A consumer finance credit risk assessment system based on a random forest algorithm

Abstract

This invention provides a consumer finance credit risk assessment system based on the random forest algorithm, belonging to the field of computer data processing technology. By combining computational resource constraints with the value of data services, this invention achieves a dynamic balance between data processing speed and risk assessment accuracy. It adaptively adjusts data processing timing based on real-time load, avoiding business timeouts caused by waiting for late data while ensuring stringent response time requirements are met. This improves service stability in high-concurrency scenarios. By quantifying the value decay of late data instead of directly discarding it, it enhances the utilization efficiency of data information, retains potential key risk signals in late events, and reasonably quantifies their impact on the final decision through time value discounting. This reduces the model misjudgment rate caused by missing features, improving the accuracy of risk identification without sacrificing system performance.

Description

Technical Field

[0001] This invention relates to the field of computer data processing technology, and in particular to a consumer finance credit risk assessment system based on the random forest algorithm. Background Technology

[0002] In the consumer finance sector, credit risk assessment systems are a core business component. They utilize computer data processing technology to quickly assess the credit risk of applicants. Typically, this involves combining users' historical behavioral data with real-time transaction or behavioral data, and using pre-built machine learning models, such as the random forest algorithm, to generate decision results. The performance of the system directly affects business efficiency and bad debt losses, thus requiring extremely high real-time performance and accuracy in data processing.

[0003] In existing technologies, risk assessment systems typically employ a fixed-time-window-based data collection strategy to integrate historical and real-time data. The system sets a predefined time period, such as several hundred milliseconds, as a buffer window to wait for real-time events that may arrive out of order due to network jitter or distributed system latency. Once the time window closes, the system aggregates all data collected within the window to form a feature vector, which is then input into a random forest model for computation. However, if the fixed time window is set too long, while more late-arriving data can be collected, the overall decision-making time increases, making it difficult to meet the millisecond-level response time requirements typically required in financial operations. Conversely, if the time window is set too short to ensure response speed, events that arrive slightly later but may contain important risk information will be frequently discarded, resulting in incomplete data features input to the model and affecting the accuracy of risk assessment. Therefore, there is room for improvement. Summary of the Invention

[0004] This invention provides a consumer finance credit risk assessment system based on the random forest algorithm, which uses a method of generating event truncation time thresholds and performing value decay processing on late events, thereby improving the accuracy of risk assessment while meeting low latency requirements.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a consumer finance credit risk assessment system based on the random forest algorithm is provided, including: The acquisition module is used to acquire the remaining computing time budget of the current computing node and the stacking density of the real-time event queue, generate system status parameters, and access the real-time event stream carrying event timestamps and arrival timestamps and user historical feature snapshots. The calculation module is used to calculate the resource constraints based on the system state parameters and obtain the expected value loss tolerance used to characterize the acceptable risk loss of the business. The threshold generation module is used to generate an event cutoff time threshold based on the resource constraints and the expected value loss tolerance by evaluating the expected risk recovery value that can be recovered by waiting for the real-time event stream under the premise of satisfying the resource constraints. The processing module is used to perform time-series determination and value decay processing on events in the real-time event stream using the event truncation time threshold, generate a time-series calibration feature vector, and determine the time delay weight in the real-time event stream through the real-time feature vector carrying the time decay weight. The fusion module is used to fuse the time-series calibration feature vector with the user's historical feature snapshot to obtain a fused feature vector; The evaluation module is used to input the fused feature vector into the random forest model and output the risk assessment result corresponding to the real-time event stream.

[0006] Optionally, the threshold generation module is specifically used for: The maximum wait time window is calculated based on the aforementioned resource constraints. Obtain a value prediction model trained based on the relationship between the risk value and delay duration of historical late events, input the maximum waitable time window into the value prediction model, and output the expected risk recovery value. When the expected risk recovery value is greater than the expected value loss tolerance, the event cutoff time threshold is generated based on the current system time and the maximum waitable time window. When the expected risk recovery value is less than or equal to the expected value loss tolerance, the current system time is used as the event cutoff time threshold.

