Resource allocation method, device, storage medium and electronic device

Abstract

The present invention discloses a resource allocation method, device, storage medium, and electronic device. The method includes: obtaining first resource change data of a first account within a first time period, and second resource change data of a second account within the first time period; performing feature extraction on the first resource change data and the second resource change data, respectively, to obtain a first feature corresponding to the first resource change data, and a second feature corresponding to the second resource change data; integrating and processing the first feature and the second feature based on the resource interaction data generated during the resource interaction operation to obtain a target feature; obtaining a resource allocation coefficient corresponding to the first account based on the target feature; and determining a target resource value allocated to the first account according to the resource allocation coefficient. This method can be applied in artificial intelligence scenarios, and specifically involves technologies such as machine learning. The present invention solves the technical problem of low resource allocation accuracy.

Description

Technical Field

[0001] The present invention relates to the field of computers, and in particular to a resource allocation method, device, storage medium and electronic equipment. Background Art

[0002] Existing account scoring methods only consider the consumption data of the lending account, without considering the trustworthiness of the lending account, which can also be reflected in other lending accounts with which it transacts. This incompleteness leads to technical issues such as low resource allocation accuracy.

[0003] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0004] Embodiments of the present invention provide a resource allocation method, apparatus, storage medium, and electronic device to at least solve the technical problem of low resource allocation accuracy.

[0005] According to one aspect of an embodiment of the present invention, a resource allocation method is provided, comprising: obtaining first resource change data of a first account within a first time period, and second resource change data of a second account within the above-mentioned first time period, wherein the above-mentioned second account is an account that jointly completes a resource interaction operation with the above-mentioned first account within the second time period; performing feature extraction on the above-mentioned first resource change data and the above-mentioned second resource change data respectively to obtain a first feature corresponding to the above-mentioned first resource change data, and a second feature corresponding to the above-mentioned second resource change data; integrating the above-mentioned first feature and the above-mentioned second feature according to the resource interaction data generated in the above-mentioned resource interaction operation to obtain a target feature; obtaining a resource allocation coefficient corresponding to the above-mentioned first account based on the above-mentioned target feature; and determining a target resource value allocated to the above-mentioned first account according to the above-mentioned resource allocation coefficient.

[0006] According to another aspect of an embodiment of the present invention, a resource allocation device is also provided, including: a first acquisition unit, used to acquire first resource change data of a first account within a first time period, and second resource change data of a second account within the above-mentioned first time period, wherein the above-mentioned second account is an account that jointly completes resource interaction operations with the above-mentioned first account within the second time period; an extraction unit, used to perform feature extraction on the above-mentioned first resource change data and the above-mentioned second resource change data, respectively, to obtain a first feature corresponding to the above-mentioned first resource change data, and a second feature corresponding to the above-mentioned second resource change data; an integration unit, used to integrate the above-mentioned first feature and the above-mentioned second feature according to the resource interaction data generated in the above-mentioned resource interaction operation, to obtain a target feature; a second acquisition unit, used to obtain the resource allocation coefficient corresponding to the above-mentioned first account based on the above-mentioned target feature; and a determination unit, used to determine the target resource value allocated to the above-mentioned first account according to the above-mentioned resource allocation coefficient.

[0007] As an optional solution, the above-mentioned integration module includes: a first determination submodule, used to determine the above-mentioned first account as the first node, and use the above-mentioned first feature as the node attribute of the above-mentioned first node; a second determination submodule, used to determine the above-mentioned second account as the second node, and use the above-mentioned second feature as the node attribute of the above-mentioned second node; a third determination submodule, used to connect the above-mentioned first node and the above-mentioned second node through the target link, and determine the link weight of the above-mentioned target link based on the above-mentioned resource interaction data; an integration submodule, used to integrate the above-mentioned first node, the above-mentioned second node, and the above-mentioned target link to obtain the above-mentioned target image data.

[0008] As an optional solution, it includes: a third acquisition unit, used to acquire the image data of the above-mentioned multiple sample accounts before acquiring the first resource change data of the first account in the first time period and the second resource change data of the second account in the first time period; a marking unit, used to mark the image data of each of the above-mentioned sample accounts before acquiring the first resource change data of the first account in the first time period and the second resource change data of the second account in the first time period, so as to obtain the marked image data of the above-mentioned multiple sample accounts; an input unit, used to input the marked image data of the above-mentioned multiple sample accounts into the initial graph neural network structure before acquiring the first resource change data of the first account in the first time period and the second resource change data of the second account in the first time period, so as to train the above-mentioned graph neural network structure.

[0009] As an optional solution, the above-mentioned input unit includes: a repetition module, which is used to repeatedly execute the following steps until the above-mentioned graph neural network structure is obtained: a first determination module, which is used to determine the image data of the current sample account from the image data of the above-mentioned multiple sample accounts after labeling, and determine the current graph neural network structure, wherein the image data of the above-mentioned current sample account is marked as image data belonging to the black / white list account; a first acquisition module, which is used to obtain the current output result through the above-mentioned current graph neural network structure, wherein the above-mentioned current output result is used to indicate that the above-mentioned current sample account is an evaluation value of the above-mentioned black / white list account; a second acquisition module, which is used to obtain the image data of the next sample account as the image data of the above-mentioned current sample account when the above-mentioned current output result does not meet the training convergence condition; a second determination module, which is used to determine that the above-mentioned current graph neural network structure is the above-mentioned graph neural network structure when the above-mentioned current output result meets the above-mentioned training convergence condition.

[0010] As an optional solution, the above-mentioned second acquisition unit includes: a second input module, used to input the above-mentioned target feature into the first output structure to obtain the above-mentioned resource allocation coefficient output by the above-mentioned first output structure, wherein the above-mentioned first output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

[0011] As an optional solution, the above-mentioned extraction unit includes: a third input module, used to input the above-mentioned first resource interaction data and the above-mentioned second resource interaction data into the feature extraction structure respectively to obtain the above-mentioned first feature and the above-mentioned second feature output by the above-mentioned feature extraction structure, wherein the above-mentioned feature extraction structure is used to extract the resource interaction features of the input data.

[0012] As an optional solution, the above-mentioned second acquisition unit includes: a fourth input module, used to input the above-mentioned first feature and the above-mentioned target feature into the second output structure to obtain the above-mentioned resource allocation coefficient output by the above-mentioned second output structure, wherein the above-mentioned second output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

[0013] As an optional solution, it includes: a third acquisition module, which is used to obtain the sample target data of the above-mentioned multiple sample accounts before obtaining the first resource change data of the first account in the first time period and the second resource change data of the second account in the first time period, wherein the above-mentioned sample target data includes sample resource change data and sample resource interaction data; a marking module, which is used to mark the sample target data of each of the above-mentioned sample accounts before obtaining the first resource change data of the first account in the first time period and the second resource change data of the second account in the first time period, so as to obtain the marked sample target data of the above-mentioned multiple sample accounts; a fifth input module, which is used to input the marked sample target data of the above-mentioned multiple sample accounts into the initial target neural network structure before obtaining the first resource change data of the first account in the first time period and the second resource change data of the second account in the first time period, so as to train and obtain the above-mentioned target neural network structure, wherein the above-mentioned target neural network structure includes the above-mentioned graph neural network structure, the above-mentioned second output structure, and the above-mentioned feature extraction structure.

[0014] As an optional solution, the above-mentioned determination unit includes: a third determination module, which is used to determine the first resource value corresponding to the above-mentioned first threshold allocated to the above-mentioned first account when the above-mentioned resource allocation coefficient reaches the first threshold; a fourth determination module, which is used to determine the second resource value corresponding to the above-mentioned second threshold allocated to the above-mentioned first account when the above-mentioned resource allocation coefficient reaches the second threshold, wherein the above-mentioned second threshold is greater than the above-mentioned first threshold, and the above-mentioned second resource value is greater than the above-mentioned first resource value.

