Resource allocation method, system and equipment and storage medium

Abstract

The invention provides a resource allocation method, system and device and a storage medium, and relates to the technical field of computer application. The resource allocation method comprises the following steps: collecting multi-dimensional initial data of a user; constructing a risk control model according to the multi-dimensional initial data of the user; determining an initial resource allocation strategy according to the customer information, the product information and the resource information; acquiring real-time behavior data of the user, and acquiring risk information according to the real-time behavior data of the user and the risk control model; and adjusting the allocable resource quantity in the initial resource allocation strategy according to the risk information. By adopting the method and the device, the intelligent risk control model can be constructed according to the multi-dimensional initial data of the user, so that the allocatable resource quantity is dynamically adjusted based on the real-time behavior data and risk information of the user, and the instantaneity and the accuracy of resource allocation are ensured.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and more specifically, to a resource allocation method, system, device and storage medium. Background Art

[0002] Traditional resource allocation usually adopts a fixed resource allocation method, which cannot flexibly respond to the diverse needs of users and provide accurate resource allocation. For example, in the field of financial technology, various financial service platforms such as banks, payment institutions, and credit cards manage and control user credit limits. Current financial platforms usually adopt a fixed credit limit method. However, this traditional credit limit management method cannot flexibly respond to the diverse needs of users and potential credit risks. Especially when the user's consumption behavior or credit status changes, the existing system often finds it difficult to respond quickly and adjust the credit limit, which leads to the risk of over-credit or under-credit.

[0003] In addition, existing credit management systems often have the following problems: (1) Rigid system architecture: The architecture of many systems is single and centralized. This design lacks flexibility and cannot quickly adapt to changing market demands or user behaviors. With the growth of user numbers and transaction volumes, traditional centralized system architectures face bottlenecks in scalability and response speed, making it difficult to meet the needs of high concurrency and real-time processing. (2) Low modularity: Existing systems often tightly couple credit management with other financial service modules and lack independent credit control modules. This architectural design limits the upgrade and optimization space of the system, resulting in the possibility of large-scale system reconstruction when making local improvements. (3) Insufficient microservice architecture: In modern financial systems, microservice architecture can improve the flexibility and scalability of the system, but many existing credit management systems still rely on a monolithic architecture and cannot support independent upgrades and flexible deployment of modules, affecting the overall performance and development efficiency of the system. Summary of the invention

[0004] In order to solve the problems existing in the above-mentioned prior art or at least part of the problems, the embodiments of the present invention provide a resource allocation method, system, storage medium, device and computer program product, which build an intelligent risk control model according to the user's multi-dimensional initial data, and then dynamically adjust the amount of allocatable resources based on the user's real-time behavior data and risk information, so as to ensure the immediacy and accuracy of resource allocation.

[0005] According to a first aspect of the present invention, an embodiment of the present invention provides a resource allocation method, which includes: collecting multi-dimensional initial data of users; building a risk control model based on the multi-dimensional initial data of users; generating a first tree structure, a second tree structure, and a third tree structure based on customer information, product information, and resource information, respectively; determining an initial resource allocation strategy based on the first tree structure, the second tree structure, and the third tree structure; obtaining real-time behavior data of users, and obtaining risk information based on the real-time behavior data of users and the risk control model; and adjusting the amount of allocatable resources in the initial resource allocation strategy based on the risk information.

[0006] According to the above-mentioned embodiment of the present invention, an intelligent risk control model is constructed based on the user's multi-dimensional initial data to provide accurate risk assessment, and then the amount of allocatable resources is dynamically adjusted based on the user's real-time behavior data, which can quickly respond to user changes and provide accurate allocatable resources.

[0007] In some embodiments of the present invention, constructing a risk control model based on the multidimensional initial data of the user includes: cleaning and converting the multidimensional initial data of the user; obtaining feature data based on the converted data; training an initial model based on the feature data, and performing training and optimization processing to obtain the risk control model.

[0008] In some embodiments of the present invention, the resource allocation method also includes: generating a resource management and control table based on the second tree structure and the third tree structure, the resource management and control table including: a node storage table and a relationship storage table; extracting the resource category, resource allocation amount, and management and control parameters of each node in the second tree structure and the third tree structure; and filling the resource category, resource allocation amount, and management and control parameters into the resource management and control table.

[0009] In some implementations of the present invention, the resource allocation method further includes: generating a customer ID index and a resource category index respectively according to the resource management table.

[0010] According to the above-mentioned implementation mode of the present invention, the client ID index and the resource category index are generated based on the resource management table, so that the query efficiency can be improved.