[0007] Optionally, the processing module is specifically used for: The event timestamp of each event in the real-time event stream is compared with the event truncation time threshold; If the event timestamp is not later than the event truncation time threshold, the corresponding event is marked as a timely event and given full weight. If the event timestamp is later than the event truncation time threshold, the corresponding event is marked as a late event, and the late time difference between it and the event truncation time threshold is calculated. The time value decay rate of the corresponding late event is calculated by using a time value discounting function that maps the lateness time difference to a value discount. The feature value of the late event is discounted using the time value decay rate to obtain the discounted feature.

[0008] Optionally, the processing module is further configured to: Obtain a feature influence weight matrix to characterize the differentiated impact weights of different features on risk assessment results under different degrees of lateness; The time-series calibration feature vector is generated by weighting the feature values ​​of the on-time event, the discounted features, and the feature influence weight matrix. The introduction of the feature influence weight matrix enables the time-series calibration feature vector to synergistically reflect the combined impact of the event's intrinsic feature values ​​and time-series value on the final risk.

[0009] Optionally, the system further includes a model feedback optimization module, used for: Monitor the accuracy of the risk assessment results output by the random forest model within a specified time window; When the accuracy metric falls below its baseline, the event samples accumulated within this time window are used as incremental training data to dynamically adjust the internal parameters of the value prediction model, thereby enabling the value prediction model to adapt to changing risk patterns in assessing the expected risk recovery value.

[0010] Optionally, the model feedback optimization module is further used for: While dynamically adjusting the internal parameters of the value prediction model, the decay coefficient of the time value discount function is also dynamically adjusted in sync with the direction and magnitude of the change in the accuracy index. This sync adjustment enables the value assessment and value discount processes to be optimized in a coordinated manner.

[0011] Optionally, the time value discount function is a non-linear discount function, and the time value decay rate output by it is negatively correlated with the late arrival time difference, and the rate of value decay changes dynamically as the late arrival time difference increases.

[0012] Optionally, the fusion module is specifically used for: The fusion of the time-series calibration feature vector and the user historical feature snapshot is completed in memory by performing a sparse matrix addition operation that quickly updates the index positions of non-zero feature values ​​in the user historical feature snapshot.

[0013] Optionally, the evaluation module is specifically used for: The risk assessment result is obtained by simultaneously distributing the fused feature vector to multiple decision trees within the random forest model for parallel computation and aggregating the outputs of all decision trees.

[0014] Secondly, a consumer finance credit risk assessment method based on the random forest algorithm is provided, including the following steps: Obtain the remaining computing time budget and the stacking density of the real-time event queue of the current computing node, generate system status parameters, and access the real-time event stream carrying event timestamps and arrival timestamps and user historical feature snapshots. Based on the system state parameters, resource constraints are calculated, and the expected value loss tolerance, which characterizes the acceptable risk loss of the business, is obtained. Based on the resource constraints and the expected value loss tolerance, an event cutoff time threshold is generated by evaluating the expected risk recovery value that can be recovered by waiting for the real-time event stream under the premise of satisfying the resource constraints. The event truncation time threshold is used to perform time sequence determination and value decay processing on events in the real-time event stream, generating a time sequence calibration feature vector. The time delay weight in the real-time event stream is determined by the real-time feature vector carrying the time decay weight. The time-series calibration feature vector is fused with the user's historical feature snapshot to obtain a fused feature vector; The fused feature vector is input into the random forest model, and the risk assessment result corresponding to the real-time event stream is output.

[0015] Thirdly, an electronic device is provided, comprising: a processor and a memory; the memory is used to store a computer program, which, when executed by the processor, causes the electronic device to perform the consumer finance credit risk assessment system based on the random forest algorithm described in the first aspect.

[0016] In one possible design, the electronic device described in the third aspect may further include a transceiver. This transceiver may be a transceiver circuit or an interface circuit. The transceiver can be used for communication between the electronic device described in the third aspect and other electronic devices.

[0017] In the embodiments of the present invention, the electronic device described in the third aspect may be a terminal, or a chip (system) or other component or assembly disposed in the terminal, or a system containing the terminal.

[0018] Fourthly, a computer-readable storage medium is provided, comprising: a computer program or instructions; when the computer program or instructions are run on a computer, the computer causes the computer to execute the consumer finance credit risk assessment system based on the random forest algorithm described in the first aspect.

