Multi-risk control strategy parallel backtracking method, electronic equipment, storage medium and product

Abstract

The invention discloses a multi-risk-control-strategy parallel backtracking method, electronic equipment, a storage medium and a product, and relates to the technical field of financial risk control, and the multi-risk-control-strategy parallel backtracking method comprises the steps: determining a plurality of sample layers and a plurality of sample groups contained in each sample layer; multiple groups of financial sample data are distributed to each sample layer and each sample group, and each sample layer comprises full-amount financial sample data; associating a plurality of preset financial risk control strategy models with the sample layers and the sample groups; and inputting the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model, and outputting a backtracking result corresponding to each financial risk control strategy model in parallel. The risk control strategy backtracking efficiency is improved, the defect of mutual interference between strategies during backtracking of multiple risk control strategies is effectively overcome, and the assessment precision of the risk control model and the risk control strategy is improved.

Description

Technical Field

[0001] The present application relates to the field of financial risk control technology, and in particular to a multi-risk control strategy parallel backtracking method, electronic equipment, computer-readable storage medium, and computer program product. Background Art

[0002] In the field of financial risk control, financial risk control strategy models are often used when conducting risk control assessments for users. To improve performance and forecasting accuracy, financial risk control strategy models require continuous iterative optimization and updating. Risk control strategy backtesting is a common method for financial risk control strategy models. This involves simulating and verifying the past performance of risk control strategies (or models) using historical data to assess their effectiveness, stability, and actual business impact, providing data support for strategy optimization.

[0003] Current backtracking solutions typically use a single strategy for backtracking analysis when processing large amounts of data. Single-strategy backtracking requires processing data samples one by one, which is time-consuming. Especially with large data volumes, the backtracking process can take hours or even days. Furthermore, due to the limitations of serial processing, system resource utilization is low, leading to wasted computing resources. When backtracking multiple strategies simultaneously, data samples may overlap between different strategies, disrupting model evaluation results and affecting the independence and effectiveness of the strategies.

[0004] The above content is only used to assist in understanding the technical solution of this application and does not constitute an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a parallel backtracking method for multiple risk control strategies, an electronic device, a computer-readable storage medium, and a computer program product, aiming to solve the technical problem that when multiple risk control strategies are backtracked, there is mutual interference between strategies, which affects the evaluation accuracy of the risk control model.

[0006] To achieve the above objectives, the present application proposes a multi-risk control strategy parallel backtracking method, which includes:

[0007] determining a plurality of sample layers and a plurality of sample groups contained in each sample layer;

[0008] Allocating multiple sets of financial sample data to each of the sample layers and each of the sample groups, wherein each sample layer includes a full amount of financial sample data, and the financial sample data includes a financial user's behavior log, transaction record, and / or historical risk events;

[0009] Associating a plurality of preset financial risk control strategy models with each of the sample layers and sample groups, respectively, wherein each financial risk control strategy model is associated with at least two sample groups in one sample layer;

[0010] The financial sample data of each sample group in each sample layer are respectively input into the corresponding financial risk control strategy model, and the backtracking results corresponding to each financial risk control strategy model are output in parallel.

[0011] In one embodiment, the step of determining multiple sample layers and multiple sample groups contained in each sample layer includes:

[0012] Generate multiple sample layers according to the preset number of sample layers;

[0013] According to the preset number of sample groups, multiple sample groups are generated in each sample layer.

[0014] In one embodiment, the step of allocating the plurality of groups of financial sample data to the sample layers and the sample groups includes:

[0015] Adding each set of financial sample data to each sample layer;

[0016] Each group of financial sample data in each sample layer is respectively allocated to a plurality of sample groups, wherein, in the same sample layer, the financial sample data in different sample groups are mutually exclusive.

[0017] In one embodiment, the step of associating the preset multiple financial risk control strategy models with each of the sample layers and sample groups includes:

[0018] According to the correlation between the financial risk control strategy models, the financial risk control strategy models are divided into multiple groups, wherein each group of financial risk control strategy models includes one financial risk control strategy model or multiple financial risk control strategy models that are correlated with each other;

[0019] Associating each group of financial risk control strategy models with each of the sample layers in a one-to-one correspondence;

[0020] Each financial risk control strategy model in each group of financial risk control strategy models is associated with at least two sample groups in the corresponding sample layer, wherein each financial risk control strategy model in the same group is associated with different sample groups.

[0021] In one embodiment, the step of inputting the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model and outputting the backtracking results corresponding to each financial risk control strategy model in parallel includes:

[0022] Generate multiple backtracking tasks, where each backtracking task corresponds to a financial risk control strategy model;

[0023] Each of the backtracking tasks is assigned to a corresponding computing node, and each of the computing nodes executes the backtracking task in parallel according to the financial sample data in the sample group associated with each of the financial risk control strategy models, and outputs the backtracking results corresponding to each financial risk control strategy model.