[0011] According to the second aspect of the present invention, an embodiment of the present invention provides a resource allocation system, which includes: a user data collection module, used to collect users' multi-dimensional initial data and users' real-time behavior data; a model construction module, used to build a risk control model based on the users' multi-dimensional initial data; an initial strategy acquisition module, used to generate a first tree structure, a second tree structure, and a third tree structure according to customer information, product information, and resource information, respectively, and determine an initial resource allocation strategy based on the first tree structure, the second tree structure, and the third tree structure; a resource adjustment module, used to obtain risk information based on the users' real-time behavior data and the risk control model, and adjust the amount of allocatable resources in the initial resource allocation strategy based on the risk information.

[0012] According to the above-mentioned embodiment of the present invention, an intelligent risk control model is constructed based on the user's multi-dimensional initial data to provide accurate risk assessment, and then the amount of allocatable resources is dynamically adjusted based on the user's real-time behavior data, which can quickly respond to user changes and provide accurate allocatable resources.

[0013] In some embodiments of the present invention, constructing a risk control model based on the multidimensional initial data of the user includes: cleaning and converting the multidimensional initial data of the user; obtaining feature data based on the converted data; training an initial model based on the feature data, and performing training and optimization processing to obtain the risk control model.

[0014] In some embodiments of the present invention, the resource allocation system also includes a resource management and control table generation module, which is used to perform the following steps: generate a resource management and control table based on the second tree structure and the third tree structure, the resource management and control table including: a node storage table and a relationship storage table; extract the resource category, resource allocation amount, and management and control parameters of each node in the second tree structure and the third tree structure; and fill the resource category, resource allocation amount, and management and control parameters into the resource management and control table.

[0015] In some implementations of the present invention, the resource allocation system further includes an index module, which is used to generate a customer ID index and a resource category index according to the resource management and control table.

[0016] According to the above-mentioned implementation mode of the present invention, the client ID index and the resource category index are generated based on the resource management table, so that the query efficiency can be improved.

[0017] According to the third aspect of the present invention, an embodiment of the present invention provides a computer-readable storage medium having computer-readable instructions stored thereon. When the computer-readable instructions are executed by a processor, the computer performs the following operations: the operations include the steps included in the resource allocation method described in any one of the above embodiments.

[0018] According to the fourth aspect of the present invention, an embodiment of the present invention provides a computer device including a memory and a processor, wherein the memory is used to store one or more computer-readable instructions, wherein the one or more computer-readable instructions, when executed by the processor, can implement the resource allocation method described in any one of the above embodiments.

[0019] According to a fifth aspect of the present invention, an embodiment of the present invention provides a computer program product including a computer program, and when the computer program is executed by a processor, the resource allocation method as described in any one of the above embodiments is implemented.

[0020] From the above, it can be seen that the resource allocation method, system, storage medium, device and computer program product provided in the embodiments of the present invention build an intelligent risk control model based on the user's multi-dimensional initial data to provide accurate risk assessment, and then dynamically adjust the amount of allocatable resources based on the user's real-time behavior data, so as to quickly respond to user changes and provide accurate allocatable resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a schematic diagram of a flow chart of a resource allocation method according to Embodiment 1 of the present invention;

[0022] Figure 2 is a schematic diagram of a quota control model according to an embodiment of the present invention;

[0023] Figure 3 According to an embodiment of the present invention Figure 2 Schematic diagram of the relationship between the customer credit limits at different levels;

[0024] Figure 4 is a schematic flow chart of a method for generating a scoring table according to an embodiment of the present invention;

[0025] Figure 5 2 is a schematic diagram of the architecture of a resource allocation system according to Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0026] The various aspects of the present invention are described in detail below in conjunction with the accompanying drawings and specific embodiments. Among them, well-known modules, units and their connections, links, communications or operations are not shown or described in detail. In addition, the described features, architectures or functions can be combined in any way in one or more embodiments. It should be understood by those skilled in the art that the various embodiments described below are only used for illustration and are not intended to limit the scope of protection of the present invention. It can also be easily understood that the modules or units or processing methods in each embodiment described herein and shown in the accompanying drawings can be combined and designed in various different configurations.

[0027] The following is a brief description of the terms used in this article.

[0028] Tree structure: A non-linear data structure consisting of a root node and several child nodes, and there may be a hierarchical relationship between the child nodes.

[0029] [Example 1]

[0030] Figure 1 It is a flowchart of the resource allocation method according to embodiment 1 of the present invention.

[0031] like Figure 1 As shown, in Embodiment 1 of the present invention, the resource allocation method may include: step S11, step S12, step S13, step S14, step S15 and step S16. The above steps are described in detail below.