[0019] In summary, the above methods and systems have the following technical effects: This invention achieves a dynamic balance between data processing speed and risk assessment accuracy by combining computational resource constraints with data business value. It adaptively adjusts data processing timing based on real-time load, avoiding business timeouts caused by waiting for late data while ensuring that stringent response time requirements are met. This improves service stability in high-concurrency scenarios. By quantifying the value decay of late data instead of discarding it directly, it improves the utilization efficiency of data information, preserves potential key risk signals in late events, and reasonably quantifies their impact on the final decision through time value discounting. This reduces the model misjudgment rate caused by feature loss and improves the accuracy of risk identification without sacrificing system performance. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the consumer finance credit risk assessment system based on the random forest algorithm provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0022] In this embodiment of the invention, "instruction" can include direct and indirect instructions, as well as explicit and implicit instructions. The information indicated by a certain piece of information is called the information to be instructed. In specific implementation, there are many ways to instruct the information to be instructed, such as, but not limited to, directly instructing the information to be instructed, such as the information to be instructed itself or its index. It can also indirectly instruct the information to be instructed by instructing other information, where there is a correlation between the other information and the information to be instructed. It can also instruct only a part of the information to be instructed, while the other parts are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing instruction overhead to some extent. Simultaneously, common parts of various pieces of information can be identified and uniformly indicated to reduce the instruction overhead caused by individually indicating the same information.

[0023] Furthermore, the specific indication method can also be any existing indication method, such as, but not limited to, the above-mentioned indication methods and their various combinations. Specific details of various indication methods can be found in existing technologies, and will not be elaborated upon here. As described above, for example, when multiple pieces of information of the same type need to be indicated, the indication methods for different pieces of information may differ. In specific implementation, the required indication method can be selected according to specific needs. This embodiment of the invention does not limit the selected indication method; therefore, the indication methods involved in this embodiment of the invention should be understood to cover various methods that enable the party to be indicated to obtain the information to be indicated.

[0024] It should be understood that the information to be indicated can be sent as a whole or divided into multiple sub-information messages sent separately, and the sending period and / or timing of these sub-information messages can be the same or different. The specific sending method is not limited in this embodiment of the invention. The sending period and / or timing of these sub-information messages can be predefined, for example, according to a protocol, or configured by the sending device by sending configuration information to the receiving device.

[0025] "Predefined" or "pre-configured" can be achieved by pre-saving corresponding codes, tables, or other means that can be used to indicate relevant information in the device. This embodiment of the invention does not limit the specific implementation method. "Saving" can refer to saving in one or more memories. These memories can be separate installations or integrated into the encoder, decoder, processor, or electronic device. Alternatively, some memories can be separately installed, while others are integrated into the decoder, processor, or electronic device. The type of memory can be any form of storage medium, and this embodiment of the invention does not limit this.

[0026] In the embodiments of this invention, "protocol" may refer to a protocol family in the field of communication, a standard protocol with a similar protocol family frame structure, or a related protocol applied to a future consumer finance credit risk assessment system based on the random forest algorithm. The embodiments of this invention do not specifically limit this.

[0027] In this embodiment of the invention, descriptions such as "when," "under the circumstances," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a specific time. They do not require the device to make a judgment action during implementation, nor do they imply any other limitations.

[0028] In the description of the embodiments of the present invention, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in the embodiments of the present invention is merely a description of the relationship between the related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of the embodiments of the present invention, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Additionally, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or order of execution, and that "first," "second," etc., are not necessarily different. Furthermore, in the embodiments of this invention, words such as "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this invention should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0029] The network architecture and business scenarios described in the embodiments of this invention are for the purpose of more clearly illustrating the technical solutions of the embodiments of this invention, and do not constitute a limitation on the technical solutions provided by the embodiments of this invention. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided by the embodiments of this invention are also applicable to similar technical problems.