[0024] In one embodiment, the sample group associated with the financial risk control strategy model includes at least an experimental group;

[0025] The step of executing the backtracking task in parallel based on the financial sample data in the sample groups associated with each of the financial risk control strategy models and outputting the backtracking results corresponding to each of the financial risk control strategy models comprises:

[0026] Reading financial sample data from the sample group in the sample layer corresponding to each financial risk control strategy model based on each backtracking task;

[0027] Inputting the financial sample data from the experimental group into the corresponding financial risk control strategy model, processing it through the financial risk control strategy model, and obtaining the backtracking results corresponding to each of the financial risk control strategy models;

[0028] Output each of the backtracking results to a preset file system or database.

[0029] In one embodiment, the sample group associated with the financial risk control strategy model also includes a control group;

[0030] After the steps of inputting the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model and outputting the backtracking results corresponding to each financial risk control strategy model in parallel, the method further includes:

[0031] Calculating performance evaluation indicators corresponding to the financial risk control strategy models based on the backtracking results corresponding to the financial risk control strategy models, wherein the performance evaluation indicators include at least one of precision and recall;

[0032] Evaluate the risk control effects of each of the financial risk control strategy models based on the backtracking results corresponding to each of the financial risk control strategy models and the financial sample data of the control group;

[0033] Obtaining data processing time and resource usage information corresponding to each of the backtracking tasks;

[0034] A backtracking report corresponding to each of the financial risk control strategy models is generated based on the performance evaluation indicators, the risk control effect, the data processing time and the resource usage information corresponding to each of the financial risk control strategy models.

[0035] In addition, to achieve the above objectives, the present application also proposes a multi-risk control strategy parallel backtracking device, which includes:

[0036] A sample layer determination module, used to determine multiple sample layers and multiple sample groups contained in each sample layer;

[0037] a sample allocation module, configured to allocate multiple sets of financial sample data to each of the sample layers and each of the sample groups, wherein each sample layer includes the full amount of financial sample data, and the financial sample data includes the behavior logs, transaction records and / or historical risk events of financial users;

[0038] A model association module, configured to associate a plurality of preset financial risk control strategy models with each of the sample layers and sample groups, wherein each financial risk control strategy model is associated with at least two sample groups in a sample layer;

[0039] The model backtracking module is used to input the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model, and output the backtracking results corresponding to each financial risk control strategy model in parallel.

[0040] In addition, to achieve the above-mentioned purpose, the present application also proposes an electronic device, which includes: a memory, a processor, and a computer program stored on the memory and runnable on the processor, wherein the computer program is configured to implement the steps of the multi-risk control strategy parallel backtracking method as described above.

[0041] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the multi-risk control strategy parallel backtracking method as described above are implemented.

[0042] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the multi-risk control strategy parallel backtracking method as described above.

[0043] The present application proposes a multi-risk control strategy parallel backtracking method. In the multi-risk control strategy parallel backtracking method, multiple sample layers and multiple sample groups contained in each sample layer are first determined, and then multiple groups of financial sample data are allocated to each of the sample layers and each of the sample groups, wherein each sample layer includes the full amount of financial sample data, and the financial sample data includes the behavior logs, transaction records and / or historical risk events of financial users, and then multiple preset financial risk control strategy models are associated with each of the sample layers and sample groups, respectively, wherein each financial risk control strategy model is associated with at least two sample groups in a sample layer, and finally, the financial sample data of each sample group in each of the sample layers are respectively input into the corresponding financial risk control strategy model, and the backtracking results corresponding to each of the financial risk control strategy models are output in parallel. This application uses a sample stratification and grouping method that includes all samples to stratify and group samples, and associates them with each risk control model one by one. The financial sample data of each sample layer are independent of each other, so the backtracking process of the associated risk control model is also independent, avoiding the mutual interference caused by backtracking multiple risk control models at the same time, ensuring that the effects of the risk control models can be independently evaluated when multiple risk control strategies are backtracked in parallel, and ensuring the accuracy and independence of each backtracking result. Moreover, in this application, multiple risk control models are backtracked simultaneously, which significantly reduces the backtracking time. The parallel processing method also improves the utilization rate of system resources and reduces the waste of computing resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0046] Figure 1 A flowchart illustrating the first embodiment of the multi-risk control strategy parallel backtracking method of this application;

[0047] Figure 2 This is a schematic diagram of two financial risk control strategy models that affect each other and are associated with sample groups in an embodiment of the present application;

[0048] Figure 3 This is a schematic diagram of allocating all samples to sample layers and sample groups in an embodiment of the present application;

[0049] Figure 4 This is a schematic diagram of the structure of a multi-risk control strategy parallel backtracking device in an embodiment of the present application;

[0050] Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the multi-risk control strategy parallel backtracking method in the embodiment of the present application.

[0051] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0052] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0053] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0054] The execution subject of this embodiment can be a computing service device with data processing, network communication, and program execution capabilities, such as a tablet computer, personal computer, mobile phone, server, etc., or an electronic device or control device capable of implementing the above functions. This embodiment and the following embodiments will be described below using a computer as the execution subject.

[0055] In the field of financial risk control, risk backtesting of financial risk control strategy models involves using historical financial sample data combined with the model to predict past user performance in order to assess its effectiveness, stability, and actual business impact. Its core goal is to determine whether the model can achieve the expected risk control effect in real-world scenarios and provide data support for risk control strategy optimization.