[0032] In step S11, multi-dimensional initial data of the user is collected.

[0033] In an exemplary embodiment, the resource allocation method of the present invention is used to adjust credit limits, etc. Correspondingly, in step S11, the system collects multi-dimensional data of users from multiple sources, including but not limited to: user personal information (such as age, gender, occupation, etc.), transaction records (such as transaction amount, transaction time, transaction frequency, etc.), credit records (such as loan records, repayment status, etc.), social media interaction data (such as likes, comments, shares, etc.), and external data sources (such as data from credit reporting agencies, industry reports, etc.).

[0034] In step S12, a risk control model is constructed according to the multi-dimensional initial data of the user.

[0035] In some embodiments, constructing a risk control model based on the multidimensional initial data of the user includes: cleaning and converting the multidimensional initial data of the user; obtaining feature data based on the converted data; training an initial model based on the feature data, and performing training and optimization processing to obtain the risk control model.

[0036] In an exemplary embodiment, the risk control model is constructed after preprocessing the multi-dimensional initial data of the user. Specifically, the preprocessing operation of the multi-dimensional initial data of the user includes the following steps:

[0037] (1) Data cleaning: remove duplicate data, handle missing values ​​(for example, by interpolation, filling with default values, or deletion), and correct outliers (for example, by setting thresholds, using statistical methods, or using machine learning methods to detect and correct outliers).

[0038] (2) Data conversion: Convert non-numeric data into numerical data (for example, process categorical variables through one-hot encoding, label encoding, etc.), and normalize or standardize the data to eliminate the impact of different dimensions.

[0039] (3) Feature selection: Extract features that are useful for risk control models from raw data. For example, evaluate the predictive power of each feature based on corresponding domain knowledge, statistical methods, or machine learning methods, and select the most discriminative features.

[0040] Through the above preprocessing steps, problems such as missing, abnormal, and inconsistent formats in the collected original data can be solved.

[0041] In a further embodiment, after the data preprocessing is completed, a risk control model is constructed using a machine learning algorithm. Specifically, the risk control model is constructed by the following steps:

[0042] (1) Algorithm selection: Select the appropriate machine learning algorithm based on the specific business problem type and goal to be solved and the characteristics of the data. For example, in the field of risk control, commonly used algorithms include but are not limited to: logistic regression, decision tree, random forest, neural network, etc. Common business problems include but are not limited to:

[0043] a. Binary classification problem. Its goal is to determine whether a certain event will occur, for example: whether a user will default (predict default risk); whether a transaction is fraudulent (detect fraud); whether a credit application is approved (approved / rejected), etc.

[0044] b. Multi-classification problem. Its goal is to classify samples into multiple categories, such as: user risk level classification (low risk, medium risk, high risk); product matching classification (matching product types with different credit limits), etc.

[0045] c. Regression problem. Its goal is to predict continuous variables, such as: the user's default probability (default rate prediction); credit recovery amount or loan amount (amount prediction), etc.

[0046] d. Time series prediction problem. Its goal is to predict the changing trend of risk based on the time dimension, such as: predicting the number of future defaults; changes in a user’s credit score in the next month, etc.

[0047] e. Anomaly detection problem. Its goal is to identify abnormal samples, for example: detecting abnormal behaviors in transactions (such as abnormal withdrawals and transfers); marking abnormal user behavior trajectories (such as large-amount consumption in a short period of time), etc.

[0048] (2) Model training: Divide the preprocessed data into a training set and a test set (which may also include a validation set), and use the training set data to train the machine learning algorithm selected in step (1) to optimize the parameters of the model. During the training process, cross-validation, grid search and other techniques are used to improve the generalization ability of the model.

[0049] (3) Feature Engineering: Feature engineering is performed during the model training process, such as feature construction (creating new features by combining existing features) and feature selection (further screening out the features that are most helpful in improving model performance).

[0050] (4) Model evaluation: Use the test set data to evaluate the trained model to verify its performance in practical applications. Common evaluation indicators include accuracy, recall, F1 value, AUC, ROC curve, etc. Based on the evaluation results, the model can be tuned or other algorithms can be selected for retraining.

[0051] Among them, accuracy refers to the proportion of correctly classified samples to the total number of samples; recall refers to the proportion of samples correctly predicted as positive by the model among all truly positive samples; F1-Score is the harmonic mean of accuracy and recall, which comprehensively considers the effects of accuracy and recall; ROC curve (Receiver Operating Characteristic Curve) is a curve drawn with false positive rate (False Positive Rate, FPR) as the horizontal axis and true positive rate (True Positive Rate, TPR) as the vertical axis. The ROC curve shows the performance of the classifier at different thresholds; AUC (Area Under the Curve) AUC is the area under the ROC curve, and its value range is from 0.5 to 1. The larger the AUC value, the better the performance of the model.