[0030] Figure 1 This is a flowchart illustrating the method provided in an embodiment of the present invention. The consumer finance credit risk assessment system based on the random forest algorithm, applied on the server side, is configured to include: The acquisition module is used to acquire the remaining computing time budget of the current computing node and the stacking density of the real-time event queue, generate system status parameters, and access the real-time event stream carrying event timestamps and arrival timestamps and user historical feature snapshots. The calculation module is used to calculate the resource constraints based on the system state parameters and obtain the expected value loss tolerance used to characterize the acceptable risk loss of the business. The threshold generation module is used to generate an event cutoff time threshold based on the resource constraints and the expected value loss tolerance by evaluating the expected risk recovery value that can be recovered by waiting for the real-time event stream under the premise of satisfying the resource constraints. The processing module is used to perform time-series determination and value decay processing on events in the real-time event stream using the event truncation time threshold, generate a time-series calibration feature vector, and determine the time delay weight in the real-time event stream through the real-time feature vector carrying the time decay weight. The fusion module is used to fuse the time-series calibration feature vector with the user's historical feature snapshot to obtain a fused feature vector; The evaluation module is used to input the fused feature vector into the random forest model and output the risk assessment result corresponding to the real-time event stream.

[0031] The threshold generation module is specifically used for: The maximum wait time window is calculated based on the aforementioned resource constraints. Obtain a value prediction model trained based on the relationship between the risk value and delay duration of historical late events, input the maximum waitable time window into the value prediction model, and output the expected risk recovery value. When the expected risk recovery value is greater than the expected value loss tolerance, the event cutoff time threshold is generated based on the current system time and the maximum waitable time window. When the expected risk recovery value is less than or equal to the expected value loss tolerance, the current system time is used as the event cutoff time threshold.

[0032] Specifically: When generating dynamic event truncation time thresholds, the remaining computation time budget is obtained by subtracting the time already consumed by the current request from the total time quota stipulated in the business service level agreement.

[0033] The stacking density is calculated by combining the number of unprocessed events in the real-time event queue with the arrival rate to determine the maximum waitable time window. The formula for calculating the maximum waitable time window is as follows: ; in, The maximum waiting time window, Budget for the remaining computation time. The estimated time for processing the current backlog of events and performing model inference.

[0034] Input the maximum wait time window into the preset value prediction model, and output the expected risk recovery value: ; in, To recover value from anticipated risks, This is a function mapping for the value prediction model.

[0035] If the expected risk recovery value is greater than the preset expected value loss tolerance, the current system time plus the maximum waiting time window is used as the event cutoff time threshold; if it is less than or equal to the tolerance, the current system time is used as the event cutoff time threshold.

[0036] The processing module is specifically used for: The event timestamp of each event in the real-time event stream is compared with the event truncation time threshold; If the event timestamp is not later than the event truncation time threshold, the corresponding event is marked as a timely event and given full weight. If the event timestamp is later than the event truncation time threshold, the corresponding event is marked as a late event, and the late time difference between it and the event truncation time threshold is calculated. The time value decay rate of the corresponding late event is calculated by using a time value discounting function that maps the lateness time difference to a value discount. The feature value of the late event is discounted using the time value decay rate to obtain the discounted feature.

[0037] The processing module is also used for: Obtain a feature influence weight matrix to characterize the differentiated impact weights of different features on risk assessment results under different degrees of lateness; The time-series calibration feature vector is generated by weighting the feature values ​​of the on-time event, the discounted features, and the feature influence weight matrix. The introduction of the feature influence weight matrix enables the time-series calibration feature vector to synergistically reflect the combined impact of the event's intrinsic feature values ​​and time-series value on the final risk.

[0038] The time value discount function is a non-linear discount function. The time value decay rate output by the function is negatively correlated with the time difference of late arrival, and the rate of value decay changes dynamically as the time difference of late arrival increases.

[0039] Specifically: Iterate through the events in the real-time event stream, determine the events whose timestamps are no later than the event truncation time threshold as on-time events, and assign them a first preset weight of 1.0.

[0040] Events whose timestamps are later than the event truncation time threshold are classified as late events, and their lateness time difference is calculated.

[0041] The time value decay rate is calculated using a nonlinear time value discount function in the form of exponential decay: ; in, The time value decay rate, For the time difference of being late, This is the preset attenuation coefficient.

[0042] The original feature value of the late event is multiplied by the calculated time value decay rate to obtain the discounted feature.

