Risk control

Abstract

A risk control method. The method comprises: determining each user to be subjected to risk control, and obtaining user data of said users; on the basis of specified attributes of the users to be subjected to risk control, grouping said users to obtain each user group, separately inputting user data corresponding to the user groups into a risk control model, and obtaining first features corresponding to the user groups; inputting the user data of the users to be subjected to risk control into the risk control model, and obtaining initial features corresponding to said users; on the basis of attribute intervals corresponding to the specified attributes, performing feature scaling on the initial features, and obtaining second features corresponding to the users to be subjected to risk control; fusing the obtained first features and second features to obtain fused features; inputting the fused features into the risk control model, and obtaining a risk control result outputted by the risk control model; and, on the basis of the risk control result, performing risk control on the users to be subjected to risk control.

Description

Risk Control Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a risk control method, device, storage medium, and electronic device. Background Art

[0002] With the development of science and technology, privacy data has attracted increasing attention from the public, and artificial intelligence technology has developed rapidly. Among them, machine learning technology is widely used. In the field of risk control, the platform can use machine learning models to conduct risk control on users to ensure the health and sustainable development of the platform.

[0003] Typically, during model training, sample data can be bucketed, also known as binned, based on features across different dimensions. Alternatively, users can be grouped based on their various attributes to discretize features, thereby enhancing feature interpretability, improving model performance, and preventing overfitting. For example, when conducting risk management on users of a platform, a model can be trained based on user data on the platform. User data on the platform can include multiple feature dimensions, such as age and gender for personal information and transaction time and amount for transaction data. In other words, users can be grouped based on age, e.g., 0-30 years old is bin 1, 31-60 years old is bin 2, and 61-100 years old is bin 3. Models can then be trained based on these binned data, allowing them to learn the relationship between age and user risk or the degree of risk.

[0004] However, grouping the features of sample data will lose the continuity information of the features. Therefore, how to ensure that the continuity information of the features is not lost while grouping the features is a difficult problem.

[0005] Based on this, this application specification provides a risk control method.

[0006] Summary of the Invention

[0007] This specification provides a risk control method, device, storage medium and electronic device to at least partially solve the above-mentioned problems existing in the prior art.

[0008] This manual adopts the following technical solutions.

[0009] This specification provides a risk control method, which includes: determining each user to be risk controlled and obtaining user data of each user to be risk controlled; grouping each user to be risk controlled according to specified attributes of each user to be risk controlled to obtain each user group, and inputting the user data corresponding to each user group into a risk control model to obtain a first feature corresponding to each user group; inputting the user data of each user to be risk controlled into the risk control model to obtain an initial feature corresponding to each user to be risk controlled; performing feature scaling on the initial feature according to the attribute interval corresponding to the specified attribute to obtain a second feature corresponding to each user to be risk controlled; fusing the obtained first features and the second features to obtain a fused feature; inputting the fused feature into the risk control model to obtain a risk control result output by the risk control model; and performing risk control on each user to be risk controlled based on the risk control result.

[0010] Optionally, the initial features are feature scaled to obtain the second features corresponding to the users to be risk controlled, specifically including: determining the standard value of the initial features corresponding to the users to be risk controlled based on the attribute interval and the initial features; for each user to be risk controlled, obtaining the second feature corresponding to the user to be risk controlled based on the standard value of the initial features corresponding to the user to be risk controlled and the initial features corresponding to the user to be risk controlled.

[0011] Optionally, determining the standard value of the initial feature corresponding to each of the users to be risk controlled specifically includes: normalizing the initial feature based on a Min-Max normalization method to obtain the standard value of the initial feature corresponding to each of the users to be risk controlled.

[0012] Optionally, based on the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled, the second feature corresponding to the user to be risk controlled is obtained, specifically including: taking the product of the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled as the second feature corresponding to the user to be risk controlled.

[0013] Optionally, the obtained first features and the second features are fused to obtain a fused feature, specifically including: for each user to be risk controlled, determining the user group to which the user to be risk controlled belongs, and determining the third feature corresponding to the user to be risk controlled based on the first feature corresponding to the user group to which the user to be risk controlled belongs; fusing the third feature corresponding to the user to be risk controlled and the second feature corresponding to the user to be risk controlled to obtain the fused feature of the user to be risk controlled.

[0014] Optionally, the risk control model is trained using the following method: obtaining sample user data of each sample user and risk control labels corresponding to the sample user data; grouping the sample users to obtain sample user groups, and inputting the sample user data corresponding to each sample user group into the risk control model to be trained respectively, to obtain the first sample features corresponding to the sample user groups output by the risk control model to be trained; inputting the sample user data of each sample user into the risk control model to be trained to obtain the initial sample features corresponding to the sample users, and performing feature scaling on the initial sample features to obtain the second sample features corresponding to the sample users; fusing the obtained first sample features and the second sample features to obtain fused sample features; inputting the fused sample features into the risk control model to be trained to obtain the prediction results output by the risk control model to be trained; and training the risk control model to be trained based on the prediction results and the risk control labels.