[0052] Through big data and machine learning technology, we use multi-dimensional data such as user consumption, transactions, and credit to build an intelligent risk control model and provide accurate risk assessment. Subsequently, we adjust the credit limit in real time according to the user's credit changes and transaction behavior to avoid the problem of excessive or low credit limits.

[0053] In step S13, a first tree structure, a second tree structure, and a third tree structure are generated respectively according to the customer information, the product information, and the resource information.

[0054] In some implementations, the hierarchical relationship between customers, products, quota categories, etc. and the information of each node in the quota management process are represented by a tree structure.

[0055] Among them, customer hierarchical relationships include: there may be superior-subordinate relationships between customers of financial institutions (including different types such as individuals and enterprises), such as group companies and their subsidiaries. Through the customer tree structure, these hierarchical relationships can be clearly expressed, which is convenient for unified credit control of group customers. Product hierarchical relationships include: there may be inclusion or complementary relationships between the wide variety of products provided by financial institutions. The product tree structure can be used to clearly divide the product hierarchy, for example, loan products are divided into personal loans and corporate loans, and further subdivided into subcategories such as mortgages and car loans. The credit category hierarchical relationship includes: there may be priority or mutually exclusive relationships between the various credit categories in credit control (such as credit limit, credit limit, risk exposure, etc.). The tree structure can easily express the relationship between credit categories and formulate corresponding credit control strategies.

[0056] In step S14, an initial resource allocation strategy is determined according to the first tree structure, the second tree structure, and the third tree structure.

[0057] In an exemplary embodiment, the resource allocation method of the present invention is used to adjust the credit limit, etc., and determine the credit limit allocation strategy according to the first tree structure, the second tree structure, the third tree structure and the credit limit control model. Figure 2 As shown, the hierarchical division structure of customer credit limits includes from first-level customers to fifth-level customers (product level), where each level represents a more detailed credit limit category. Specifically, first-level customers have an overall upper limit, which represents the overall credit limit control of customers; second-level customer credit limit control includes low-risk credit limits and high-risk credit limits, which are used to control credit limit allocations for customers of different risk categories; third-level customer credit limit control is segmented credit limit control, including credit limit types such as mortgage, Internet credit, Internet guarantee, traditional credit, and traditional guarantee; fourth-level customer credit limit control is a specific control item, such as mortgage and pledge in mortgage, and consumer credit and business credit credit allocation / control in Internet credit; fifth-level customer credit limit control is a product-level credit limit, such as credit limits for different financial products such as mortgages and car mortgages. In addition, Figure 2 "Unlimited" for some nodes means that there is no limit on the amount under this node, but it may be controlled by the parent node.

[0058] Figure 2 The relationship between the customer credit limits at different levels is as follows: Figure 3As shown, the root node (A0) represents the upper limit of the total amount of customers, which is controlled by the variable limitA0 (total limit amount of customers); the intermediate nodes (B1, B2, B3) are subdivided into different amount categories, each of which is controlled by its own amount (such as limitB1, limitB2, limitB3); the leaf nodes (C1, C2, ...) represent product-level or more specific amounts, which are controlled by the most detailed amount limitC1, etc., where limitA0, limitB1, limitB2, limitB3, limitC1 ... are adjusted through the resource / quota control table. Among them, the amount limit is carried out step by step from the leaf node to the root node. For example, the amount value of each leaf node (such as C1) must be less than or equal to its own limit (limitC1), and the sum of the leaf amounts (such as C1+C2) meets the amount limit (limitB1) of its parent node (such as B1). Finally, the total amount of the entire tree is controlled by the root node A0. In other words, the sum of the amount of each level is limited by the current level and all upper nodes.

[0059] In step S15, the real-time behavior data of the user is obtained, and risk information is obtained based on the real-time behavior data of the user and the risk control model.

[0060] In step S16, the allocable resource amount in the initial resource allocation strategy is adjusted according to the risk information. In this embodiment, the allocable resources (e.g., allocable quota) in the resource management table are dynamically modified according to the risk information. In the resource management table, the resource allocation strategy for the corresponding customer is updated based on the customer ID index.

[0061] By adopting the above-mentioned resource allocation method of Example 1 of the present invention, an intelligent risk control model is constructed according to the user's multi-dimensional initial data to provide accurate risk assessment, and then the amount of allocatable resources is dynamically adjusted based on the user's real-time behavior data, which can quickly respond to user changes and provide accurate allocatable resources.