[0043] Obtain a preset feature influence weight matrix, the elements of which represent the contribution weight of a specific feature to risk assessment within a specific lateness time interval.

[0044] Aggregate the feature values ​​of all on-time events with the discounted features into a preliminary feature vector, and then perform a Hadamard product operation with the dynamically extracted feature influence weight vector: ; Where Vcal is the time-series calibration feature vector, Vpre is the preliminary feature vector, and Wt is the feature influence weight vector.

[0045] The system also includes a model feedback optimization module, used for: Monitor the accuracy of the risk assessment results output by the random forest model within a specified time window; When the accuracy metric falls below its baseline, the event samples accumulated within this time window are used as incremental training data to dynamically adjust the internal parameters of the value prediction model, thereby enabling the value prediction model to adapt to changing risk patterns in assessing the expected risk recovery value.

[0046] The model feedback optimization module is also used for: While dynamically adjusting the internal parameters of the value prediction model, the decay coefficient of the time value discount function is also dynamically adjusted in sync with the direction and magnitude of the change in the accuracy index. This sync adjustment enables the value assessment and value discount processes to be optimized in a coordinated manner.

[0047] The system aggregates risk assessment results with real business tags within a preset time window, calculates an accuracy index, and triggers a parameter adjustment process when the accuracy index decreases from the benchmark value by more than a preset sensitivity threshold.

[0048] The samples accumulated within the time window are used as incremental training data. The value weight of the late data is increased for the bad samples that are missed, and the value weight of the misjudged samples is decreased, so as to update the internal parameters of the value prediction model.

[0049] The false positive and false negative rates attributable to late events in the synchronously statistically analyzed misjudged samples are calculated, and the adjustment amount is updated to the decay coefficient of the time value discount function. The update formula for the decay coefficient is as follows: ; ; in, For the new attenuation coefficient, This is the current attenuation coefficient. For learning rate, and $ are normalized measures of the false positive rate and false negative rate attributable to late events, respectively.

[0050] The fusion module is specifically used for: The fusion of the time-series calibration feature vector and the user historical feature snapshot is completed in memory by performing a sparse matrix addition operation that quickly updates the index positions of non-zero feature values ​​in the user historical feature snapshot.

[0051] The evaluation module is specifically used for: The risk assessment result is obtained by simultaneously distributing the fused feature vector to multiple decision trees within the random forest model for parallel computation and aggregating the outputs of all decision trees.

[0052] Specifically: In memory, user history feature snapshots are constructed as a sparse structure that stores only non-zero features and their index positions.

[0053] The non-zero features in the time-series calibration feature vector are traversed, and a search is performed in the sparse structure in memory using the feature index. If the index exists, the feature values ​​are added to update it; if it does not exist, a new key-value pair is inserted, thereby completing the sparse matrix addition operation in memory and generating the fused feature vector.

[0054] The thread pool technique is used to distribute the fused feature vectors to multiple worker threads in the random forest model for parallel traversal and computation. The main control thread collects the outputs of all decision trees and performs aggregation computation to obtain the final result.

[0055] In another embodiment, a consumer finance credit risk assessment method based on a random forest algorithm is also included, comprising the following steps: Obtain the remaining computing time budget and the stacking density of the real-time event queue of the current computing node, generate system status parameters, and access the real-time event stream carrying event timestamps and arrival timestamps and user historical feature snapshots. Based on the system state parameters, resource constraints are calculated, and the expected value loss tolerance, which characterizes the acceptable risk loss of the business, is obtained. Based on the resource constraints and the expected value loss tolerance, an event cutoff time threshold is generated by evaluating the expected risk recovery value that can be recovered by waiting for the real-time event stream under the premise of satisfying the resource constraints. The event truncation time threshold is used to perform time sequence determination and value decay processing on events in the real-time event stream, generating a time sequence calibration feature vector. The time delay weight in the real-time event stream is determined by the real-time feature vector carrying the time decay weight. The time-series calibration feature vector is fused with the user's historical feature snapshot to obtain a fused feature vector; The fused feature vector is input into the random forest model, which outputs the risk assessment result corresponding to the real-time event stream. In summary, by establishing a closed-loop decision-making framework that dynamically couples real-time computing resources, business risk value, and data timeliness, a dynamic resource constraint boundary is established by sensing the current computing load status of the system. Then, the purely time-related issue of data waiting is transformed into an economic decision based on risk value. Specifically, by quantitatively assessing the potential recoverable risk value of continuing to wait and the preset business tolerance, a critical point is dynamically generated to determine when to stop waiting—the event truncation time threshold. Based on this threshold, all real-time events are time-series determined, and the value of late events is innovatively quantified and decayed rather than simply discarded, generating a time-series calibration feature vector containing a time value dimension. This real-time feature is fused with the user's historical feature snapshot, providing optimized input to the downstream random forest model that balances timeliness and information integrity. Through a dynamic watermark mechanism, the system can adaptively adjust the waiting strategy according to the real-time load, ensuring that stringent response time requirements are met while avoiding business timeouts caused by rigid waiting.