[0015] As an optional solution, it includes: a receiving unit for receiving a resource allocation request triggered by the first account after the target resource value allocated to the first account is determined according to the resource allocation coefficient, wherein the resource allocation request is used to request allocation of a third resource value to the first account; a responding unit for responding to the resource allocation request after the target resource value allocated to the first account is determined according to the resource allocation coefficient, and allocating the third resource value to the first account if the target resource value is greater than or equal to the third resource value; or a display unit for displaying a prompt message after the target resource value allocated to the first account is determined according to the resource allocation coefficient and if the target resource value is less than the third resource value, wherein the prompt message is used to prompt that the third resource value requested to be allocated by the first account is higher than the resource allocation authority corresponding to the first account.

[0016] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is configured to execute the above-mentioned resource allocation method when running.

[0017] According to another aspect of an embodiment of the present invention, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the resource allocation method through the computer program.

[0018] In an embodiment of the present invention, first resource change data of a first account within a first time period and second resource change data of a second account within the first time period are obtained, wherein the second account is an account that has jointly completed resource interaction operations with the first account within the second time period; feature extraction is performed on the first resource change data and the second resource change data to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data; the first feature and the second feature are integrated based on the resource interaction data generated by the resource interaction operations to obtain a target feature; a resource allocation coefficient corresponding to the first account is obtained based on the target feature; and a target resource value to be allocated to the first account is determined according to the resource allocation coefficient. By applying the principle of "birds of a feather flock together" and combining resource change data of other accounts that have recently jointly completed resource interaction operations with the account to be evaluated during the calculation of the resource allocation coefficient of the account to be evaluated, the technical purpose of expanding the calculation dimension of the resource allocation coefficient is achieved, thereby achieving the technical effect of improving the accuracy of resource allocation and solving the technical problem of low resource allocation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0020] Figure 1 is a schematic diagram of an application environment of an optional resource allocation method according to an embodiment of the present invention;

[0021] Figure 2 is a schematic diagram of a process of an optional resource allocation method according to an embodiment of the present invention;

[0022] Figure 3 is a schematic diagram of an optional resource allocation method according to an embodiment of the present invention;

[0023] Figure 4is a schematic diagram of another optional resource allocation method according to an embodiment of the present invention;

[0024] Figure 5 is a schematic diagram of another optional resource allocation method according to an embodiment of the present invention;

[0025] Figure 6 is a schematic diagram of another optional resource allocation method according to an embodiment of the present invention;

[0026] Figure 7 is a schematic diagram of another optional resource allocation method according to an embodiment of the present invention;

[0027] Figure 8 is a schematic diagram of another optional resource allocation method according to an embodiment of the present invention;

[0028] Figure 9 is a schematic diagram of an optional resource allocation device according to an embodiment of the present invention;

[0029] Figure 10 FIG. 4 is a schematic structural diagram of an optional electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] First, to facilitate understanding of the embodiments of the present invention, some of the terms or nouns involved in the present invention are explained below:

[0033] Graph Convolutional Network (GCN) is a neural network used to process graph data, which draws on the ideas of convolutional neural networks, recurrent networks, and deep autoencoders.

[0034] Logistic regression model: Logistic regression is a generalized linear regression analysis model, which is often used in data mining, automatic disease diagnosis, economic forecasting and other fields.

[0035] Application Card: In credit risk control, when a user applies for a loan for the first time, the customer's credit is assessed. If the user's credit risk is too high, the customer's loan application may be directly rejected.

[0036] Artificial Intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also involves studying the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0037] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0038] Machine learning (ML) is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. It specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of AI. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.

[0039] With the research and advancement of artificial intelligence technology, artificial intelligence technology has been studied and applied in many fields, such as common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned driving, autonomous driving, drones, robots, smart medical care, smart customer service, etc. It is believed that with the development of technology, artificial intelligence technology will be applied in more fields and play an increasingly important role.

[0040] The solutions provided in the embodiments of this application involve technologies such as machine learning based on artificial intelligence, and are specifically described through the following embodiments:

[0041] According to one aspect of an embodiment of the present invention, a resource allocation method is provided. Optionally, as an optional implementation, the resource allocation method can be applied to, but is not limited to, Figure 1 In the environment shown, it may include, but is not limited to, a user device 102, a network 110, and a server 112. The user device 102 may include, but is not limited to, a display 108, a processor 106, and a memory 104.

[0042] The specific process can be as follows:

[0043] In step S102, the user device 102 obtains, on the client, relevant data of the first account serving as the target account, wherein the relevant data includes at least first resource change data of the first account within a first time period and account data of resource interaction operations performed together with the first account within a second time period.

[0044] Steps S104-S106, the user device 102 sends the relevant data to the server 112 via the network 110;

[0045] In steps S108-S112, server 112 processes the relevant data through processing engine 116 to obtain first resource change data of the first account in the first time period and a second account that performed resource interaction operations with the first account in the second time period. Processing engine 116 then searches database 114 for second resource change data of the second account in the first time period. Processing engine 116 further processes the first and second resource change data to generate a resource allocation coefficient corresponding to the first account.

[0046] In steps S114-S116, the server 112 sends the resource allocation coefficient to the user device 102 via the network 110. The user device 102 processes the resource allocation coefficient via the processor 106 to obtain the target resource value allocated to the first account, displays the target resource value on the display 108, and stores the target resource value in the memory 104.

[0047] remove Figure 1 In addition to the examples shown, the above steps can be independently completed by user device 102. That is, user device 102 performs steps such as image processing and determining whether the captured posture is normal or abnormal, thereby reducing the processing pressure on the server. User device 102 includes but is not limited to handheld devices (such as mobile phones), laptops, desktop computers, and in-vehicle devices. The present invention does not limit the specific implementation of user device 102.

[0048] Alternatively, as an optional implementation, as Figure 2 As shown, the resource allocation method includes:

[0049] S202, obtaining first resource change data of a first account within a first time period, and second resource change data of a second account within the first time period, wherein the second account is an account that performs resource interaction operations with the first account within the second time period;

[0050] S204, performing feature extraction on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data;

[0051] S206, integrating the first feature and the second feature according to the resource interaction data generated in the resource interaction operation to obtain a target feature;

[0052] S208, obtaining a resource allocation coefficient corresponding to the first account based on the target feature;

[0053] S210: Determine a target resource value to be allocated to the first account according to the resource allocation coefficient.

[0054] Optionally, in this embodiment, the above-mentioned resource allocation method can be applied in, but is not limited to, the financial field. By studying the transaction flow data of the account to be evaluated, and combining the transaction flow data of other accounts that have recently completed transactions with the account to be evaluated, the transaction performance of the account to be evaluated is comprehensively counted, and based on the above-mentioned comprehensive transaction performance, a more accurate resource allocation coefficient is provided for the account to be evaluated. Compared with the traditional technology of single-dimensional data mining in the financial field, the above-mentioned resource allocation method is more comprehensive and can allocate more accurate target resource values ​​to accounts, so as to reduce the overdue probability of the account returning to the target resource value within expectations.