[0056] This application embodiment provides a multi-risk control strategy parallel backtracking method, referring to Figure 1 , Figure 1 This is a flowchart of the first embodiment of the multi-risk control strategy parallel backtracking method of the present application. The multi-risk control strategy parallel backtracking method applied to the local end includes:

[0057] Step S10, determining multiple sample layers and multiple sample groups contained in each sample layer;

[0058] In order to realize the parallel backtracking of multiple risk control strategies in the embodiment of the present application, hardware and software environment preparations need to be carried out in advance before executing step S10. It should be noted that sufficient computing resources are ensured, including multiple servers, storage devices and network devices. Install and configure the big data processing framework database management system and necessary development tools. Before obtaining sample data, design a data storage solution to ensure efficient reading and writing of data. Distributed file systems and distributed databases can be used to store large-scale data. You can also select and configure a big data computing framework to ensure that multiple tasks can be executed in batches. In addition, it is necessary to pre-design and implement multiple random ID (Identity) generators to ensure that the ID generated by each generator is independent and highly random, so as to be used to mark sample layers, sample groups and each sample data.

[0059] In addition, before step S10, data collection is required. For example, financial sample data (such as financial users' behavior logs, transaction records, historical risk events, etc.) must be collected from various data sources. The integrity and accuracy of the data must also be ensured, and necessary data cleaning and preprocessing must be performed. The collected financial sample data is then stored in a distributed file system or database to ensure high data availability and scalability. Finally, a database management system can be used to index and shard the financial sample data to improve data query efficiency.

[0060] After completing the aforementioned preparatory work, we first determine multiple sample layers and the multiple sample groups contained in each sample layer. These sample layers are orthogonal, independent of each other and non-interfering. Furthermore, each sample layer includes multiple sample groups, which are used to group financial sample data for risk control strategy retrospective analysis. It should be noted that within the same sample layer, the sample groups are mutually exclusive and do not contain the same sample data.

[0061] Among them, when determining the sample layers and sample groups, the number of sample layers and sample groups can be set according to needs, and each sample layer and sample group can be given an independent ID (i.e., identifier), so as to facilitate the subsequent correspondence between the sample layers, sample groups and samples, and with the financial risk control strategy model.

[0062] In a feasible embodiment, the sample layer may be stored in the server in the form of a folder, a table, or structured data, and the sample group may also be stored in the server in the form of a folder, a table, or structured data.

[0063] Step S20: Allocate multiple sets of financial sample data to sample layers and sample groups, wherein each sample layer includes the full amount of financial sample data, and the financial sample data includes the financial user's behavior log, transaction record and / or historical risk events;

[0064] When assigning each group of collected financial sample data to each sample stratum, it is necessary to ensure that each sample stratum includes the full set of financial sample data. In other words, each sample stratum includes all collected financial sample data. Multiple sample groups within the same sample stratum each include all collected financial sample data. Some of the financial sample data within a sample stratum belongs to mutually exclusive sample groups.

[0065] The distributed financial sample data is stored in a distributed file system. This allows for the dispersed storage of financial sample data across multiple physical servers (nodes), but presents it as a single logical file system to users responsible for executing multi-risk control strategy backtesting tasks. This provides high-capacity, high-throughput storage services while reducing storage pressure on individual servers.

[0066] In another feasible embodiment, the financial sample data may be text data or image data.

[0067] Step S30: Associating multiple preset financial risk control strategy models with each sample layer and sample group, wherein each financial risk control strategy model is associated with at least two sample groups in one sample layer;

[0068] Before formally conducting strategy backtracking, each sample layer and sample group is first assigned to each financial risk control strategy model. The financial risk control strategy model is a pre-established model used to predict the user's risk level or risk control result (such as pass or fail) based on the user's financial sample data. Different financial risk control strategy models have different functions and can be applied in different scenarios. Generally, each financial risk control strategy model corresponds to a sample layer. However, there may be situations where the strategies of some financial risk control strategy models affect each other. In this case, multiple related financial risk control strategy models (referring to models whose strategies affect each other) need to be assigned to the same sample layer and assigned to different sample groups to ensure that the applied sample data are mutually exclusive and do not interfere with each other.

[0069] In addition, in order to accurately evaluate the model effect of the financial risk control strategy model, when associating the financial risk control strategy model with the sample group, it is necessary to associate at least two sample groups, namely at least one experimental group and at least one control group. The experimental group can be calculated and processed by the financial risk control strategy model to output the corresponding prediction results, while the control group does not need to be processed and is used to compare and evaluate the risk control effect of the financial risk control strategy model.

[0070] Step S40: input the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model, and output the backtracking results corresponding to each financial risk control strategy model in parallel.

[0071] Finally, the financial sample data from the experimental group within each sample group within the corresponding sample layer, associated with each financial risk control strategy model, is input into the financial risk control strategy model. Parallel computation and processing are performed by the financial risk control strategy model to obtain backtesting results corresponding to the multiple sets of financial sample data. The backtesting results output by each financial risk control strategy model include the user assessment results corresponding to the multiple sets of financial sample data contained in the corresponding sample group (experimental group). The backtesting results can be either "pass" or "fail," can be simplified to 1 or 0, or can represent the user's risk level, without limitation.