[0015] This specification provides a risk control device, including: an acquisition module, used to determine each user to be risk controlled and obtain user data of each user to be risk controlled; a first input module, used to group each user to be risk controlled according to specified attributes of each user to be risk controlled, to obtain each user group, and input the user data corresponding to each user group into a risk control model to obtain a first feature corresponding to each user group; a second input module, used to input the user data of each user to be risk controlled into the risk control model to obtain an initial feature corresponding to each user to be risk controlled; a feature scaling module, used to perform feature scaling on the initial feature according to an attribute interval corresponding to the specified attribute, to obtain a second feature corresponding to each user to be risk controlled; a feature fusion module, used to fuse the obtained first features and the second features to obtain a fused feature; an output module, used to input the fused feature into the risk control model to obtain a risk control result output by the risk control model; and a risk control module, used to perform risk control on each user to be risk controlled based on the risk control result.

[0016] Optionally, the feature scaling module is specifically used to determine the standard value of the initial feature corresponding to each user to be risk controlled based on the attribute interval and the initial feature; for each user to be risk controlled, the second feature corresponding to the user to be risk controlled is obtained based on the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled.

[0017] Optionally, the feature scaling module is specifically configured to normalize the initial features based on a Min-Max normalization method to obtain standard values ​​of the initial features corresponding to each user to be risk controlled.

[0018] Optionally, the feature scaling module is specifically configured to use a product of a standard value of an initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled as the second feature corresponding to the user to be risk controlled.

[0019] Optionally, the feature fusion module is specifically used to determine, for each user to be risk controlled, the user group to which the user to be risk controlled belongs, and determine the third feature corresponding to the user to be risk controlled based on the first feature corresponding to the user group to which the user to be risk controlled belongs; and fuse the third feature corresponding to the user to be risk controlled and the second feature corresponding to the user to be risk controlled to obtain the fused feature of the user to be risk controlled.

[0020] Optionally, the device also includes a training module; the training module is specifically used to obtain sample user data of each sample user and risk control labels corresponding to the sample user data; group the sample users to obtain sample user groups, input the sample user data corresponding to each sample user group into the risk control model to be trained, and obtain the first sample features corresponding to each sample user group output by the risk control model to be trained; input the sample user data of each sample user into the risk control model to be trained to obtain the initial sample features corresponding to each sample user, perform feature scaling on the initial sample features to obtain the second sample features corresponding to each sample user; fuse the obtained first sample features and the second sample features to obtain fused sample features; input the fused sample features into the risk control model to be trained to obtain the prediction results output by the risk control model to be trained; train the risk control model to be trained based on the prediction results and the risk control labels.

[0021] This specification provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned risk control method is implemented.

[0022] This specification provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned risk control method when executing the program.

[0023] At least one of the above technical solutions adopted in this specification can achieve the following beneficial effects.

[0024] It can be seen from the risk control method provided in this specification that when risk control is performed on each user to be risk controlled, each user to be risk controlled can first be grouped and processed, and based on the user's specified attributes, the user data corresponding to each user group can be input into the risk control model to obtain the first feature of each user group. Then, the data of each user to be risk controlled can be input into the risk control model to obtain the initial feature of each user to be risk controlled, and the initial feature can be scaled based on the attribute interval corresponding to the user's specified attribute to obtain the second feature. Finally, the first feature and the second feature can be fused to obtain a fused feature, and the fused feature can be input into the risk control model to obtain a risk control result, so as to perform risk control on each user to be risk controlled based on the risk control result. This method groups each user, obtains the features of each user under its corresponding group, and at the same time scales the initial features of the user data corresponding to each user, and fuses them with the features of each user under its corresponding group, so as to adjust the discrete degree of the features of the grouped users, correct the expression of the importance of the features, and improve the accuracy of the results output by the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide further understanding of this specification and constitute a part of this specification. The illustrative embodiments of this specification and their descriptions are used to explain this specification and do not constitute improper limitations on this specification.

[0026] FIG1 is a flow chart of a risk control method in this specification.

[0027] FIG2a is a schematic diagram of the continuity representation of information provided in this specification.

[0028] FIG2 b is a schematic diagram of continuity representation of grouped information provided in this specification.

[0029] FIG3 is a schematic diagram of a wind control device provided in this specification.

