actual needs.

[0100] If a user's interaction channels change multiple times, the user's interaction behavior type is determined as an interaction channel change type. The interaction channel change type includes an interaction channel change rate threshold.

[0101] If BEE > 0.8 and the interaction channel change rate > 30%, and the user's interaction intent is determined to be the first interaction channel intent, then the target resource matching the first interaction channel intent can be to push new channel offers to the user (such as a pop-up window guiding the user to jump to the stock trading page).

[0102] If the correlation between product types across multiple channels interacted with by a user increases, then the user interaction behavior type is determined to be an interaction channel association type. Interaction channel association types include those with increased correlation between interaction channels.

[0103] If BEE > 0.8 and the correlation between interaction channels is enhanced, and the user's interaction intent is determined to be the intent of the second interaction channel, then the target resource matching the intent of the second interaction channel can be a dormant product that activates the channel (such as recommending related index fund services to money market fund users).

[0104] The second preset threshold can be 0.2, while the first preset threshold can be set according to actual needs.

[0105] If the frequency of a user's multiple interactions matches the frequency of their historical interactions, then the user's interaction behavior type is determined to be an interaction frequency type.

[0106] If BEE < 0.2 and the user interaction behavior type is the interaction frequency type, and the user's interaction intent is determined to be the interaction frequency intent, then the target resource that matches the interaction frequency intent can be a push strategy (such as a do-not-disturb policy for users with fixed deposits).

[0107] In addition, behavioral characteristic data can be analyzed before pushing pushes to users to determine the probability of different users responding to pushes. Based on the probability of different users responding to pushes, the priority of pushes for different users can be determined so that users with a high probability of responding to pushes can receive the target resources pushed in a timely manner.

[0108] According to embodiments of this application, different judgment strategies based on the numerical range of behavioral evolution data and the types of user interaction behaviors can determine multiple interaction intentions, enrich the diversity of judgment results, and enable diversified and reasonable push to different users, thereby improving the user's push response rate.

[0109] According to an embodiment of this application, obtaining a user-related user behavior data stream includes: for any interaction channel, obtaining the user's interaction behavior performed at multiple specified times; determining the user-related user behavior data stream based on the user's interaction behavior performed at multiple specified times; wherein, in the absence of obtained interaction behavior, the user-related user behavior data stream is determined based on historical interaction behavior.

[0110] Acquiring user-related user behavior data streams can involve collecting data from different channels, including the interaction behaviors of different users on any given interaction channel and the specific time when those interactions occurred.

[0111] In addition, if an anomaly occurs during the collection process, or if it is necessary to collect user behavior data streams at times when no interaction occurs, i.e., if no interaction is obtained, then for any user on any interaction channel, the historical interaction behavior of that user on that interaction channel will be obtained.

[0112] After obtaining historical interaction behavior, if the historical interaction behavior includes multiple historical transaction values, the average of the multiple historical transaction values ​​is taken as the interaction behavior related to the user, thereby determining the user behavior data stream related to the user.

[0113] According to embodiments of this application, user behavior data streams related to users are acquired, and the interactive behaviors of different users on different channels are captured, thereby enabling the user behavior tensor to capture multi-dimensional behavioral attributes and thus improve the accuracy of users' interactive intentions.

[0114] Figure 3 A schematic diagram of a resource push method according to an embodiment of this application is shown.

[0115] like Figure 3 As shown, in operation S310, user behavior data streams related to the user are acquired. In operation S320, orthogonal projection processing is performed on the (n-1)th user behavior tensor and the nth user behavior tensor to determine the nth target user behavior tensor. In operation S330, incremental tensor decomposition is performed on the nth target user behavior tensor to obtain the mth core tensor and the mth time-varying factor matrix. In operation S340, it is determined whether the rank of the mth time-varying factor matrix meets the preset condition. If it does not meet the condition, operation S330 is continued; if it does meet the condition, operation S350 is executed to determine the nth time-varying factor matrix of the nth target user behavior tensor. In operation S360, time factor data and behavioral feature data related to the nth time are determined. In operation S370, the user's interaction intent is determined based on the behavior evolution data and behavior feature data, and target resources matching the interaction intent are pushed to the user.

[0116] Based on the above resource push method, this application also provides a resource push device. The following will combine... Figure 4 The device is described in detail.

[0117] Figure 4 A structural block diagram of a resource push device according to an embodiment of this application is shown.

[0118] like Figure 4 As shown, the resource push device 400 in this embodiment includes a first acquisition module 410, a first decomposition module 420, a first analysis module 430, and a first push module 440.