[0072] These backtracking results reflect the ability of each financial risk control strategy model to conduct risk assessment on users based on the collected financial sample data. By comparing these assessment results with the real labels of each user, the performance of the financial risk control strategy model can be determined, which facilitates the subsequent optimization of the parameters of the financial risk control strategy model or the corresponding risk control rules, thereby improving the prediction accuracy and risk control effect.

[0073] Furthermore, in a feasible implementation manner, the step of determining multiple sample layers and multiple sample groups included in each sample layer may include:

[0074] Step S11, generating multiple sample layers according to a preset number of sample layers;

[0075] The number of corresponding sample layers is determined based on actual needs (such as the number of financial risk control strategy models that need to be backdated). For example, it can be divided into 10 orthogonal sample layers, and each layer is configured with a random ID generator. The random ID generator can be used to generate the sample layer ID, and the IDs between the sample layers are orthogonal, that is, the UD of each sample layer in this layer is independent of the IDs of other sample layers.

[0076] Step S12: generating multiple sample groups in each sample layer according to a preset number of sample groups.

[0077] Within each orthogonal sample layer, multiple mutually exclusive sample groups can be further divided according to requirements. For example, each sample layer can be divided into 5 mutually exclusive sample groups.

[0078] For example, a method such as divisor remainder can be used to generate a unique sample group ID for each sample group based on the sample layer ID. It should be noted that the samples in each sample group are mutually exclusive, that is, each sample belongs to only one sample group.

[0079] After determining each sample layer and sample group and their corresponding IDs, they are stored in the database. The ID of each sample layer and sample group is used to represent their identity, and corresponding indexes and shards are established to facilitate the establishment of mapping relationships in the subsequent process of associating with the financial risk control strategy model.

[0080] Furthermore, in a feasible embodiment, the step of allocating the multiple groups of financial sample data to the sample layers and the sample groups includes:

[0081] Step S21, adding each group of financial sample data to each sample layer;

[0082] Step S22 : Allocate each group of financial sample data in each sample layer to multiple sample groups, wherein, in the same sample layer, the financial sample data in different sample groups are mutually exclusive.

[0083] In the process of allocating financial sample data, it is first allocated to the sample layer. It should be noted that each sample layer includes all financial sample data, that is, the full amount of sample data.

[0084] Specifically, a corresponding ID can be assigned to each group of financial sample data through the ID generator of each sample layer. The ID is related to the ID of the sample layer and the ID of the sample group. According to the ID of each financial sample data, it can be assigned to the corresponding sample layer and sample group.

[0085] It should be noted that during the allocation of financial sample data, it is necessary to ensure the orthogonality of each sample layer and the mutual exclusivity of each sample group. For example, data validation can be used to ensure that each sample layer contains the full set of financial sample data and that each sample group within each layer is mutually exclusive. For example, SQL (Structured Query Language) queries or Spark (an open source distributed computing system) DataFrame (a distributed dataset) operations and verification can be used.

[0086] In a feasible embodiment, the step of associating the preset multiple financial risk control strategy models with each sample layer and sample group may include:

[0087] Step S31: Divide the financial risk control strategy models into a plurality of groups according to the correlation between the financial risk control strategy models, wherein each group of financial risk control strategy models includes one financial risk control strategy model or a plurality of interrelated financial risk control strategy models;

[0088] Step S32: Associating each group of financial risk control strategy models with each sample layer in a one-to-one correspondence;

[0089] Step S33 : Associating each financial risk control strategy model in each group of financial risk control strategy models with at least two sample groups in the corresponding sample layer, wherein each financial risk control strategy model in the same group is associated with different sample groups.

[0090] It should be noted that multiple financial risk control strategy models can be pre-defined based on risk control requirements. Each model can have different algorithms and parameter settings, such as a user behavior analysis model based on machine learning, a transaction risk assessment model based on rules, etc.

[0091] Generally, there's a one-to-one correspondence between financial risk control strategy models and sample layers, with each financial risk control strategy model corresponding to one sample layer. It's important to note that in financial risk control scenarios, models that influence each other's risk control strategies typically refer to models that rely on or share data within the business process. Adjustments to these models can directly or indirectly alter each other's input data, customer group distribution, or risk exposure levels. For example, for credit scoring and anti-fraud models, the rigor of the anti-fraud model directly impacts the distribution of customer samples fed to the credit scoring model. Adjustments to the credit scoring model's pass rate can alter fraudsters' attack strategies, thereby impacting the anti-fraud model's detection effectiveness. The two interact.

[0092] If there are financial risk control strategy models that affect each other, they can be considered as the same financial risk control strategy model group, and each financial risk control strategy model group corresponds to a sample layer. Financial risk control strategy models in the same group are assigned to different sample groups in the same sample layer.

[0093] For example, Figure 2 As shown, for sample layer A, it includes four different mutually exclusive sample groups a, sample group b, sample group c, and sample group d. There are mutually influencing financial risk control strategy model A and financial risk control strategy model B. Each mutually exclusive sample group in sample layer A needs to be assigned to financial risk control strategy model A and financial risk control strategy model B respectively, such as assigning sample groups a and b to financial risk control strategy model A, and assigning sample groups c and d to financial risk control strategy model B.