[0062] In a further implementation, the resource allocation method further includes: generating a resource management table according to the second tree structure and the third tree structure, the resource management table including: a node storage table and a relationship storage table; extracting the resource category, resource allocation amount, and management parameters of each node in the second tree structure and the third tree structure; and filling the resource category, resource allocation amount, and management parameters into the resource management table. Among them, the corresponding node storage table is generated according to all the node information in the second tree structure; and the relationship storage table is generated according to the second tree structure and the third tree structure.

[0063] The resource control table is used to dynamically update the customer's resource allocation strategy, execute the resource allocation strategy in real time, and monitor and feedback the resource allocation situation. Through the dynamic maintenance and flexible application of the resource control table, accurate and efficient resource allocation management can be achieved, thereby improving resource utilization and reducing allocation risks. For example, when a customer's risk level is adjusted from medium risk to low risk, the following resource allocation adjustment process is triggered by the risk control model:

[0064] (1) In the resource management table, update the customer's resource allocation strategy based on the customer ID index, such as increasing the credit limit.

[0065] (2) Update the resource allocation corresponding to the customer in the node storage table, and modify the weight value associated with the product resource in the relationship storage table to reflect the new allocation logic.

[0066] (3) Generate instructions based on the updated resource allocation strategy, allocate new credit limits to customers, and monitor customer usage in real time.

[0067] In an exemplary embodiment, the amount of resources is adjusted according to the real-time behavior data of the user and the risk information generated by the risk control model, for example, the amount (such as the credit limit) is adjusted, that is, the process of dynamically modifying the amount of allocable resources / amount in the resource control table specifically includes the following steps:

[0068] (1) Trigger adjustment conditions. For example, the adjustment of the credit limit is triggered when the following conditions are met: a. The user's real-time behavior data (such as transaction amount, frequency, and overdue status) changes; b. The risk control model assesses the change in the user's risk level; c. The overall allocatable amount of the resource pool changes (for example, the overall allocatable amount changes due to insufficient system resources or policy adjustments).

[0069] (2) Adjustment calculation: The system calculates the new allocation quota based on the risk level, user behavior, historical records, and resource pool status. For example, the adjustment formula for the allocation quota is as follows: a. When the user's risk level decreases or the usage rate is high, the quota is increased proportionally, that is, the new quota is calculated according to "new quota = current quota + Delta quota", where Delta quota is a variable quota, and the Delta quota = min (benchmark increase \ times weight coefficient, available resource pool remaining amount); b. When the user's risk level increases or the allocated quota is not effectively used, the quota is reduced according to the formula "new quota = current quota - Delta quota", where Delta quota = current quota \ times risk attenuation coefficient.

[0070] (3) Update the resource control table: Update the relevant fields in the resource control table according to the new allocation amount (such as credit limit), including: updating the resource category and allocation amount fields corresponding to the user node in the node storage table; updating the weights between nodes in the relationship storage table according to the quota adjustment results to ensure that the new strategy is reflected in the upstream and downstream resources.

[0071] (4) Real-time push adjustment instructions: Generate adjusted resource allocation instructions and push them to the execution module to update the quota. Specifically, when the quota increases, immediately open a new quota and notify the user; when the quota decreases, limit the user's overuse and prompt the user with the updated quota range.

[0072] (5) Behavior monitoring and backtracking: After adjusting the quota, the user's resource usage is continuously monitored. If the user frequently exceeds the limit or the risk level increases again, a strategy to further reduce the quota is triggered. If the user's quota usage is stable and the behavior is good, the quota is appropriately increased in the next evaluation.

[0073] In a further embodiment, the resource allocation method further includes: generating a customer ID index and a resource category index respectively according to the resource management table. Generating a customer ID index and a resource category index based on the resource management table can improve query efficiency.

[0074] The node storage table is generated by creating a node table and storing all node information in the second tree structure. Optionally, the node storage table includes but is not limited to the following fields: unique identifier of the node, node name, node level, parent node identifier, controlled amount, etc. The controlled amount is the limit amount corresponding to the node, for example, the maximum credit limit corresponding to the node.

[0075] The relationship storage table is generated by creating a relationship table and storing the hierarchical relationship between all nodes in the first tree structure, the second tree structure, and the third tree structure. Optionally, the relationship storage table includes but is not limited to the following fields: child node identifier, parent node identifier, path ID, path code value level, etc. Among them, the path ID is the node ID passed from the bottom node to the top node in the tree structure; the path code value level is the hierarchical code value corresponding to the path ID passed in the tree structure. Based on the relationship storage table, the information of any node and its child nodes can be easily queried.