[0119] The first acquisition module 410 is used to acquire user behavior data streams related to the user. The user behavior tensors corresponding to multiple specified times in the user behavior data stream represent the interactive behaviors performed by the user at those multiple specified times. In one embodiment, the first acquisition module 410 can be used to execute the operation S210 described above, which will not be repeated here.

[0120] The first decomposition module 420 is used to perform incremental tensor decomposition on a specified user behavior tensor based on the differences between at least two adjacent user behavior tensors, to obtain time factor data and behavioral feature data related to a specified time. The behavioral feature data represents multiple types of behavioral attributes related to the interaction behavior. In one embodiment, the first decomposition module 420 can be used to perform the operation S220 described above, which will not be repeated here.

[0121] The first analysis module 430 is used to perform interactive behavior evolution analysis on users based on time factor data corresponding to multiple specified times, and obtain behavioral evolution data characterizing the changing trend of user interactive behavior. In one embodiment, the first analysis module 430 can be used to execute the operation S230 described above, which will not be repeated here.

[0122] The first push module 440 is used to determine the user's interaction intent based on behavioral evolution data and behavioral feature data, and to push target resources that match the interaction intent to the user. In one embodiment, the first push module 440 can be used to perform the operation S240 described above, which will not be repeated here.

[0123] According to embodiments of this application, a user behavior data stream related to the user is acquired; based on the differences between at least two adjacent user behavior tensors, an incremental tensor decomposition is performed on a specified user behavior tensor to obtain time factor data and behavioral feature data related to a specified time; based on the time factor data corresponding to multiple specified times, an interaction behavior evolution analysis is performed on the user to obtain behavioral evolution data characterizing the trend of user interaction behavior changes; based on the behavioral evolution data and behavioral feature data, the user's interaction intent is determined, and target resources matching the interaction intent are pushed to the user. Because the specified user behavior tensor is incrementally decomposed through multiple rounds of decomposition, the complexity of the decomposed user behavior tensor is reduced, resulting in low-rank behavioral evolution data. This allows for the analysis of large volumes of behavioral evolution data and behavioral feature data during the process of determining the user's interaction intent based on behavioral evolution data and behavioral feature data, thereby improving analysis efficiency. Furthermore, since the behavioral feature data is multi-dimensional data determined based on the user behavior data stream related to the user, it can enrich the dimensions of analysis and improve the accuracy and rationality of the interaction intent while improving analysis efficiency during the process of determining the user's interaction intent based on behavioral evolution data and behavioral feature data.

[0124] According to an embodiment of this application, the first decomposition module 420 includes a first projection submodule and a first processing submodule.

[0125] The first projection submodule is used to perform orthogonal projection processing on the (n-1)th user behavior tensor and the nth user behavior tensor among at least two adjacent user behavior tensors to obtain a difference time factor matrix, which represents the difference in time factor data between the (n-1)th time and the nth time.

[0126] The first processing submodule is used to process the difference time factor matrix and the (n-1)th user behavior tensor to obtain time factor data and behavior feature data related to the nth time.

[0127] According to an embodiment of this application, the first processing submodule includes a first obtaining unit, a second obtaining unit, and a first determining unit.

[0128] The first unit is used to multiply the difference time factor matrix and the (n-1)th user behavior tensor to obtain the nth target user behavior tensor.

[0129] The second obtaining unit is used to perform incremental tensor decomposition on the nth target user behavior tensor for the mth round of M rounds, to obtain the mth core tensor and the mth time-varying factor matrix. Wherein, if the rank of the mth time-varying factor matrix satisfies a preset condition, the mth time-varying factor matrix is ​​determined as the nth time-varying factor matrix of the nth target user behavior tensor.

[0130] The first determining unit is used to determine the time factor data and behavioral feature data related to the nth time based on the nth time-varying factor matrix.

[0131] According to an embodiment of this application, the second obtaining unit includes a first decomposition subunit and a second decomposition subunit.

[0132] The first decomposition subunit is used to perform singular value decomposition on the core matrix corresponding to the (m-1)th core tensor to obtain the m-th core tensor.

[0133] The second decomposition subunit is used to perform incremental tensor decomposition on the nth target user behavior tensor based on the mth core tensor, to obtain the mth time-varying factor matrix.

[0134] According to an embodiment of this application, the first analysis module 430 includes a first determination submodule and a second determination submodule.

[0135] The first determination submodule is used to determine the time factor data difference value at time n based on the difference between the time factor data at time n-1 and the time factor data at time n, for a given time among multiple specified times.

[0136] The second determination submodule is used to determine the behavioral evolution data at time n based on the time factor data difference value and time decay coefficient at time n.