[0055] Optionally, in this embodiment, the resource change data may be, but is not limited to, the transaction flow data of the account, and the transaction flow data includes at least one of the following: transaction expenditure data, transaction income data, transaction loan data, transaction loan repayment data, etc. For example, if the first account is an account of a catering enterprise, then the transaction flow data may be, but is not limited to, the number of catering industry transactions, the catering industry transaction amount, the maximum catering industry transaction amount, the number of days for catering industry transactions, etc. in the last month (first time period), and the second account may be, but is not limited to, the account with the largest transaction amount completed together with the account of the catering enterprise in the last two months (second time period). For example, if the second account is an account of a food enterprise, then the transaction flow data such as the number of food transaction, food transaction amount, food maximum transaction amount, and food transaction days for the second account in the last month (first time period) are further obtained. There is no limit on the number of second accounts here, and the accounts with the amount (resource interaction data) of transactions completed together with the account of the catering enterprise in the last two months (second time period) can be determined as second accounts, where N is an integer greater than or equal to 1.

[0056] Optionally, in this embodiment, the resource interaction data generated during the resource interaction operation may be, but is not limited to, resource values ​​that change during the resource interaction operation. Optionally, the second account may be, but is not limited to, determined based on the resource values ​​that change during the resource interaction operation. For example, resource values ​​that change during the resource interaction operation are sorted from largest to smallest, and the accounts corresponding to the first N resource values ​​are determined as the second account.

[0057] Optionally, in this embodiment, the resource allocation coefficient can be, but is not limited to, a probability value between 0 and 1. The closer it is to 1, the higher the probability that the first account is overdue. Applied to the loan approval stage, an approval strategy can be formulated based on the score. For example, for low-risk customers, a target resource value with higher authority can be assigned, and if the loan resource value is lower than the target resource value, the application will be directly approved. For medium-risk customers, the application will be manually reviewed by credit review personnel. If the loan resource value is lower than, but close to, the target resource value, the application will be manually reviewed by credit review personnel. For high-risk customers, if the loan resource value is higher than the target resource value, their loan approval will be directly rejected.

[0058] Optionally, in this embodiment, the second account may be, but is not limited to, an account that directly interacts with the first account, for example Figure 3 As shown, within the first time period, account A, which is the first account 302, transfers its resources to account B, and account C transfers its resources to account A. Then, both account B and account C can be used as the second account 404, but are not limited to; in addition, the second account can also include, but is not limited to, an account that has indirect resource interaction operations with the first account, such as Figure 4As shown, the first account 402 is account A, and assuming that within a first time period, account A transfers the resources held by it to account B, and account B transfers the resources held by it to account C, then both account B and account C can be used as the second account 404, but are not limited to;

[0059] For further example, accounts that directly interact with a first account within a first time period are designated as first-degree accounts; accounts that indirectly interact with a first-degree account within a first time period are designated as second-degree accounts; accounts that indirectly interact with a second-degree account within a first time period are designated as third-degree accounts; and accounts that indirectly interact with an N-degree account within a first time period are designated as (N+1)-degree accounts. Alternatively, first-degree accounts and / or second-degree accounts and / or (N+1)-degree accounts can be flexibly designated as second accounts.

[0060] Optionally, in this embodiment, the target feature may be used to characterize, but is not limited to, at least one of the following: the transaction performance of the first account during the first time period, the transaction performance of the second account during the first time period, the association between the first account and the second account, etc. Specifically, if the target feature is used to characterize the association between the first account and the second account, if the transaction performance of the second account during the first time period indicates that the account is at risk, the credit assessment of the first account will also be lowered.

[0061] Optionally, in this embodiment, the resource change data may also be, but is not limited to, resource lending data or resource repayment data, etc. Optionally, the resource lending data or resource repayment data may also be, but is not limited to, the same occurrence object. For example, if the first account is customer A and the second account is customer B, then the resource lending data / resource repayment data (first resource change data) of customer A to financial institution C during the first time period is obtained, and the resource lending data / resource repayment data (second resource change data) of customer B to financial institution C during the first time period is obtained.

[0062] Optionally, in this embodiment, the resource allocation coefficient can be used but is not limited to represent the evaluation result of the first account, and the evaluation result can be used to determine the target resource value allocated to the first account. The target resource value can be used but is not limited to represent the upper limit of the resource value that can be allocated to the first account. For example, if the target resource value allocated to the first account is determined to be 1 million according to the resource allocation coefficient, it can be understood that when the loan amount requested by the first account is less than 1 million, the loan can be granted to the first account; conversely, when the loan amount requested by the first account is more than 1 million, the loan to the first account will be refused.

[0063] It should be noted that the above-mentioned resource allocation method obtains the capital relationship of the first account by constructing a capital relationship network of the first account to be allocated resources, and then incorporates the second account that has a capital relationship with the first account. Based on the idea that birds of a feather flock together, when evaluating the qualifications of the first account, not only the characteristics of the first account itself are utilized, but also the characteristics of the second account that has a capital relationship with the first account are used to evaluate the qualifications of the first account.

[0064] To illustrate further, the optional Figure 5 The human-computer interaction interface 502 shown first Figure 5 As shown in (a), an account evaluation request corresponding to the first account is triggered on the human-computer interaction interface 502, wherein the account evaluation request is used to evaluate the loanable amount of the first account; Figure 5 As shown in (b) , in response to the account evaluation request, the target resource value 504 allocated to the first account is displayed on the human-computer interaction interface 502 .

[0065] Through the embodiments provided by the present application, first resource change data of a first account within a first time period and second resource change data of a second account within the first time period are obtained, wherein the second account is an account that jointly completes resource interaction operations with the first account within the second time period; feature extraction is performed on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data; based on the resource interaction data generated in the resource interaction operation, the first feature and the second feature are integrated and processed to obtain a target feature; based on the target feature, a resource allocation coefficient corresponding to the first account is obtained; according to the resource allocation coefficient, a target resource value allocated to the first account is determined, and by using the method of "birds of a feather flock together", in the process of calculating the resource allocation coefficient of the account to be evaluated, the resource change data of other accounts that have recently completed resource interaction operations with the account to be evaluated are combined, thereby achieving the technical purpose of expanding the calculation dimension of the resource allocation coefficient, thereby achieving the technical effect of improving the accuracy of resource allocation.

[0066] As an optional solution, the first feature and the second feature are integrated and processed based on the resource interaction data generated in the resource interaction operation to obtain the target feature, including:

[0067] S1, integrating the first feature and the second feature according to the resource interaction data generated in the resource interaction operation to obtain target image data;

[0068] S2, input the target image data into the graph neural network structure to obtain the target features output by the graph neural network structure, wherein the graph neural network structure is a neural network structure for recognizing images obtained after training using image data from multiple sample accounts.

[0069] Optionally, in this embodiment, the target image data may refer to, but is not limited to, a set of grayscale values ​​of individual pixels represented by numerical values. Real-world images are generally identified by the intensity and spectrum (color) of light at each point on the image. When converting image information into data information, the image needs to be decomposed into many small areas. These small areas are called pixels, and their grayscale can be represented by a numerical value. For color images, the three primary colors (trichromatic) of red, green, and blue are often used to represent them.

[0070] Optionally, in this embodiment, the hidden layer after convolution of the graph neural network structure is subjected to Softmax to obtain the final output.

[0071] Optionally, in this embodiment, the graph neural network structure can be used for, but is not limited to, deep learning of input graph data. The graph neural network structure can include, but is not limited to, at least one of the following: GCN, Graph Attention Networks (GAT), graphsage, etc., wherein the input graph data can be, but is not limited to, irregular relational data that does not have translation invariance, and the surrounding structure of each node may be unique. This type of structured data instantly renders traditional CNN and RNN ineffective. Assuming that the graph neural network structure is used as the GCN structure, GCN has designed a method for extracting features from input graph data, so that these features can be used to perform node classification, graph classification, and edge prediction on the input graph data, and can also obtain graph embedding.

[0072] In addition, GCN is a first-order local approximation of spectral graph convolution. It is a multi-layer graph convolutional neural network. Each convolution layer only processes first-order neighborhood information. By stacking several convolution layers, multi-order neighborhood information can be transmitted.