[0094] For ease of understanding, combined with the contents of the aforementioned application embodiments, such as Figure 3 As shown, first, the ID of each sample layer is generated by random ID generator 1, random ID generator 2, random ID generator 3, and random ID generator 4 of each sample layer, and the full amount of samples are allocated to the sample layer to obtain orthogonal samples 1, orthogonal samples 2, orthogonal samples 3, and orthogonal samples 4 corresponding to each sample layer respectively. Then, each orthogonal sample is grouped by grouper 1, grouper 2, grouper 3, and grouper 4 to obtain the grouped orthogonal sample 1. Finally, the two groups of orthogonal samples 1 in the same sample layer are allocated to the same strategy model (such as strategy models 1 and 2), where the strategy model is equivalent to the aforementioned financial risk control strategy model.

[0095] In a feasible embodiment, the step of inputting the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model and outputting the backtracking results corresponding to each financial risk control strategy model in parallel may include:

[0096] Step S41: generating multiple backtracking tasks, wherein each backtracking task corresponds to a financial risk control strategy model;

[0097] Step S42: assign each backtracking task to the corresponding computing node, and execute the backtracking task in parallel through each computing node according to the financial sample data in the sample group associated with each financial risk control strategy model, and output the backtracking results corresponding to each financial risk control strategy model.

[0098] First, multiple backtracking tasks can be generated based on a parallel computing framework (such as Spark). Each backtracking task corresponds to a financial risk control strategy model, and also corresponds to a sample layer and at least two sample groups associated with the financial risk control strategy model.

[0099] It should be noted that when generating backtracking tasks, it is necessary to ensure that the input data of each backtracking task are independent of each other to avoid data conflicts.

[0100] A preconfigured task scheduler ensures efficient execution of each backtracking task across computing resources. For example, Spark's scheduling mechanism distributes tasks across multiple computing nodes, monitors task execution status, and ensures smooth progress. Upon completion, each task node outputs the backtracking results corresponding to each financial risk control strategy model. Each computing node can correspond to a different server or processor, enabling the simultaneous execution of backtracking tasks for multiple risk control strategies using distributed computing resources, improving computing resource utilization and task execution efficiency.

[0101] Through the parallel backtracking method, computing resources can be effectively utilized, backtracking efficiency can be improved, and computing time can be reduced, which greatly improves the backtracking efficiency in scenarios where multiple financial risk control strategy models need to be backtracked.

[0102] Furthermore, the sample groups associated with the financial risk control strategy models include at least an experimental group; and the step of executing the backtracking task in parallel based on the financial sample data in the sample groups associated with each financial risk control strategy model and outputting the backtracking results corresponding to each financial risk control strategy model includes:

[0103] Step S421 , reading financial sample data from the sample groups in the sample layer corresponding to each financial risk control strategy model based on each backtracking task;

[0104] Step S422: Input the financial sample data from the experimental group into the corresponding financial risk control strategy model, process it through the financial risk control strategy model, and obtain the backtracking results corresponding to each financial risk control strategy model;

[0105] Step S423: Output each backtracking result to a preset file system or database.

[0106] When executing, each backtracking task will read the corresponding financial sample data from the sample group in the sample layer corresponding to the corresponding financial risk control strategy model, where the read financial sample data includes at least the financial sample data corresponding to the experimental group and the control group respectively, but only the financial sample data of the experimental group will be processed.

[0107] Each backtracking task inputs the financial sample data of the experimental group into the corresponding financial risk control strategy model, which then performs calculations, processing, and analysis. The financial risk control strategy model finally outputs the backtracking results (pass or fail) corresponding to each group of financial sample data in the corresponding experimental group. The backtracking results represent the risk assessment results of the users corresponding to each group of financial sample data. The backtracking results of each financial risk control strategy model are then stored in a specified output path. Specifically, a distributed file system or database can be used for data storage. The integrity and consistency of the backtracking results must be ensured to avoid data loss or errors. Finally, the backtracking results in each output path are aggregated into a unified data set for result aggregation and evaluation.

[0108] Furthermore, in a feasible embodiment, the sample group associated with the financial risk control strategy model also includes a control group; after the steps of inputting the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model and outputting the backtracking results corresponding to each financial risk control strategy model in parallel, the method further includes:

[0109] Step S50: Calculate the performance evaluation index corresponding to each financial risk control strategy model based on the backtracking results corresponding to each financial risk control strategy model, where the performance evaluation index includes at least one of precision and recall;

[0110] Step S60: Evaluate the risk control effects of each financial risk control strategy model based on the backtracking results corresponding to each financial risk control strategy model and the financial sample data of the control group;

[0111] Step S70: Obtain data processing time and resource usage information corresponding to each backtracking task;

[0112] Step S80: Generate a backtracking report corresponding to each financial risk control strategy model based on the performance evaluation indicators, risk control effects, data processing time, and resource usage information corresponding to each financial risk control strategy model.