[0137] According to an embodiment of this application, the first push module 440 includes a first analysis submodule, a third determination submodule, and a first push submodule.

[0138] The first analysis submodule is used to analyze user information feature data, interaction information feature data, and interaction channel feature data of behavioral feature data to obtain the user interaction behavior type.

[0139] The third determination submodule is used to determine the user's interaction intent based on the numerical range of the behavior evolution data and the type of user interaction behavior.

[0140] The first push submodule is used to push target resources that match the user's interaction intent to the user.

[0141] According to an embodiment of this application, the third determining submodule includes a second determining unit, a third determining unit, and a fourth determining unit.

[0142] The second determining unit is used to determine the user's interaction intent as the first interaction channel intent when the numerical range of the behavior evolution data is greater than the first preset threshold and the user interaction behavior type is the interaction channel change type.

[0143] The third determining unit is used to determine the user's interaction intent as the second interaction channel intent when the numerical range of the behavior evolution data is greater than the first preset threshold and the user interaction behavior type is the interaction channel association type.

[0144] The fourth determining unit is used to determine the user's interaction intent as an interaction frequency intent when the numerical range of the behavior evolution data is less than the second preset threshold and the user interaction behavior type is an interaction frequency type.

[0145] According to an embodiment of this application, the first acquisition module 410 includes a first acquisition submodule and a fourth determination submodule.

[0146] The first acquisition submodule is used to acquire the user's interactive behaviors at multiple specified times for any interaction channel;

[0147] The fourth determination submodule is used to determine the user behavior data stream related to the user based on the user's interactive behavior performed at multiple specified times; wherein, in the absence of interactive behavior, the user behavior data stream related to the user is determined based on historical interactive behavior.

[0148] According to embodiments of this application, any multiple modules among the first acquisition module 410, first decomposition module 420, first analysis module 430, and first push module 440 can be merged into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module. According to embodiments of this application, at least one of the first acquisition module 410, first decomposition module 420, first analysis module 430, and first push module 440 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging the circuitry, or implemented in any one of the three implementation methods of software, hardware, and firmware, or in a suitable combination of any of these. Alternatively, at least one of the first acquisition module 410, the first decomposition module 420, the first analysis module 430, and the first push module 440 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0149] Figure 5 A block diagram of an electronic device suitable for implementing a resource push method according to an embodiment of this application is shown.

[0150] like Figure 5 As shown, an electronic device 500 according to an embodiment of this application includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage portion 508 into a random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include onboard memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0151] RAM 503 stores various programs and data required for the operation of electronic device 500. Processor 501, ROM 502, and RAM 503 are interconnected via bus 504. Processor 501 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 502 and / or RAM 503. It should be noted that the programs may also be stored in one or more memories other than ROM 502 and RAM 503. Processor 501 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in said one or more memories.

[0152] According to embodiments of this application, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to a bus 504. The electronic device 500 may also include one or more of the following components connected to the input / output (I / O) interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 510 as needed so that computer programs read from it can be installed into the storage section 508 as needed.

[0153] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0154] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include ROM 502 and / or RAM 503 and / or one or more memories other than ROM 502 and RAM 503 described above.

[0155] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to enable the computer system to implement the resource push method provided in the embodiments of this application.

[0156] When the computer program is executed by the processor 501, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0157] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 509, and / or installed from a removable medium 511. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0158] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by processor 501, it performs the functions defined in the system of this application embodiment. According to embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0159] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0160] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0161] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

Claims

1. A resource push method, characterized in that, The method includes: Acquire user behavior data streams related to users, wherein user behavior tensors corresponding to multiple specified times in the user behavior data streams represent the interactive behaviors performed by the user at multiple specified times; Based on the difference between at least two adjacent user behavior tensors, an incremental tensor decomposition is performed on the specified user behavior tensor to obtain time factor data and behavior feature data related to the specified time. The behavior feature data characterizes multiple types of behavior attributes related to the interaction behavior. Based on time factor data corresponding to multiple specified times, the user's interactive behavior evolution analysis is performed to obtain behavioral evolution data characterizing the trend of the user's interactive behavior changes; Based on the behavioral evolution data and the behavioral feature data, the user's interaction intent is determined, and target resources matching the interaction intent are pushed to the user.