[0073] The propagation rule of each convolutional layer is shown in the following formula (1):

[0074]

[0075] Among them, σ() is the activation function, It is the adjacency matrix of the undirected graph G plus the self-connection (that is, each vertex adds an edge to itself), IN is the identity matrix, : is the degree matrix Right now H (l) It is The activation unit matrix of the layer, H 0 =X,W (l) is the parameter matrix of each layer;

[0076] In addition, the number of times the convolution operation is performed can be, but is not limited to, based on the capital chain relationship of the first account. For example, if the second account is a first-degree account, a convolution operation is performed once. For another example, if the second account is an N-degree account, N convolution operations are performed.

[0077] It should be noted that the use of a graph neural network structure to process irregular relational data to clarify the transaction relationship between the first account and the second account has more advantages than the ineffective ordinary neural network structure.

[0078] Through the embodiments provided by the present application, the first feature and the second feature are integrated and processed according to the resource interaction data generated in the resource interaction operation to obtain target image data; the target image data is input into the graph neural network structure to obtain the target features output by the graph neural network structure, wherein the graph neural network structure is a neural network structure for recognizing images obtained after training using image data of multiple sample accounts, which achieves the purpose of accurately outputting features used to express the correlation between accounts, and realizes the effect of improving the output accuracy of the target features.

[0079] As an optional solution, the first feature and the second feature are integrated and processed according to the resource interaction data generated in the resource interaction operation to obtain the target image data, including:

[0080] S1, determining the first account as the first node, and using the first feature as the node attribute of the first node;

[0081] S2, determining the second account as a second node, and using the second feature as a node attribute of the second node;

[0082] S3, connecting the first node and the second node through the target link, and determining the link weight of the target link according to the resource interaction data;

[0083] S4, integrating the first node, the second node, and the target link to obtain target image data.

[0084] Optionally, in this embodiment, the target image data can be used but is not limited to representing the financial relationship network of the first account in the second time period. Taking the resource interaction operation as red envelope transfer as an example, a financial relationship network of the red envelope transfer relationship of the first account in the second time period is constructed. First, the account is used as the node, the red envelope transfer amount is used as the edge weight, and the characteristics corresponding to the resource change data are used as the attributes of each node.

[0085] It should be noted that after constructing the image data of the first account through the capital relationship chain, the topological structure information of the first account in the graph is first obtained. This is information that cannot be obtained by traditional consumption feature aggregation. Further use of the topological structure information of the image data can be used to compare the structural similarities between different accounts in the image data; in addition, in the process of traditional credit development, only the attributes of the first account itself are often considered. After introducing the relevant algorithms of image data, in addition to considering the attributes of the first account itself, the attributes of the second account that has a capital relationship with the first account are also introduced, so that the attributes of the second account can be used to evaluate the first account.

[0086] To illustrate further, the optional Figure 6 As shown, Figure 6 Node 1 corresponding to the first account 602 shown in (a) is determined to be Figure 6 The first node 608 shown in (b) in FIG, and the first feature is used as the node attribute of the first node 608; Figure 6 Nodes 2 and 3 corresponding to the second account 604 shown in (a) are determined as Figure 6 The second node 610 shown in (b) is selected, and the second feature is used as the node attribute of the second node 610; the first node 608 and the second node 610 are connected through the target link 612, and the link weight of the target link 612 is determined according to the resource interaction data; the first node 608, the second node 610, and the target link 612 are integrated to obtain the target image data 606, wherein the target link 612 may also carry, but is not limited to, link information, and the link information may be used, but is not limited to, to indicate the interaction direction of the resource interaction data, such as whether the second account transfers to the first account, or whether the first account transfers to the second account, etc.

[0087] Through the embodiments provided in the present application, the first account is determined as the first node, and the first feature is used as the node attribute of the first node; the second account is determined as the second node, and the second feature is used as the node attribute of the second node; the first node and the second node are connected through the target link, and the link weight of the target link is determined based on the resource interaction data; the first node, the second node, and the target link are integrated to obtain target image data, thereby achieving the purpose of obtaining more comprehensive image data and realizing the effect of improving the comprehensiveness of the image data.

[0088] As an optional solution, before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period, the method includes:

[0089] S1, obtain image data of multiple sample accounts;

[0090] S2, labeling the image data of each sample account to obtain labeled image data of multiple sample accounts;

[0091] S3: Input the labeled image data of multiple sample accounts into the initial graph neural network structure to train the graph neural network structure.

[0092] It should be noted that before the target features are output through the graph neural network structure, it is necessary to obtain image data of multiple sample accounts and label the image data of each sample account to obtain the labeled image data of multiple sample accounts; and input the labeled image data of multiple sample accounts into the initial graph neural network structure to train the graph neural network structure, and then use the graph neural network structure to output the corresponding features.

[0093] Through the embodiments provided in this application, image data of multiple sample accounts are obtained; the image data of each sample account is labeled to obtain the labeled image data of multiple sample accounts; the labeled image data of multiple sample accounts are input into the initial graph neural network structure to train the graph neural network structure, thereby achieving the purpose of efficiently acquiring target features and realizing the effect of improving the output efficiency of target features.

[0094] As an optional solution, the labeled image data of multiple sample accounts is input into the initial graph neural network structure to train the graph neural network structure, including:

[0095] S1, repeat the following steps until the graph neural network structure is obtained:

[0096] S2, determining the image data of the current sample account from the labeled image data of the multiple sample accounts, and determining the current graph neural network structure, wherein the image data of the current sample account is labeled as image data belonging to a blacklist / whitelist account;

[0097] S3, obtain the current output result through the current graph neural network structure, where the current output result is used to indicate the evaluation value of the current sample account as a blacklist / whitelist account;

[0098] S4, if the current output result does not meet the training convergence condition, obtain the image data of the next sample account as the image data of the current sample account;

[0099] S5. When the current output result reaches the training convergence condition, determine that the current graph neural network structure is a graph neural network structure.

[0100] Optionally, in this embodiment, a group of loan applicants can be selected as sample accounts, but is not limited to, and the historical behavior data of each sample account can be used as a tag definition, such as whether each sample account has 30+ overdue payments within 6 months after the loan is borrowed as a tag definition for the blacklist / whitelist account.

[0101] It should be noted that the image data of the current sample account is determined from the image data of multiple labeled sample accounts, and the current graph neural network structure is determined; the current output result is obtained through the current graph neural network structure; when the current output result does not meet the training convergence conditions, the image data of the next sample account is obtained as the image data of the current sample account; when the current output result meets the training convergence conditions, the current graph neural network structure is determined to be a graph neural network structure.

[0102] Through the embodiment provided by the present application, the following steps are repeatedly performed until a graph neural network structure is obtained: the image data of the current sample account is determined from the image data of multiple labeled sample accounts, and the current graph neural network structure is determined, wherein the image data of the current sample account is marked as image data belonging to a black / white list account; the current output result is obtained through the current graph neural network structure, wherein the current output result is used to indicate the evaluation value of the current sample account as a black / white list account; when the current output result does not meet the training convergence condition, the image data of the next sample account is obtained as the image data of the current sample account; when the current output result meets the training convergence condition, the current graph neural network structure is determined to be a graph neural network structure, thereby achieving the purpose of efficiently acquiring the target features and realizing the effect of improving the output efficiency of the target features.

[0103] As an optional solution, obtaining the resource allocation coefficient corresponding to the first account based on the target feature includes:

[0104] The target feature is input into the first output structure to obtain the resource allocation coefficient output by the first output structure, wherein the first output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

[0105] Optionally, in this embodiment, the first output structure may be, but is not limited to, a logistic regression model, wherein logistic regression may be, but is not limited to, a generalized linear regression analysis model, which is commonly used in data mining, automatic disease diagnosis, economic forecasting and other fields.