[0113] After obtaining the backtracking results, the performance and risk control effects of each financial risk control strategy model can be evaluated so as to subsequently optimize the financial risk control strategy model and improve the level of financial risk control.

[0114] Specifically, the corresponding performance evaluation indicators are first calculated based on the backtracking results of each financial risk control strategy model. The backtracking results reflect the model's predicted evaluation results for the user, and the financial sample data carries the user's true label. Based on the backtracking results and the corresponding true labels, the accuracy of the financial risk control strategy model's backtracking results for each user can be determined, thereby further calculating the precision and recall rates. Common model performance evaluation indicators such as the F1 score can also be included. The calculation of these model performance evaluation indicators can also be completed by the backtracking tasks on the computing nodes in parallel, further improving the efficiency of multi-model performance evaluation.

[0115] On the other hand, after the financial risk control strategy model generates corresponding backtracking results for the user, a corresponding risk control strategy will usually be adopted, so that the user passes or fails the loan approval or other procedures, bringing about certain risk control effects or consequences.

[0116] Specifically, the strategy backtesting process applies the current financial risk control strategy model (including credit scoring models, anti-fraud rules, and credit limit policies) to historical financial sample data, simulating the risk control strategy's operational results over past time periods and analyzing its performance in identifying risk, controlling losses, and balancing returns. For example, the effectiveness of the risk control strategy model is verified, confirming whether it accurately identifies historical risk events (such as delinquencies and fraud); business impact is assessed, quantifying the impact of the risk control model strategy on business indicators such as approval rates, bad debt rates, and profits; it helps identify potential problems and identifies thresholds where the risk control model strategy is too loose (missing risks) or too strict (accidentally killing good customers); it supports iterative optimization, and provides a data basis for adjusting rule thresholds, model parameters, or features. When evaluating risk control performance, the financial sample data of the control group is used as a reference for the evaluation. The risk control performance of the financial risk control strategy model can include qualitative results, such as excellent, fair, or poor, or can be quantitatively expressed as a score (e.g., a percentage relative to the control group).

[0117] Furthermore, the resource utilization of the financial risk control strategy model can be evaluated based on factors such as data processing time and resource usage during the execution of the backtracking task. Finally, the performance evaluation indicators, risk control effectiveness, data processing time, and resource usage information identified above are aggregated in the backtracking report, presenting a comprehensive and multi-faceted assessment and analysis of the financial risk control strategy model. The report can be stored in a database or file system for reference and strategy optimization by the risk control team.

[0118] In one feasible embodiment, the performance of the underlying system is further optimized based on the execution status of the backtracking tasks and the backtracking reports. For example, the parameters of the computing framework are adjusted, and the data storage and reading mechanisms are optimized. Based on the data processing time and resource usage information corresponding to the aforementioned backtracking tasks, load balancing and resource scheduling can be used to ensure efficient system operation.

[0119] Furthermore, you can design appropriate fault handling mechanisms to ensure automatic recovery if a failure occurs during task execution. Use a logging system to record detailed task execution information for easier troubleshooting and analysis. Regular system maintenance, including data backups, system updates, and performance monitoring, ensures system stability and security, preventing data leaks and system crashes.

[0120] The technical solutions of the embodiments of this application are not limited to the fields of credit approval, anti-fraud, and market risk management in the financial industry. They can also be applied to scenarios such as e-commerce that require efficient risk assessment and control. For example, in the financial industry, they can be used to evaluate the effectiveness of different credit strategies and quickly identify potential fraudulent behavior; in e-commerce, they can be used to monitor and assess transaction risks and improve platform security and user experience.

[0121] In summary, the multi-risk control strategy parallel backtracking method provided in this embodiment of the present application can efficiently and accurately complete the backtracking task of multiple financial risk control strategy models, significantly improving the performance of the risk control system and the accuracy of effect evaluation. Furthermore, its high scalability and flexibility enable it to adapt to the risk management needs of different industries and scenarios.

[0122] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the parallel backtracking method of multiple risk control strategies of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0123] This application also provides a multi-risk control strategy parallel backtracking device, refer to Figure 4 The multi-risk control strategy parallel backtracking device includes:

[0124] A sample layer determination module M10 is used to determine multiple sample layers and multiple sample groups contained in each sample layer;

[0125] a sample allocation module M20 for allocating multiple sets of financial sample data to the sample layers and the sample groups, wherein each sample layer includes the full amount of financial sample data, and the financial sample data includes the behavior logs, transaction records and / or historical risk events of financial users;

[0126] A model association module M30 is configured to associate a plurality of preset financial risk control strategy models with each of the sample layers and sample groups, wherein each financial risk control strategy model is associated with at least two sample groups in a sample layer;

[0127] The model backtracking module M40 is used to input the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model, and output the backtracking results corresponding to each financial risk control strategy model in parallel.

[0128] In one embodiment, the sample layer determination module M10 is further configured to:

[0129] Generate multiple sample layers according to the preset number of sample layers;

[0130] According to the preset number of sample groups, multiple sample groups are generated in each sample layer.