[0030] FIG4 is a schematic diagram of an electronic device corresponding to FIG1 provided in this specification. DETAILED DESCRIPTION

[0031] To make the purpose, technical solutions, and advantages of this specification more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the specific embodiments of this specification and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0032] The following describes in detail the technical solutions provided by various embodiments of the present application in conjunction with the accompanying drawings.

[0033] FIG1 is a flow chart of a risk control method provided in this specification, which may specifically include the following steps.

[0034] S100: Determine each user to be subject to risk control, and obtain user data of each user to be subject to risk control.

[0035] The execution entity of the technical solution of this specification can be any electronic device with computing capabilities, such as a server, terminal, etc.

[0036] When using a risk control model to conduct risk control on users on a platform, the computing device may first identify users to be risk controlled. These users may be selected from users currently performing business on the platform, or may be users identified based on historical business data on the platform. Furthermore, the computing device may obtain user data for each user to be risk controlled. This user data may be structured data, such as tabular data, and may include: the user's personal information, which may specifically include age, gender, etc.; the user's transaction data, which may specifically include transaction time, transaction amount, and number of transactions, etc.; and other user behavior data, which may specifically include items browsed and browsing time on the platform, etc.

[0037] S102: Grouping the users to be risk controlled according to designated attributes of the users to be risk controlled to obtain user groups, inputting user data corresponding to the user groups into a risk control model to obtain first features corresponding to the user groups.

[0038] In order to improve the accuracy of the risk control results output by the risk control model, enhance the interpretability of features, and avoid overfitting of the risk control model, the computing device can group the users to be risk controlled to obtain user groups. Specifically, the computing device can group the users to be risk controlled according to the specified attributes of the users to be risk controlled, where the user attributes can be the user's age, transaction time, transaction amount, etc., which are not limited in this specification. For example, the users to be risk controlled are user A, user B, user C, and user D, and the user data are the age, gender, transaction amount, number of transactions, and transaction amount of users A to D. Users A to D can be grouped based on the attribute of age.

[0039] It should be noted that when grouping users to be risk-controlled, there is no specific limit on the number of specified attributes and it can be set according to specific needs and business scenarios. Continuing with the above example, users A to D can be grouped based on age and gender.

[0040] The computing device may then input the user data corresponding to each group and each user group into the risk control model respectively to obtain the first feature corresponding to each user group.

[0041] S104: Inputting the user data of each user to be risk-controlled into the risk control model to obtain initial features corresponding to each user to be risk-controlled.

[0042] S106: Perform feature scaling on the initial feature according to the attribute interval corresponding to the designated attribute to obtain a second feature corresponding to each of the users to be risk controlled.

[0043] If users are simply grouped and prediction results are obtained based on the characteristics of each user group, the continuity information of the characteristics of the user data will be lost, so that the risk control model lacks the expression of the importance of the continuity information of the characteristics. As shown in Figure 2a, it is a schematic diagram of the continuity representation of the information provided in the specification of this application, and as shown in Figure 2b, it is a schematic diagram of the continuity representation of the grouped information provided in the specification of this application. It can be seen that when users are grouped, each user group has the continuity of its own characteristics, and on the whole, the continuity information of the information between the user groups is lost. Therefore, the computing device can input the user data of each user to be risk controlled into the risk control model to obtain the initial characteristics corresponding to each user to be risk controlled, and can scale the initial characteristics according to the attribute interval corresponding to the specified attribute to obtain the second characteristic, and fuse the first characteristic and the second characteristic in the subsequent steps to eliminate the loss of continuity information caused by grouping, thereby improving the accuracy of the risk control results.

[0044] Specifically, when performing feature scaling on the initial features, the computing device may normalize the initial features corresponding to each user to be risk controlled based on the attribute interval and the initial features to obtain a standard value for the initial features corresponding to each user to be risk controlled. Furthermore, for each user to be risk controlled, the computing device may obtain a second feature corresponding to the user to be risk controlled based on the standard value of the initial features corresponding to the user to be risk controlled and the initial features corresponding to the user to be risk controlled. In one or more embodiments of this specification, the product of the standard value of the initial features corresponding to the user to be risk controlled and the initial features corresponding to the user to be risk controlled may be used as the second feature corresponding to the user to be risk controlled.

[0045] It should be noted that the normalization method used when normalizing the initial features corresponding to each user to be risk controlled, or the method used when scaling the features, is not specifically limited in this specification. For example, the Min-Max normalization method, the Scale to [-1, 1] normalization method, and the Gauss Rank normalization method, etc., can be used. That is, in one or more embodiments of this specification, the computing device can normalize the initial features corresponding to each user to be risk controlled based on the Min-Max normalization method to obtain the standard value of the initial features corresponding to each user to be risk controlled, or can normalize the initial features corresponding to each user to be risk controlled based on the Gauss Rank normalization method to obtain the standard value of the initial features corresponding to each user to be risk controlled. In other words, in one or more embodiments of this specification, the computing device can scale the initial features based on the Min-Max normalization method, or can scale the initial features based on the Gauss Rank normalization method.