[0106] It should be noted that the target feature is input into the first output structure to obtain the resource allocation coefficient output by the first output structure, wherein the first output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

[0107] Through the embodiments provided in the present application, the target feature is input into the first output structure to obtain the resource allocation coefficient output by the first output structure, wherein the first output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature, thereby achieving the purpose of efficiently outputting the resource allocation coefficient and realizing the effect of improving the output efficiency of the resource allocation coefficient.

[0108] As an optional solution, feature extraction is performed on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data, including:

[0109] The first resource interaction data and the second resource interaction data are respectively input into the feature extraction structure to obtain the first feature and the second feature output by the feature extraction structure, wherein the feature extraction structure is used to extract the resource interaction feature of the input data.

[0110] Optionally, in this embodiment, the feature extraction structure can be used for, but is not limited to, variable derivation and variable screening, wherein variable derivation mainly adopts a statistical up-aggregation method, and the derivation method is crossed with time to obtain the final derived variable, as shown in the following table (1):

[0111]

[0112] After the variables are derived, the variables are screened. The screening of variables is mainly based on the data shown in the following table (2):

[0113]

[0114] After the variable screening is completed, the remaining variables are put into the logistic regression model for training. Logistic regression is a continuous probability distribution, and its distribution function and density function are shown in the following formula (2):

[0115]

[0116]

[0117] For the binary classification problem, suppose we predict whether the user (the first account) is overdue). Given variables X1, X2, … XN, Y, which are 0 or 1 labels indicating whether the user is overdue, we have the following formula (3):

[0118]

[0119] Furthermore, the optimal solution of the parameters is obtained through gradient descent.

[0120] It should be noted that the first resource interaction data and the second resource interaction data are respectively input into the feature extraction structure to obtain the first feature and the second feature output by the feature extraction structure.

[0121] Through the embodiments provided in the present application, the first resource interaction data and the second resource interaction data are respectively input into the feature extraction structure to obtain the first feature and the second feature output by the feature extraction structure, wherein the feature extraction structure is used to extract the resource interaction features of the input data, thereby achieving the purpose of efficiently extracting the first feature and the second feature, and realizing the effect of improving the feature extraction efficiency.

[0122] As an optional solution, obtaining the resource allocation coefficient corresponding to the first account based on the target feature includes:

[0123] The first feature and the target feature are input into the second output structure to obtain the resource allocation coefficient output by the second output structure, wherein the second output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

[0124] Optionally, in this embodiment, the first output structure may be, but is not limited to, a logistic regression model, wherein logistic regression may be, but is not limited to, a generalized linear regression analysis model, which is commonly used in data mining, automatic disease diagnosis, economic forecasting and other fields.

[0125] It should be noted that the first feature and the target feature are input into the second output structure to obtain the resource allocation coefficient output by the second output structure.

[0126] Through the embodiment provided by the present application, the first feature and the target feature are input into the second output structure to obtain the resource allocation coefficient output by the second output structure, wherein the second output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature, thereby achieving the purpose of efficiently outputting the resource allocation coefficient and realizing the effect of improving the output efficiency of the resource allocation coefficient.

[0127] As an optional solution, before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period, the method includes:

[0128] S1, obtaining sample target data of multiple sample accounts, wherein the sample target data includes sample resource change data and sample resource interaction data;

[0129] S2, labeling the sample target data of each sample account to obtain labeled sample target data of multiple sample accounts;

[0130] S3, inputting the labeled sample target data of the multiple sample accounts into the initial target neural network structure to train and obtain the target neural network structure, wherein the target neural network structure includes a graph neural network structure, a second output structure, and a feature extraction structure.

[0131] Optionally, in this embodiment, the outputs of the feature extraction structure and the graph neural network structure can be added as variables to the original logistic regression (LR) model respectively, but not limited to. In the traditional original scoring card, there are XN variables, so the probability value of the graph neural network structure is added as the variable XN+1 to the logistic regression model, and the overall model is retrained to obtain the optimal solution for the parameters.

[0132] It should be noted that sample target data of multiple sample accounts are obtained, wherein the sample target data include sample resource change data and sample resource interaction data; the sample target data of each sample account are labeled to obtain sample target data of multiple sample accounts after labeling; the sample target data of multiple sample accounts after labeling are input into the initial target neural network structure to train and obtain the target neural network structure, wherein the target neural network structure includes a graph neural network structure, a second output structure, and a feature extraction structure.

[0133] To illustrate further, the optional Figure 7 As shown, sample target data of multiple sample accounts are obtained; the sample target data of each sample account is labeled to obtain sample target data of multiple labeled sample accounts; the sample target data of multiple labeled sample accounts are input into the initial target neural network structure 702, and feature extraction is performed through the feature extraction structure 704, and the extracted resource features are combined and further obtained through the graph neural network structure to obtain corresponding graph features, and then the extracted resource features and graph features are input into the second output structure 708, and the output result of the second output structure 708 is combined with the labeling information to complete the training process of the target neural network structure 702.

[0134] Through the embodiments provided by the present application, sample target data of multiple sample accounts are obtained, wherein the sample target data include sample resource change data and sample resource interaction data; the sample target data of each sample account are labeled to obtain sample target data of multiple sample accounts after labeling; the sample target data of multiple sample accounts after labeling are input into the initial target neural network structure to train and obtain the target neural network structure, wherein the target neural network structure includes a graph neural network structure, a second output structure, and a feature extraction structure, thereby achieving the purpose of training a complete neural network structure and realizing the effect of improving the training completeness of the neural network.

[0135] As an optional solution, determining the target resource value allocated to the first account according to the resource allocation coefficient includes:

[0136] S1, when the resource allocation coefficient reaches a first threshold, determining a first resource value corresponding to the first threshold allocated to the first account;

[0137] S2. When the resource allocation coefficient reaches a second threshold, determine a second resource value corresponding to the second threshold allocated to the first account, wherein the second threshold is greater than the first threshold, and the second resource value is greater than the first resource value.

[0138] It should be noted that, when allocating the resource value corresponding to the resource allocation coefficient to the first account, the resource allocation coefficient is positively correlated with the corresponding resource value.

[0139] Through the embodiments provided in the present application, when the resource allocation coefficient reaches a first threshold, a first resource value corresponding to the first threshold allocated to the first account is determined; when the resource allocation coefficient reaches a second threshold, a second resource value corresponding to the second threshold allocated to the first account is determined, wherein the second threshold is greater than the first threshold, and the second resource value is greater than the first resource value, thereby achieving the purpose of refining the resource allocation method and realizing the effect of improving the resource allocation granularity.

[0140] As an optional solution, after determining the target resource value to be allocated to the first account according to the resource allocation coefficient, the following steps are included:

[0141] S1, receiving a resource allocation request triggered by a first account, wherein the resource allocation request is used to request allocation of a third resource value to the first account;

[0142] S2, responding to the resource allocation request and allocating the third resource value to the first account if the target resource value is greater than or equal to the third resource value; or,

[0143] S3. When the target resource value is less than the third resource value, a prompt message is displayed, wherein the prompt message is used to prompt that the third resource value requested for allocation by the first account is higher than the resource allocation authority corresponding to the first account.

[0144] It should be noted that the resource allocation request triggered by the first account is received; the resource allocation request is responded to, and the third resource value is allocated to the first account when the target resource value is greater than or equal to the third resource value; or, when the target resource value is less than the third resource value, a prompt message is displayed.