[0131] In one embodiment, the sample allocation module M20 is further configured to:

[0132] Adding each set of financial sample data to each sample layer;

[0133] The financial sample data in each sample layer are respectively allocated to a plurality of sample groups, wherein in the same sample layer, the financial sample data in different sample groups are mutually exclusive.

[0134] In one embodiment, the model association module M30 is further configured to:

[0135] According to the correlation between the financial risk control strategy models, the financial risk control strategy models are divided into multiple groups, wherein each group of financial risk control strategy models includes one financial risk control strategy model or multiple financial risk control strategy models that are correlated with each other;

[0136] Associating each group of financial risk control strategy models with each of the sample layers in a one-to-one correspondence;

[0137] Each financial risk control strategy model in each group of financial risk control strategy models is associated with at least two sample groups in the corresponding sample layer, wherein each financial risk control strategy model in the same group is associated with different sample groups.

[0138] In one embodiment, the model backtracking module M40 is further configured to:

[0139] Generate multiple backtracking tasks, where each backtracking task corresponds to a financial risk control strategy model;

[0140] Each of the backtracking tasks is assigned to a corresponding computing node, and each of the computing nodes executes the backtracking task in parallel according to the financial sample data in the sample group associated with each of the financial risk control strategy models, and outputs the backtracking results corresponding to each financial risk control strategy model.

[0141] In one embodiment, the sample group associated with the financial risk control strategy model includes at least an experimental group;

[0142] The model backtracking module M40 is also used to:

[0143] Reading financial sample data from the sample group in the sample layer corresponding to each financial risk control strategy model based on each backtracking task;

[0144] Inputting the financial sample data from the experimental group into the corresponding financial risk control strategy model, processing it through the financial risk control strategy model, and obtaining the backtracking results corresponding to each of the financial risk control strategy models;

[0145] Output each of the backtracking results to a preset file system or database.

[0146] In one embodiment, the sample group associated with the financial risk control strategy model also includes a control group;

[0147] The multi-risk control strategy parallel backtracking device also includes a report generation module, which is used to:

[0148] Calculating performance evaluation indicators corresponding to the financial risk control strategy models based on the backtracking results corresponding to the financial risk control strategy models, wherein the performance evaluation indicators include at least one of precision and recall;

[0149] Evaluate the risk control effects of each of the financial risk control strategy models based on the backtracking results corresponding to each of the financial risk control strategy models and the financial sample data of the control group;

[0150] Obtaining data processing time and resource usage information corresponding to each of the backtracking tasks;

[0151] A backtracking report corresponding to each of the financial risk control strategy models is generated based on the performance evaluation indicators, the risk control effect, the data processing time and the resource usage information corresponding to each of the financial risk control strategy models.

[0152] The multi-risk control strategy parallel backtracking device provided in this application adopts the multi-risk control strategy parallel backtracking method in the above-mentioned embodiment, which can solve the technical problem that when backtracking multiple risk control strategies, there is mutual interference between strategies, which affects the evaluation accuracy of the risk control model. Compared with the existing technology, the beneficial effects of the multi-risk control strategy parallel backtracking device provided in this application are the same as the beneficial effects of the multi-risk control strategy parallel backtracking method provided in the above-mentioned embodiment, and the other technical features of the multi-risk control strategy parallel backtracking device are the same as the features disclosed in the previous embodiment method, which will not be repeated here.

[0153] The present application provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the multi-risk control strategy parallel backtracking method in the above-mentioned embodiment one.

[0154] Reference below Figure 5 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present application. The electronic devices in the embodiments of the present application may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (such as in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0155] like Figure 5As shown, the electronic device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage device 1003 into a random access memory (RAM) 1004. Various programs and data required for the operation of the electronic device are also stored in RAM 1004. The processing device 1001, ROM 1002, and RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows an electronic device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have instead.

[0156] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0157] The electronic device provided by this application adopts the multi-risk control strategy parallel backtracking method in the above embodiment, which can solve the technical problem that when multiple risk control strategies are backtracked, there is mutual interference between strategies, which affects the evaluation accuracy of the risk control model. Compared with the existing technology, the beneficial effects of the electronic device provided by this application are the same as the beneficial effects of the multi-risk control strategy parallel backtracking method provided by the above embodiment, and the other technical features of the electronic device are the same as the features disclosed in the previous embodiment method, which will not be repeated here.

[0158] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0159] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0160] The present application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, and the computer-readable program instructions are used to execute the multi-risk control strategy parallel backtracking method in the above-mentioned embodiment.

[0161] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0162] The computer-readable storage medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0163] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by an electronic device, the electronic device executes: determining multiple sample layers and multiple sample groups contained in each sample layer; allocating multiple groups of financial sample data to each of the sample layers and each of the sample groups, wherein each sample layer includes a full amount of financial sample data, and the financial sample data includes the behavior log, transaction record and / or historical risk events of the financial user; associating multiple preset financial risk control strategy models with each of the sample layers and sample groups, wherein each financial risk control strategy model is associated with at least two sample groups in a sample layer; inputting the financial sample data of each sample group in each of the sample layers into the corresponding financial risk control strategy model, and outputting the backtracking results corresponding to each of the financial risk control strategy models in parallel.