[0076] Furthermore, the data in the tree structure is migrated to the node storage table and the relationship storage table. Specifically, for each node, a record is inserted into the node storage table, and a record related to its parent node is inserted into the relationship storage table. Furthermore, an index is established on the node storage table and the relationship storage table to quickly locate the target node. Optionally, the table structure is optimized by using the partitioning function provided by the database engine to reduce the query range. Therefore, the query efficiency can be improved by establishing an index, using a suitable database engine, etc.

[0077] In addition, special relationships existing in the tree structure, such as mutually exclusive relationships, priority relationships, etc., are stored. For example, an additional additional information table is created to store the special relationships existing in the tree structure, or an additional field is added to the node storage table to represent the special relationships existing in the tree structure.

[0078] At the same time, the present invention provides an example of constructing a node storage table and a relationship storage table:

[0079] The table structure of the node storage table includes the following information: NodeID: node unique identifier, NodeName: node name, NodeLevel: node level, ParentNodeID: parent node identifier, QuotaValue: node quota value, ControlParameters: control parameters (such as approval restrictions, priority, etc.). By inserting into the node storage table and traversing the tree structure, for each node, extract the basic information of the node (such as NodeID, NodeName, NodeLevel, etc.) and insert it into the node storage table to obtain the node storage table data.

[0080] The table structure of the relationship storage table includes the following information: ChildNodeID: child node identifier, ParentNodeID: parent node identifier, PathID: path identifier, RelationshipType: relationship type (such as hierarchical relationship, mutually exclusive relationship, priority relationship, etc.). The relationship storage table data is obtained by inserting into the relationship storage table, extracting the relationship between the child node and the parent node for each node, determining the type of the relationship (such as hierarchical relationship) and inserting it into the relationship storage table.

[0081] For example, for the following tree structure:

[0082] Root(NodeID=1,QuotaValue=100000)

[0083] ├──Node A(NodeID=2,QuotaValue=50000)

[0084] │ ├──Node A1(NodeID=4,QuotaValue=25000)

[0085] │ └──Node A2(NodeID=5,QuotaValue=25000)

[0086] └──Node B(NodeID=3,QuotaValue=50000)

[0087] The node storage table insertion records obtained according to the above insertion operation are shown in Table 1 below:

[0088] Table 1

[0089]

[0090]

[0091] The relational storage table insertion records obtained according to the above insertion operation are shown in Table 2 below:

[0092] Table 2

[0093] ChildNodeID ParentNodeID PathID RelationshipType 2 1 1-2 Hierarchical 3 1 1-3 Hierarchical 4 2 1-2-4 Hierarchical 5 2 1-2-5 Hierarchical

[0094] In an optional embodiment, the tree structure is converted into a table structure of a quota control table through the following steps:

[0095] (1) Traversing the tree structure: First, traverse the original tree structure, starting from the root node and traversing layer by layer to the leaf nodes.

[0096] (2) Extract information: During the traversal process, extract the quota category, quota value and related control parameters of each node. Among them, the quota value is the quota limit controlled by the node; the related control parameters are auxiliary conditions used to limit or optimize quota allocation, and the related control parameters include but are not limited to one or more of the following: priority (priority), quota type (quota_type), validity_period. Among them, the priority is used to determine the resource allocation priority of different nodes. When resources are insufficient, high-priority nodes are given priority; the quota type is used to set whether the node has upper and lower limit constraints (such as the maximum and minimum values ​​when dynamically adjusting the quota) to prevent abnormal changes in the quota due to real-time data fluctuations; the validity period is used to set the effective time of the quota, so as to automatically recycle resources that have exceeded the validity period and reallocate them to other nodes.

[0097] (3) Fill in the table structure: Fill the extracted information into the quota management table (node ​​storage table and relationship storage table) to generate one or more records for each customer. The field value in each record corresponds to the node information in the tree structure. The converted table structure can completely save all the information in the tree structure, including the hierarchical relationship between nodes, attribute information, etc., to ensure data integrity and consistency between the tree structure and the table structure, avoiding data loss or duplication.

[0098] (4) Create indexes: Create indexes for the credit limit control table, such as customer ID index, credit limit category index, etc., to improve query efficiency.

[0099] Through the above steps, the tree structure is converted into the table structure of the quota management table, making the management and query of quota information more efficient and flexible. In addition, the table structure has good scalability, which is convenient for adding new nodes or modifying the properties of existing nodes.

[0100] In some embodiments, the resource allocation method further includes: obtaining customer application information, and obtaining customer-related data based on the customer application information; generating a scoring table based on the customer-related data, wherein the scoring table includes the resource allocation amounts in the resource management and control table.