[0145] To illustrate further, the optional Figure 8 As shown, a resource allocation request triggered by a first account on the human-computer interaction interface 802 is received, wherein the resource allocation request is used to request allocation of a third resource value of "500,000" to the first account, for example Figure 8 As shown in (a) in the figure; in response to the resource allocation request, since it is detected that the target resource value "200,000" is less than the third resource value "500,000", a prompt message 806 "Rejected" is displayed on the human-computer interaction interface 802 to inform the first account that the loan amount currently applied for is greater than the maximum loan amount it can borrow, for example Figure 8 As shown in (b) in .

[0146] Through the embodiments provided in the present application, a resource allocation request triggered by a first account is received, wherein the resource allocation request is used to request allocation of a third resource value to the first account; the resource allocation request is responded to, and the third resource value is allocated to the first account when the target resource value is greater than or equal to the third resource value; or, when the target resource value is less than the third resource value, a prompt message is displayed, wherein the prompt message is used to prompt that the third resource value requested to be allocated by the first account is higher than the resource allocation authority corresponding to the first account, thereby achieving the purpose of promptly prompting the first account whether the resource allocation request can be responded to successfully, and realizing the effect of improving the timeliness of resource allocation.

[0147] It should be noted that for the aforementioned method embodiments, for simplicity of description, they are all expressed as a series of action combinations. However, those skilled in the art should be aware that the present invention is not limited by the order of the actions described, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0148] According to another aspect of the embodiment of the present invention, a resource allocation device for implementing the above resource allocation method is also provided. Figure 9 As shown, the device includes:

[0149] A first acquiring unit 902 is configured to acquire first resource change data of a first account within a first time period, and second resource change data of a second account within the first time period, wherein the second account is an account that performs resource interaction operations with the first account within the second time period;

[0150] An extraction unit 904 is configured to perform feature extraction on the first resource change data and the second resource change data, respectively, to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data;

[0151] An integration unit 906 is configured to integrate the first feature and the second feature according to the resource interaction data generated in the resource interaction operation to obtain a target feature;

[0152] A second acquiring unit 908 is configured to acquire a resource allocation coefficient corresponding to the first account based on the target feature;

[0153] The determining unit 910 is configured to determine a target resource value to be allocated to the first account according to the resource allocation coefficient.

[0154] Optionally, in this embodiment, the above-mentioned resource allocation method can be applied in, but is not limited to, the financial field. By studying the transaction flow data of the account to be evaluated, and combining the transaction flow data of other accounts that have recently completed transactions with the account to be evaluated, the transaction performance of the account to be evaluated is comprehensively counted, and based on the above-mentioned comprehensive transaction performance, a more accurate resource allocation coefficient is provided for the account to be evaluated. Compared with the traditional technology of single-dimensional data mining in the financial field, the above-mentioned resource allocation method is more comprehensive and can allocate more accurate target resource values ​​to accounts, so as to reduce the overdue probability of the account returning to the target resource value within expectations.

[0155] Optionally, in this embodiment, the resource change data may be, but is not limited to, the transaction flow data of the account, and the transaction flow data includes at least one of the following: transaction expenditure data, transaction income data, transaction loan data, transaction loan repayment data, etc. For example, if the first account is an account of a catering enterprise, then the transaction flow data may be, but is not limited to, the number of catering industry transactions, the catering industry transaction amount, the maximum catering industry transaction amount, the number of days for catering industry transactions, etc. in the last month (first time period), and the second account may be, but is not limited to, the account with the largest transaction amount completed together with the account of the catering enterprise in the last two months (second time period). For example, if the second account is an account of a food enterprise, then the transaction flow data such as the number of food transaction, food transaction amount, food maximum transaction amount, and food transaction days for the second account in the last month (first time period) are further obtained. There is no limit on the number of second accounts here, and the accounts with the amount (resource interaction data) of transactions completed together with the account of the catering enterprise in the last two months (second time period) can be determined as second accounts, where N is an integer greater than or equal to 1.

[0156] Optionally, in this embodiment, the resource interaction data generated during the resource interaction operation may be, but is not limited to, resource values ​​that change during the resource interaction operation. Optionally, the second account may be, but is not limited to, determined based on the resource values ​​that change during the resource interaction operation. For example, resource values ​​that change during the resource interaction operation are sorted from largest to smallest, and the accounts corresponding to the first N resource values ​​are determined as the second account.

[0157] Optionally, in this embodiment, the resource allocation coefficient can be, but is not limited to, a probability value between 0 and 1. The closer it is to 1, the higher the probability that the first account is overdue. Applied to the loan approval stage, an approval strategy can be formulated based on the score. For example, for low-risk customers, a target resource value with higher authority can be assigned, and if the loan resource value is lower than the target resource value, the application will be directly approved. For medium-risk customers, the application will be manually reviewed by credit review personnel. If the loan resource value is lower than, but close to, the target resource value, the application will be manually reviewed by credit review personnel. For high-risk customers, if the loan resource value is higher than the target resource value, their loan approval will be directly rejected.

[0158] Optionally, in this embodiment, the target feature may be used to characterize, but is not limited to, at least one of the following: the transaction performance of the first account during the first time period, the transaction performance of the second account during the first time period, the association between the first account and the second account, etc. Specifically, if the target feature is used to characterize the association between the first account and the second account, if the transaction performance of the second account during the first time period indicates that the account is at risk, the credit assessment of the first account will also be lowered.

[0159] Optionally, in this embodiment, the resource change data may also be, but is not limited to, resource lending data or resource repayment data, etc. Optionally, the resource lending data or resource repayment data may also be, but is not limited to, the same occurrence object. For example, if the first account is customer A and the second account is customer B, then the resource lending data / resource repayment data (first resource change data) of customer A to financial institution C during the first time period is obtained, and the resource lending data / resource repayment data (second resource change data) of customer B to financial institution C during the first time period is obtained.

[0160] Optionally, in this embodiment, the resource allocation coefficient can be used but is not limited to represent the evaluation result of the first account, and the evaluation result can be used to determine the target resource value allocated to the first account. The target resource value can be used but is not limited to represent the upper limit of the resource value that can be allocated to the first account. For example, if the target resource value allocated to the first account is determined to be 1 million according to the resource allocation coefficient, it can be understood that when the loan amount requested by the first account is less than 1 million, the loan can be granted to the first account; conversely, when the loan amount requested by the first account is more than 1 million, the loan to the first account will be refused.

[0161] It should be noted that the above-mentioned resource allocation method obtains the capital relationship of the first account by constructing a capital relationship network of the first account to be allocated resources, and then incorporates the second account that has a capital relationship with the first account. Based on the idea that birds of a feather flock together, when evaluating the qualifications of the first account, not only the characteristics of the first account itself are utilized, but also the characteristics of the second account that has a capital relationship with the first account are used to evaluate the qualifications of the first account.

[0162] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0163] Through the embodiments provided by the present application, first resource change data of a first account within a first time period and second resource change data of a second account within the first time period are obtained, wherein the second account is an account that jointly completes resource interaction operations with the first account within the second time period; feature extraction is performed on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data; based on the resource interaction data generated in the resource interaction operation, the first feature and the second feature are integrated and processed to obtain a target feature; based on the target feature, a resource allocation coefficient corresponding to the first account is obtained; according to the resource allocation coefficient, a target resource value allocated to the first account is determined, and by using the method of "birds of a feather flock together", in the process of calculating the resource allocation coefficient of the account to be evaluated, the resource change data of other accounts that have recently completed resource interaction operations with the account to be evaluated are combined, thereby achieving the technical purpose of expanding the calculation dimension of the resource allocation coefficient, thereby achieving the technical effect of improving the accuracy of resource allocation.