[0164] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0165] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0166] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0167] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., a computer program) for executing the above-mentioned multi-risk control strategy parallel backtracking method. It can solve the technical problem that when multiple risk control strategies are backtracked, there is mutual interference between strategies, which affects the evaluation accuracy of the risk control model. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the multi-risk control strategy parallel backtracking method provided in the above-mentioned embodiment, and will not be repeated here.

[0168] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned multi-risk control strategy parallel backtracking method.

[0169] The computer program product provided in this application can address the technical problem of interference between strategies during the backtracking of multiple risk control strategies, which affects the assessment accuracy of the risk control model. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the parallel backtracking method for multiple risk control strategies provided in the above-mentioned embodiments, and will not be elaborated here.

[0170] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A multi-risk control strategy parallel backtracking method, characterized in that: The multi-risk control strategy parallel backtracking method includes: determining a plurality of sample layers and a plurality of sample groups contained in each sample layer; Allocating multiple sets of financial sample data to each of the sample layers and each of the sample groups, wherein each sample layer includes a full amount of financial sample data, and the financial sample data includes a financial user's behavior log, transaction record, and / or historical risk events; Associating a plurality of preset financial risk control strategy models with each of the sample layers and sample groups, respectively, wherein each financial risk control strategy model is associated with at least two sample groups in one sample layer; The financial sample data of each sample group in each sample layer are respectively input into the corresponding financial risk control strategy model, and the backtracking results corresponding to each financial risk control strategy model are output in parallel.

2. The multi-risk control strategy parallel backtracking method according to claim 1, characterized in that: The step of determining multiple sample layers and multiple sample groups contained in each sample layer includes: Generate multiple sample layers according to the preset number of sample layers; According to the preset number of sample groups, multiple sample groups are generated in each sample layer.

3. The multi-risk control strategy parallel backtracking method according to claim 2, characterized in that: The step of allocating the plurality of groups of financial sample data to the sample layers and the sample groups comprises: Adding each set of financial sample data to each sample layer; Each group of financial sample data in each sample layer is respectively allocated to a plurality of sample groups, wherein, in the same sample layer, the financial sample data in different sample groups are mutually exclusive.

4. The multi-risk control strategy parallel backtracking method according to claim 1, characterized in that: The step of associating the preset multiple financial risk control strategy models with each of the sample layers and sample groups includes: According to the correlation between the financial risk control strategy models, the financial risk control strategy models are divided into multiple groups, wherein each group of financial risk control strategy models includes one financial risk control strategy model or multiple financial risk control strategy models that are correlated with each other; Associating each group of financial risk control strategy models with each of the sample layers in a one-to-one correspondence; Each financial risk control strategy model in each group of financial risk control strategy models is associated with at least two sample groups in the corresponding sample layer, wherein each financial risk control strategy model in the same group is associated with different sample groups.

5. The multi-risk control strategy parallel backtracking method according to claim 1, characterized in that: The step of inputting the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model and outputting the backtracking results corresponding to each financial risk control strategy model in parallel includes: Generate multiple backtracking tasks, where each backtracking task corresponds to a financial risk control strategy model; Each of the backtracking tasks is assigned to a corresponding computing node, and each of the computing nodes executes the backtracking task in parallel according to the financial sample data in the sample group associated with each of the financial risk control strategy models, and outputs the backtracking results corresponding to each financial risk control strategy model.

6. The multi-risk control strategy parallel backtracking method according to claim 5, characterized in that: The sample group associated with the financial risk control strategy model includes at least an experimental group; The step of executing the backtracking task in parallel based on the financial sample data in the sample groups associated with each of the financial risk control strategy models and outputting the backtracking results corresponding to each of the financial risk control strategy models comprises: Reading financial sample data from the sample group in the sample layer corresponding to each financial risk control strategy model based on each backtracking task; Inputting the financial sample data from the experimental group into the corresponding financial risk control strategy model, processing it through the financial risk control strategy model, and obtaining the backtracking results corresponding to each of the financial risk control strategy models; Output each of the backtracking results to a preset file system or database.

7. The multi-risk control strategy parallel backtracking method according to claim 6, characterized in that: The sample group associated with the financial risk control strategy model also includes a control group; After the steps of inputting the financial sample data of each sample group in each sample layer into the corresponding financial risk control strategy model and outputting the backtracking results corresponding to each financial risk control strategy model in parallel, the method further includes: Calculating performance evaluation indicators corresponding to the financial risk control strategy models based on the backtracking results corresponding to the financial risk control strategy models, wherein the performance evaluation indicators include at least one of precision and recall; Evaluate the risk control effects of each of the financial risk control strategy models based on the backtracking results corresponding to each of the financial risk control strategy models and the financial sample data of the control group; Obtaining data processing time and resource usage information corresponding to each of the backtracking tasks; A backtracking report corresponding to each of the financial risk control strategy models is generated based on the performance evaluation indicators, the risk control effect, the data processing time and the resource usage information corresponding to each of the financial risk control strategy models.

8. An electronic device, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the multi-risk control strategy parallel backtracking method according to any one of claims 1 to 7.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the multi-risk control strategy parallel backtracking method according to any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images