[0101] At the same time, the present invention provides a specific example of a scoring table generation method. Figure 4 For details, please refer to the following sections. Figure 4 As shown, the scoring table generation method specifically includes the following steps:

[0102] (1) Customer application submission.

[0103] (2) Internal data warehouse processing. Customer application information is transferred to the internal data warehouse, which extracts customer-related information from multiple data sources. By combining information from the internal data warehouse and external data sources, a comprehensive analysis of customer applications can be performed, improving the accuracy and personalization of scoring.

[0104] (3) Model Score analysis: The extracted customer-related information is integrated into the model score. The model score is based on the LangChain big model framework and can comprehensively analyze all aspects of the customer's application. The use of the LangChain big model framework enables the model score to handle complex data analysis tasks and provides strong technical support for the generation of the score sheet.

[0105] (4) Model Score generates a score sheet through analysis. The score sheet includes various scoring information of the customer, such as but not limited to: risk information obtained based on the user's real-time behavior data and risk control model, and credit limit in the credit limit control table.

[0106] By generating a scoring sheet, not only is a detailed credit status assessment provided to customers, but it also provides an important reference for determining loan needs.

[0107] [Example 2]

[0108] Figure 5 : is a schematic diagram of the architecture of the resource allocation system according to Example 2 of the present invention. The resource allocation system of Example 2 of the present invention adopts a microservice architecture, so that the system has high scalability and flexibility, can quickly respond to user needs and market changes, and the system of Example 2 improves the scalability and maintainability of the system through modular design, supports high concurrent processing and flexible function expansion.

[0109] like Figure 5 As shown, the resource allocation system includes: a user data collection module 210 , a model construction module 220 , an initial policy acquisition module 230 , a resource adjustment module 240 , a resource management table generation module 250 , and an index module 260 .

[0110] The user data collection module 210 is used to collect the user's multi-dimensional initial data and the user's real-time behavior data. The user data collection module is also used to perform pre-processing operations such as cleaning, conversion, and feature selection on the collected data. Furthermore, the pre-processing operation of the data also includes: by embedding the compliance review process and data security protection mechanism, ensuring that the system meets regulatory requirements and guarantees the security of user data.

[0111] In an exemplary embodiment, the resource allocation system of the present invention is applied to a financial platform for quota management. Correspondingly, the user data collection module collects multi-dimensional data of users from multiple sources, including but not limited to: user personal information (such as age, gender, occupation, etc.), transaction records (such as transaction amount, transaction time, transaction frequency, etc.), credit records (such as loan records, repayment status, etc.), social media interaction data (such as likes, comments, shares, etc.), and external data sources (such as data from credit reporting agencies, industry reports, etc.).

[0112] The model building module 220 is used to build a risk control model according to the multi-dimensional initial data of the user.

[0113] In some embodiments, constructing a risk control model based on the multidimensional initial data of the user includes: cleaning and converting the multidimensional initial data of the user; obtaining feature data based on the converted data; training an initial model based on the feature data, and performing training and optimization processing to obtain the risk control model.

[0114] The initial strategy acquisition module 230 is used to generate a first tree structure, a second tree structure, and a third tree structure according to customer information, product information, and resource information, respectively, and determine an initial resource allocation strategy according to the first tree structure, the second tree structure, and the third tree structure.

[0115] The resource adjustment module 240 is used to obtain risk information according to the real-time behavior data of the user and the risk control model, and adjust the amount of allocatable resources in the initial resource allocation strategy according to the risk information.

[0116] The resource control table generation module 250 is used to perform the following steps: generate a resource control table according to the second tree structure and the third tree structure, the resource control table including: a node storage table and a relationship storage table; extract the resource category, resource allocation amount, and control parameters of each node in the second tree structure and the third tree structure; and fill the resource category, resource allocation amount, and control parameters into the resource control table.

[0117] The index module 260 is used to generate a customer ID index and a resource category index according to the resource management table.

[0118] In some embodiments, the resource allocation system also includes a scoring table generation module for performing the following operations: obtaining customer application information, and obtaining customer-related data based on the customer application information; generating a scoring table based on the customer-related data, the scoring table including the resource allocation amounts in the resource management table.

[0119] The resource allocation system of Example 2 of the present invention is used to build an intelligent risk control model based on the user's multi-dimensional initial data to provide accurate risk assessment, and then dynamically adjust the amount of allocatable resources based on the user's real-time behavior data, so as to quickly respond to user changes and provide accurate allocatable resources.