[0164] As an optional solution, the integration unit 906 includes:

[0165] an integration module, configured to integrate the first feature and the second feature according to resource interaction data generated in the resource interaction operation to obtain target image data;

[0166] The first input module is used to input the target image data into the graph neural network structure to obtain the target features output by the graph neural network structure, wherein the graph neural network structure is a neural network structure for recognizing images obtained after training using image data from multiple sample accounts.

[0167] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0168] As an optional solution, integrate modules including:

[0169] A first determining submodule, configured to determine the first account as a first node, and use the first feature as a node attribute of the first node;

[0170] A second determining submodule, configured to determine the second account as a second node, and use the second feature as a node attribute of the second node;

[0171] a third determining submodule, configured to connect the first node and the second node via a target link, and determine a link weight of the target link according to the resource interaction data;

[0172] The integration submodule is used to integrate the first node, the second node, and the target link to obtain target image data.

[0173] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0174] As an optional solution, it includes:

[0175] a third acquiring unit, configured to acquire image data of a plurality of sample accounts before acquiring first resource change data of a first account within a first time period and second resource change data of a second account within the first time period;

[0176] a marking unit, configured to mark the image data of each sample account before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period, to obtain the marked image data of the plurality of sample accounts;

[0177] An input unit is used to input the labeled image data of multiple sample accounts into the initial graph neural network structure before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period to train the graph neural network structure.

[0178] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0179] As an optional solution, the input unit includes:

[0180] The repeating module is used to repeatedly execute the following steps until the graph neural network structure is obtained:

[0181] A first determination module is configured to determine image data of a current sample account from the labeled image data of the plurality of sample accounts and determine a current graph neural network structure, wherein the image data of the current sample account is labeled as image data belonging to a blacklist / whitelist account;

[0182] A first acquisition module is used to obtain a current output result through the current graph neural network structure, wherein the current output result is used to indicate the evaluation value of whether the current sample account is a blacklist / whitelist account;

[0183] The second acquisition module is used to acquire the image data of the next sample account as the image data of the current sample account when the current output result does not meet the training convergence condition;

[0184] The second determination module is used to determine that the current graph neural network structure is a graph neural network structure when the current output result meets the training convergence condition.

[0185] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0186] As an optional solution, the second obtaining unit 908 includes:

[0187] The second input module is used to input the target feature into the first output structure to obtain the resource allocation coefficient output by the first output structure, wherein the first output structure is a linear regression analysis structure used to calculate the trustworthiness of the account corresponding to the input feature.

[0188] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0189] As an optional solution, the extraction unit 904 includes:

[0190] The third input module is used to input the first resource interaction data and the second resource interaction data into the feature extraction structure respectively to obtain the first feature and the second feature output by the feature extraction structure, wherein the feature extraction structure is used to extract the resource interaction feature of the input data.

[0191] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0192] As an optional solution, the second obtaining unit 908 includes:

[0193] The fourth input module is used to input the first feature and the target feature into the second output structure to obtain the resource allocation coefficient output by the second output structure, wherein the second output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

[0194] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0195] As an optional solution, it includes:

[0196] a third acquisition module, configured to acquire sample target data of a plurality of sample accounts before acquiring first resource change data of a first account within a first time period and second resource change data of a second account within the first time period, wherein the sample target data includes sample resource change data and sample resource interaction data;

[0197] a marking module, configured to mark the sample target data of each sample account before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period, to obtain the marked sample target data of the plurality of sample accounts;

[0198] The fifth input module is used to input the sample target data of multiple labeled sample accounts into the initial target neural network structure before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period, so as to train the target neural network structure, wherein the target neural network structure includes a graph neural network structure, a second output structure, and a feature extraction structure.

[0199] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0200] As an optional solution, determine the unit, including:

[0201] A third determining module is configured to determine a first resource value corresponding to the first threshold value allocated to the first account when the number of resource allocation coefficients reaches the first threshold value;

[0202] The fourth determining module is configured to determine a second resource value corresponding to the second threshold allocated to the first account when the number of resource allocation systems reaches the second threshold, wherein the second threshold is greater than the first threshold and the second resource value is greater than the first resource value.

[0203] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0204] As an optional solution, it includes:

[0205] a receiving unit, configured to receive a resource allocation request triggered by the first account after determining a target resource value to be allocated to the first account according to the resource allocation coefficient, wherein the resource allocation request is used to request allocation of a third resource value to the first account;

[0206] a responding unit, configured to respond to the resource allocation request after determining the target resource value to be allocated to the first account according to the resource allocation coefficient, and allocate the third resource value to the first account if the target resource value is greater than or equal to the third resource value; or

[0207] The display unit is used to display a prompt message after determining the target resource value allocated to the first account according to the resource allocation coefficient, if the target resource value is less than the third resource value, wherein the prompt message is used to prompt that the third resource value requested to be allocated by the first account is higher than the resource allocation authority corresponding to the first account.

[0208] For specific embodiments, reference may be made to the examples shown in the above resource allocation method, which will not be described in detail in this example.

[0209] According to another aspect of the embodiment of the present invention, an electronic device for implementing the above resource allocation method is also provided. Figure 10 As shown, the electronic device includes a memory 1002 and a processor 1004. The memory 1002 stores a computer program, and the processor 1004 is configured to execute the steps in any of the above method embodiments through the computer program.

[0210] Optionally, in this embodiment, the electronic device may be located in at least one network device among a plurality of network devices of a computer network.

[0211] Optionally, in this embodiment, the processor may be configured to execute the following steps through a computer program:

[0212] S1, obtaining first resource change data of a first account within a first time period, and second resource change data of a second account within the first time period, wherein the second account is an account that performs resource interaction operations with the first account within the second time period;

[0213] S2, performing feature extraction on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data;

[0214] S3, integrating the first feature and the second feature according to the resource interaction data generated in the resource interaction operation to obtain a target feature;

[0215] S4, obtaining a resource allocation coefficient corresponding to the first account based on the target feature;

[0216] S5. Determine the target resource value allocated to the first account according to the resource allocation coefficient.

[0217] Alternatively, those skilled in the art will appreciate that Figure 10 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (MID), a PAD, or other terminal devices. Figure 10 It does not limit the structure of the above electronic device. For example, the electronic device may also include Figure 10 More or fewer components (such as network interfaces, etc.) as shown in, or with Figure 10 Different configurations shown.

[0218] Among them, the memory 1002 can be used to store software programs and modules, such as the program instructions / modules corresponding to the resource allocation method and device in the embodiment of the present invention. The processor 1004 executes various functional applications and data processing by running the software programs and modules stored in the memory 1002, that is, realizing the above-mentioned resource allocation method. The memory 1002 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 1002 may further include a memory remotely located relative to the processor 1004, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned networks include but are not limited to the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. Among them, the memory 1002 can be used to store, but is not limited to, first resource change data, second resource change data, resource interaction data, resource allocation coefficients, and target resource values ​​and other information. As an example, if Figure 10 As shown, the memory 1002 may include, but is not limited to, the first acquisition unit 902, the extraction unit 904, the integration unit 906, the second acquisition unit 908, and the determination unit 910 in the resource allocation apparatus. Furthermore, the memory 1002 may also include, but is not limited to, other module units in the resource allocation apparatus, which will not be described in detail in this example.

[0219] Optionally, the transmission device 1006 is configured to receive or send data via a network. Specific examples of the network may include a wired network and a wireless network. In one embodiment, the transmission device 1006 includes a network interface controller (NIC), which can be connected to other network devices and a router via a network cable to communicate with the Internet or a local area network. In one embodiment, the transmission device 1006 is a radio frequency (RF) module, which is configured to communicate with the Internet wirelessly.