[0056] The electronic device provided in this embodiment of the invention, exemplarily, can be a network device, or a chip (system) or other component or assembly that can be disposed in a network device. The electronic device may include a processor. Optionally, the electronic device may also include a memory and / or a transceiver. The processor is coupled to the memory and transceiver, for example, by means of a communication bus connection.

[0057] The following is a detailed introduction to the various components of the electronic device: In this context, the processor is the control center of the electronic device. It can be a single processor or a collective term for multiple processing elements. For example, a processor can be one or more central processing units (CPUs), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0058] Alternatively, the processor can perform various functions of the electronic device, such as the methods described above, by running or executing software programs stored in memory and by calling data stored in memory.

[0059] In a specific implementation, as one example, the processor may include one or more CPUs, such as CPU0 and CPU1.

[0060] In a specific implementation, as one example, the electronic device may also include multiple processors. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0061] The memory is used to store the software program that executes the solution of the present invention, and the execution is controlled by the processor. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0062] Optionally, the memory can be read-only memory (ROM) or other types of static storage devices capable of storing static information and instructions, random access memory (RAM) or other types of dynamic storage devices capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory can be integrated with the processor or exist independently and coupled to the processor through the interface circuit of the electronic device; the embodiments of the present invention do not specifically limit this.

[0063] A transceiver is used for communication with other electronic devices. For example, if the electronic device is a terminal, the transceiver can be used to communicate with a network device or with another terminal device. As another example, if the electronic device is a network device, the transceiver can be used to communicate with a terminal or with another network device.

[0064] Optionally, the transceiver may include a receiver and a transmitter. The receiver is used to implement the receiving function, and the transmitter is used to implement the sending function.

[0065] Optionally, the transceiver can be integrated with the processor or exist independently and coupled to the processor through the interface circuit of the electronic device. This embodiment of the invention does not specifically limit this.

[0066] It is understood that the structure of the electronic device in this embodiment does not constitute a limitation on the electronic device. The actual electronic device may include more or fewer components, or combine certain components, or have different component arrangements.

[0067] Furthermore, the technical effects of the electronic device can be referred to the technical effects of the method described in the above method embodiments, and will not be repeated here.

[0068] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0069] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0070] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0071] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0072] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0073] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0074] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0075] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0076] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0077] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0078] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0079] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0080] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A consumer finance credit risk assessment system based on the random forest algorithm, characterized in that, include: The acquisition module is used to acquire the remaining computing time budget of the current computing node and the stacking density of the real-time event queue, generate system status parameters, and access the real-time event stream carrying event timestamps and arrival timestamps and user historical feature snapshots. The calculation module is used to calculate the resource constraints based on the system state parameters and obtain the expected value loss tolerance used to characterize the acceptable risk loss of the business. The threshold generation module is used to generate an event cutoff time threshold based on the resource constraints and the expected value loss tolerance by evaluating the expected risk recovery value that can be recovered by waiting for the real-time event stream under the premise of satisfying the resource constraints. The processing module is used to perform time-series determination and value decay processing on events in the real-time event stream using the event truncation time threshold, generate a time-series calibration feature vector, and determine the time delay weight in the real-time event stream through the real-time feature vector carrying the time decay weight. The fusion module is used to fuse the time-series calibration feature vector with the user's historical feature snapshot to obtain a fused feature vector; The evaluation module is used to input the fused feature vector into the random forest model and output the risk assessment result corresponding to the real-time event stream.

2. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 1, characterized in that, The threshold generation module is specifically used for: The maximum wait time window is calculated based on the aforementioned resource constraints. Obtain a value prediction model trained based on the relationship between the risk value and delay duration of historical late events, input the maximum waitable time window into the value prediction model, and output the expected risk recovery value. When the expected risk recovery value is greater than the expected value loss tolerance, the event cutoff time threshold is generated based on the current system time and the maximum waitable time window. When the expected risk recovery value is less than or equal to the expected value loss tolerance, the current system time is used as the event cutoff time threshold.

3. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 1, characterized in that, The processing module is specifically used for: The event timestamp of each event in the real-time event stream is compared with the event truncation time threshold; If the event timestamp is not later than the event truncation time threshold, the corresponding event is marked as a timely event and given full weight. If the event timestamp is later than the event truncation time threshold, the corresponding event is marked as a late event, and the late time difference between it and the event truncation time threshold is calculated. The time value decay rate of the corresponding late event is calculated by using a time value discounting function that maps the lateness time difference to a value discount. The feature value of the late event is discounted using the time value decay rate to obtain the discounted feature.

4. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 3, characterized in that, The processing module is also used for: Obtain a feature influence weight matrix to characterize the differentiated impact weights of different features on risk assessment results under different degrees of lateness; The time-series calibration feature vector is generated by weighting the feature values ​​of the on-time event, the discounted features, and the feature influence weight matrix. The introduction of the feature influence weight matrix enables the time-series calibration feature vector to synergistically reflect the combined impact of the event's intrinsic feature values ​​and time-series value on the final risk.

5. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 2, characterized in that, The system also includes a model feedback optimization module, used for: Monitor the accuracy of the risk assessment results output by the random forest model within a specified time window; When the accuracy metric falls below its baseline, the event samples accumulated within this time window are used as incremental training data to dynamically adjust the internal parameters of the value prediction model, thereby enabling the value prediction model to adapt to changing risk patterns in assessing the expected risk recovery value.

6. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 5, characterized in that, The model feedback optimization module is also used for: While dynamically adjusting the internal parameters of the value prediction model, the decay coefficient of the time value discount function is also dynamically adjusted in sync with the direction and magnitude of the change in the accuracy index. This sync adjustment enables the value assessment and value discount processes to be optimized in a coordinated manner.

7. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 3, characterized in that, The time value discount function is a non-linear discount function. The time value decay rate output by the function is negatively correlated with the time difference of late arrival, and the rate of value decay changes dynamically as the time difference of late arrival increases.

8. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 1, characterized in that, The fusion module is specifically used for: The fusion of the time-series calibration feature vector and the user historical feature snapshot is completed in memory by performing a sparse matrix addition operation that quickly updates the index positions of non-zero feature values ​​in the user historical feature snapshot.

9. The consumer finance credit risk assessment system based on the random forest algorithm according to claim 1, characterized in that, The evaluation module is specifically used for: The risk assessment result is obtained by simultaneously distributing the fused feature vector to multiple decision trees within the random forest model for parallel computation and aggregating the outputs of all decision trees.

10. A consumer finance credit risk assessment method based on the random forest algorithm, characterized in that, Includes the following steps: Obtain the remaining computing time budget and the stacking density of the real-time event queue of the current computing node, generate system status parameters, and access the real-time event stream carrying event timestamps and arrival timestamps and user historical feature snapshots. Based on the system state parameters, resource constraints are calculated, and the expected value loss tolerance, which characterizes the acceptable risk loss of the business, is obtained. Based on the resource constraints and the expected value loss tolerance, an event cutoff time threshold is generated by evaluating the expected risk recovery value that can be recovered by waiting for the real-time event stream under the premise of satisfying the resource constraints. The event truncation time threshold is used to perform time sequence determination and value decay processing on events in the real-time event stream, generating a time sequence calibration feature vector. The time delay weight in the real-time event stream is determined by the real-time feature vector carrying the time decay weight. The time-series calibration feature vector is fused with the user's historical feature snapshot to obtain a fused feature vector; The fused feature vector is input into the random forest model, and the risk assessment result corresponding to the real-time event stream is output.

Images