2. The method according to claim 1, characterized in that, The step of performing incremental tensor decomposition on a specified user behavior tensor based on the difference between at least two adjacent user behavior tensors to obtain time factor data and behavioral feature data related to the specified time moment includes: For at least two adjacent user behavior tensors, the (n-1)th user behavior tensor and the nth user behavior tensor, Orthogonal projection processing is performed on the (n-1)th user behavior tensor and the nth user behavior tensor to obtain the difference time factor matrix, which represents the difference in time factor data between the (n-1)th time and the nth time. The difference time factor matrix and the (n-1)th user behavior tensor are processed to obtain time factor data and behavior feature data related to the nth time.

3. The method according to claim 3, characterized in that, The process of processing the difference time factor matrix and the (n-1)th user behavior tensor yields time factor data and behavioral feature data related to the nth time moment, including: Multiply the difference time factor matrix and the (n-1)th user behavior tensor to obtain the nth target user behavior tensor; For the m-th round of M rounds, the n-th target user behavior tensor is decomposed into an incremental tensor to obtain the m-th core tensor and the m-th time-varying factor matrix. Wherein, if the rank of the m-th time-varying factor matrix satisfies a preset condition, the m-th time-varying factor matrix is ​​determined as the n-th time-varying factor matrix of the n-th target user behavior tensor. Based on the nth time-varying factor matrix, determine the time factor data and behavioral feature data related to the nth time.

4. The method according to claim 3, characterized in that, The incremental tensor decomposition of the nth target user behavior tensor to obtain the mth core tensor and the mth time-varying factor matrix includes: Singular value decomposition is performed on the core matrix corresponding to the (m-1)th core tensor to obtain the m-th core tensor; Based on the m-th core tensor, the n-th target user behavior tensor is decomposed incrementally to obtain the m-th time-varying factor matrix.

5. The method according to claim 1, characterized in that, The step of performing an evolutionary analysis of the user's interactive behavior based on time factor data corresponding to multiple specified times to obtain behavioral evolution data characterizing the changing trend of the user's interactive behavior includes: For the nth time among the multiple specified times, Based on the difference between the time factor data at time n-1 and the time factor data at time n, determine the difference value of the time factor data at time n; The behavioral evolution data at time n is determined based on the difference in time factor data and time decay coefficient at time n.

6. The method according to claim 1, characterized in that, The step of determining the user's interaction intent based on the behavioral evolution data and the behavioral feature data, and pushing target resources matching the interaction intent to the user, includes: The user information feature data, interaction information feature data, and interaction channel feature data of the behavioral feature data are analyzed to obtain the user interaction behavior type. The user's interaction intent is determined based on the numerical range of the behavioral evolution data and the type of user interaction behavior. Based on the user's interaction intent, target resources that match the interaction intent are pushed to the user.

7. The method according to claim 6, characterized in that, Determining the user's interaction intent based on the numerical range of the behavioral evolution data and the type of user interaction behavior includes: If the numerical range of the behavior evolution data is greater than a first preset threshold, and the user interaction behavior type is an interaction channel change type, then the user's interaction intent is determined to be the first interaction channel intent. If the numerical range of the behavior evolution data is greater than the first preset threshold, and the user interaction behavior type is an interaction channel association type, then the user's interaction intent is determined to be a second interaction channel intent. If the numerical range of the behavior evolution data is less than the second preset threshold, and the user interaction behavior type is the interaction frequency type, then the user's interaction intent is determined to be the interaction frequency intent.

8. The method according to claim 1, characterized in that, The acquisition of user-related user behavior data streams includes: For any given interaction channel, obtain the user's interaction behaviors performed at multiple specified times; Based on the user's interactive behaviors performed at multiple specified times, determine the user behavior data stream related to the user; In cases where no interaction behavior is obtained, user behavior data streams related to the user are determined based on historical interaction behavior.

9. A resource delivery device, characterized in that, The device includes: The first acquisition module is used to acquire user behavior data streams related to users, wherein user behavior tensors corresponding to multiple specified times in the user behavior data streams represent the interactive behaviors performed by the user at multiple specified times. The first decomposition module is used to perform incremental tensor decomposition on a specified user behavior tensor based on the difference between at least two adjacent user behavior tensors, to obtain time factor data and behavior feature data related to the specified time, wherein the behavior feature data characterizes multiple types of behavior attributes related to the interaction behavior. The first analysis module is used to perform interactive behavior evolution analysis on the user based on time factor data corresponding to multiple specified times, to obtain behavioral evolution data characterizing the changing trend of the user's interactive behavior; and The first push module is used to determine the user's interaction intent based on the behavior evolution data and the behavior feature data, and to push target resources that match the interaction intent to the user.

10. An electronic device, comprising: One or more processors; Memory, used to store one or more computer programs. The characteristic feature is that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 8.

12. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1 to 8.

Images