[0120] Applying the above resource allocation system to the financial platform for credit limit management can utilize multi-dimensional data such as user consumption, transactions, and credit to build an intelligent risk control model, provide accurate risk assessment, and dynamically adjust credit limits based on users' real-time transaction behaviors or risk changes. This can ensure the immediacy and accuracy of credit limit management and realize event-driven credit limit management.

[0121] Through the description of the above implementation modes, those skilled in the art can clearly understand that the present invention can be implemented by combining software with a hardware platform. Based on such an understanding, all or part of the contribution of the technical solution of the present invention to the background technology can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation modes of the present invention or some parts of the implementation modes.

[0122] Correspondingly, an embodiment of the present invention further provides a computer-readable storage medium, on which computer-readable instructions or programs are stored, and when the computer-readable instructions or programs are executed by a processor, the computer performs the following operations: the operations include the steps included in the resource allocation method described in any of the above embodiments, which are not repeated here. The storage medium may include, for example, a CD, a hard disk, a floppy disk, a flash memory, a magnetic tape, etc.

[0123] In addition, an embodiment of the present invention further provides a computer device including a memory and a processor, wherein the memory is used to store one or more computer-readable instructions or programs, wherein the one or more computer-readable instructions or programs can implement the resource allocation method described in any of the above embodiments when executed by the processor. The computer device can be, for example, a server, a desktop computer, a notebook computer, a tablet computer, etc.

[0124] The embodiment of the present invention also provides a computer program product including a computer program, wherein the computer program includes a program code for executing the resource allocation method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the resource allocation method provided by the embodiment of the present disclosure.

[0125] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention is described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention. Therefore, the protection scope of the present invention shall be subject to the claims.

Claims

1. A resource allocation method, characterized in that: The resource allocation method comprises: Collect multi-dimensional initial data of users; Building a risk control model based on the multi-dimensional initial data of the user; Generate a first tree structure, a second tree structure, and a third tree structure respectively according to the customer information, the product information, and the resource information; Determine an initial resource allocation strategy according to the first tree structure, the second tree structure, and the third tree structure; Obtaining real-time behavior data of users, and obtaining risk information based on the real-time behavior data of users and risk control models; The amount of allocatable resources in the initial resource allocation strategy is adjusted according to the risk information.

2. The resource allocation method according to claim 1, characterized in that: Constructing a risk control model based on the multi-dimensional initial data of the user includes: Cleaning and converting the multi-dimensional initial data of the user; Acquire feature data according to the converted data; The risk control model is obtained by training and optimizing the initial model according to the characteristic data.

3. The resource allocation method according to claim 1, characterized in that: The resource allocation method further includes: Generate a resource management table according to the second tree structure and the third tree structure, wherein the resource management table includes: a node storage table and a relationship storage table; Extracting resource categories, resource allocation amounts, and control parameters of each node in the second tree structure and the third tree structure; The resource category, resource allocation amount, and management and control parameters are filled into the resource management and control table.

4. The resource allocation method according to claim 3, characterized in that: The resource allocation method further includes: generating a customer ID index and a resource category index respectively according to the resource management table.

5. A resource allocation system, characterized in that: The resource allocation system comprises: User data collection module, used to collect multi-dimensional initial data of users and real-time behavior data of users; A model building module, used to build a risk control model based on the multi-dimensional initial data of the user; An initial strategy acquisition module, used to generate a first tree structure, a second tree structure, and a third tree structure according to customer information, product information, and resource information, and determine an initial resource allocation strategy according to the first tree structure, the second tree structure, and the third tree structure; The resource adjustment module is used to obtain risk information based on the real-time behavior data of the user and the risk control model, and adjust the amount of allocatable resources in the initial resource allocation strategy based on the risk information.

6. The resource allocation system according to claim 5, characterized in that: The resource allocation system further includes a resource management table generation module, which is used to perform the following steps: Generate a resource management table according to the second tree structure and the third tree structure, wherein the resource management table includes: a node storage table and a relationship storage table; Extracting resource categories, resource allocation amounts, and control parameters of each node in the second tree structure and the third tree structure; The resource category, resource allocation amount, and management and control parameters are filled into the resource management and control table.

7. The resource allocation system according to claim 6, characterized in that: The resource allocation system further includes an index module, which is used to generate a customer ID index and a resource category index according to the resource management and control table.

8. A computer-readable storage medium storing computer-readable instructions, characterized in that: The computer-readable instructions are executed by a processor to implement the resource allocation method according to any one of claims 1 to 4.

9. A computer device comprising a memory and a processor, The memory stores computer-readable instructions, characterized in that: The processor executes the computer-readable instructions to implement the resource allocation method according to any one of claims 1 to 4.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the resource allocation method according to any one of claims 1 to 4 is implemented.

Images