[0220] In addition, the above-mentioned electronic device also includes: a display 1008 for displaying information such as the above-mentioned first resource change data, second resource change data, resource interaction data, resource allocation coefficient and target resource value; and a connection bus 1010 for connecting the various module components in the above-mentioned electronic device.

[0221] In other embodiments, the terminal device or server may be a node in a distributed system, wherein the distributed system may be a blockchain system, and the blockchain system may be a distributed system formed by connecting multiple nodes via network communication. The nodes may form a peer-to-peer (P2P) network, and any computing device, such as a server, terminal, or other electronic device, may become a node in the blockchain system by joining the peer-to-peer network.

[0222] According to one aspect of the present application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the aforementioned method XX, wherein the computer program is configured to perform the steps of any of the aforementioned method embodiments when executed.

[0223] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0224] S1, obtaining first resource change data of a first account within a first time period, and second resource change data of a second account within the first time period, wherein the second account is an account that performs resource interaction operations with the first account within the second time period;

[0225] S2, performing feature extraction on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data;

[0226] S3, integrating the first feature and the second feature according to the resource interaction data generated in the resource interaction operation to obtain a target feature;

[0227] S4, obtaining a resource allocation coefficient corresponding to the first account based on the target feature;

[0228] S5. Determine the target resource value allocated to the first account according to the resource allocation coefficient.

[0229] Optionally, in this embodiment, a person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing the hardware related to the terminal device through a program, and the program may be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0230] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0231] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above-mentioned computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes a number of instructions for causing one or more computer devices (such as personal computers, servers, or network devices) to execute all or part of the steps of the methods described in various embodiments of the present invention.

[0232] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0233] In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, and can be electrical or other forms.

[0234] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0235] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0236] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A resource allocation method, characterized in that: include: Obtaining first resource change data of a first account within a first time period, and second resource change data of a second account within the first time period, wherein the second account is an account that performs resource interaction operations with the first account within the second time period; performing feature extraction on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data; According to the resource interaction data generated in the resource interaction operation, the first account and the second account are determined as nodes, the resource interaction data is determined as edge weights, and the first feature and the second feature are used as attributes of the node to obtain target image data; Inputting the target image data into a graph neural network structure to obtain target features output by the graph neural network structure, wherein the graph neural network structure is a neural network structure for image recognition obtained by training using image data from multiple sample accounts; Obtaining a resource allocation coefficient corresponding to the first account based on the target feature; A target resource value allocated to the first account is determined according to the resource allocation coefficient.

2. The method according to claim 1, characterized in that The step of determining the first account and the second account as nodes, determining the resource interaction data as edge weights, and using the first feature and the second feature as attributes of the node to obtain target image data includes: Determine the first account as a first node, and use the first feature as a node attribute of the first node; Determine the second account as a second node, and use the second feature as a node attribute of the second node; Connecting the first node and the second node via a target link, and determining a link weight of the target link according to the resource interaction data; The first node, the second node, and the target link are integrated to obtain the target image data.

3. The method according to claim 1, characterized in that Before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period, the method includes: Acquiring image data of the plurality of sample accounts; performing labeling processing on the image data of each of the sample accounts to obtain labeled image data of the plurality of sample accounts; The labeled image data of the multiple sample accounts are input into the initial graph neural network structure to train the graph neural network structure.

4. The method according to claim 3, characterized in that Inputting the labeled image data of the plurality of sample accounts into an initial graph neural network structure to train the graph neural network structure includes: Repeat the following steps until the graph neural network structure is obtained: Determining image data of a current sample account from the labeled image data of the plurality of sample accounts, and determining a current graph neural network structure, wherein the image data of the current sample account is labeled as image data belonging to a blacklist / whitelist account; Obtaining a current output result through the current graph neural network structure, wherein the current output result is used to indicate an evaluation value of the current sample account being a blacklist / whitelist account; If the current output result does not meet the training convergence condition, obtaining image data of the next sample account as the image data of the current sample account; When the current output result meets the training convergence condition, the current graph neural network structure is determined to be the graph neural network structure.

5. The method according to claim 1, wherein The acquiring the resource allocation coefficient corresponding to the first account based on the target feature includes: The target feature is input into a first output structure to obtain the resource allocation coefficient output by the first output structure, wherein the first output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

6. The method according to claim 1, characterized in that The performing feature extraction on the first resource change data and the second resource change data respectively to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data includes: The first resource interaction data and the second resource interaction data are respectively input into a feature extraction structure to obtain the first feature and the second feature output by the feature extraction structure, wherein the feature extraction structure is used to extract resource interaction features of the input data.

7. The method according to claim 6, characterized in that The acquiring the resource allocation coefficient corresponding to the first account based on the target feature includes: The first feature and the target feature are input into a second output structure to obtain the resource allocation coefficient output by the second output structure, wherein the second output structure is a linear regression analysis structure for calculating the trustworthiness of the account corresponding to the input feature.

8. The method according to claim 7, characterized in that Before obtaining the first resource change data of the first account within the first time period and the second resource change data of the second account within the first time period, the method includes: Acquiring sample target data of the plurality of sample accounts, wherein the sample target data includes sample resource change data and sample resource interaction data; performing labeling processing on the sample target data of each of the sample accounts to obtain labeled sample target data of the plurality of sample accounts; The labeled sample target data of the multiple sample accounts are input into the initial target neural network structure to train the target neural network structure, wherein the target neural network structure includes the graph neural network structure, the second output structure, and the feature extraction structure.

9. The method according to any one of claims 1 to 8, characterized in that The determining the target resource value allocated to the first account according to the resource allocation coefficient includes: When the resource allocation coefficient reaches a first threshold, determining a first resource value corresponding to the first threshold allocated to the first account; When the resource allocation coefficient reaches a second threshold, a second resource value corresponding to the second threshold allocated to the first account is determined, wherein the second threshold is greater than the first threshold, and the second resource value is greater than the first resource value.

10. The method according to any one of claims 1 to 8, characterized in that After determining the target resource value allocated to the first account according to the resource allocation coefficient, the method further includes: receiving a resource allocation request triggered by the first account, wherein the resource allocation request is used to request allocation of a third resource value to the first account; responding to the resource allocation request and allocating the third resource value to the first account if the target resource value is greater than or equal to the third resource value; or In a case where the target resource value is less than the third resource value, a prompt message is displayed, wherein the prompt message is used to prompt that the third resource value requested for allocation by the first account is higher than the resource allocation authority corresponding to the first account.

11. A resource allocation device, characterized in that: include: A first acquisition unit is configured to acquire first resource change data of a first account within a first time period, and second resource change data of a second account within the first time period, wherein the second account is an account that performs resource interaction operations with the first account within the second time period; an extraction unit, configured to perform feature extraction on the first resource change data and the second resource change data, respectively, to obtain a first feature corresponding to the first resource change data and a second feature corresponding to the second resource change data; an integration unit, configured to integrate the first feature and the second feature according to the resource interaction data generated in the resource interaction operation to obtain a target feature; a second acquiring unit, configured to acquire a resource allocation coefficient corresponding to the first account based on the target feature; a determining unit, configured to determine a target resource value to be allocated to the first account according to the resource allocation coefficient; The integration unit comprises: an integration module, configured to determine, based on resource interaction data generated in the resource interaction operation, the first account and the second account as nodes, determine the resource interaction data as edge weights, and use the first feature and the second feature as attributes of the node to obtain target image data; The first input module is used to input the target image data into a graph neural network structure to obtain the target features output by the graph neural network structure, wherein the graph neural network structure is a neural network structure for recognizing images obtained after training using image data of multiple sample accounts.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method described in any one of claims 1 to 10 when executed.

13. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 10 through the computer program.

Images