[0046] S108: Fusing the obtained first features and the second features to obtain a fused feature.

[0047] Since the second feature obtained in step S106 is obtained after feature scaling of the features of the user data of all users to be risk controlled, that is, each second feature representation is a measure of the continuity of the initial feature corresponding to the user data, or a linear measure, therefore, in this specification, the computing device can fuse each first feature and the second feature to obtain a fused feature, so as to use the continuity of the second feature representation to adjust the discrete degree of the first feature obtained by grouping users in the above step S102, thereby retaining the continuity information of the feature without increasing the number of user groups, thereby improving the accuracy of the prediction result.

[0048] In one or more embodiments of the present specification, when fusing the first and second features, the specific computing device may determine the user group to which the user to be risk controlled belongs for each user to be risk controlled, and determine the third feature corresponding to the user to be risk controlled based on the first feature corresponding to the user group to which the user to be risk controlled belongs. That is, the third feature refers to the feature of the user data corresponding to the user to be risk controlled in the user group to which the user to be risk controlled belongs, or in other words, in the first feature corresponding to the user group to which the user to be risk controlled belongs, the first feature of the user data belonging to the user to be risk controlled is the third feature. Then, the third feature corresponding to the user to be risk controlled and the second feature corresponding to the user to be risk controlled may be fused to obtain the fused feature of the user to be risk controlled. In one or more embodiments of the present specification, the sum of the third feature and the second feature of each user to be risk controlled may be used as the fused feature of the user to be risk controlled.

[0049] S110: Input the fusion feature into the risk control model to obtain the risk control result output by the risk control model.

[0050] S112: Based on the risk control result, risk control is performed on each of the users to be risk controlled.

[0051] Finally, the computing device can input the fusion features into the risk control model to obtain the risk control results of the risk control output, and can perform risk control on each user to be risk controlled based on the risk control results.

[0052] It should be noted that the risk control results can be determined based on the specific scenario and the specific services provided by the platform. For example, the risk control results can be risk control levels, risk control intensity, risk probability, etc. Of course, different risk control levels or risk control intensity correspond to different risk control strategies. For example, if the platform provides payment transaction services, the risk control results can be the risk level of the user's payment transactions, such as high, medium, and low. The corresponding strategies for different levels can include limiting the user's transaction amount, number of transactions, and prohibiting the user from trading.

[0053] Based on the risk control method shown in Figure 1, when performing risk control on each user to be risk controlled, each user can first be grouped based on the user's designated attributes. The user data corresponding to each user group is then input into the risk control model to obtain the first feature of each user group. The data of each user to be risk controlled can then be input into the risk control model to obtain the initial features of each user to be risk controlled. The initial features are then scaled based on the attribute range of the user's designated attributes to obtain the second features. Finally, the first and second features can be fused to obtain a fused feature, which is then input into the risk control model to obtain a risk control result, and risk control can then be performed on each user to be risk controlled based on the risk control result. This method groups each user to obtain the features of each user in their corresponding group. At the same time, the initial features of the user data corresponding to each user are scaled and fused with the features of each user in their corresponding group to adjust the discreteness of the features of the grouped users, correct the importance expression of the features, and improve the accuracy of the results output by the model.

[0054] Typically, when grouping features of structured data, the number of groups needs to be increased to preserve the continuity information of the grouped features. However, a larger number of groups often means an increase in model parameters, more specifically, an increase in embedding layer parameters. This solution, however, standardizes the initial features of the data, such as through Min-Max normalization, to achieve feature scaling of the initial features. It also uses the grouped features as biases to modify the importance expression of the features, incorporating the continuity information of the features into the grouped features. This indirectly adjusts the discreteness of the feature groupings, ensuring that the continuity information of the features is not lost even when the number of groups is small. Experiments have shown that the prediction accuracy of the model obtained by grouping features into 64 groups, or binning 64, based on the method provided by this solution, is higher than that of the traditional model grouping features into 64 bins, reaching the same accuracy as the model grouping features into 256 groups, or binning 256. The number of parameters in the model grouping features into 64 groups is far lower than that in the model grouping features into 256 groups.

[0055] Furthermore, in one or more embodiments of the present specification, a training method for the risk control model is also provided. Specifically, the computing device can obtain the sample user data of each sample user and the risk control annotation corresponding to the sample user data. Then, the sample users are grouped to obtain each sample user group, and the sample user data corresponding to each sample user group is respectively input into the risk control model to be trained to obtain the first sample features corresponding to each sample user group output by the risk control model to be trained, and the sample user data of each sample user is input into the risk control model to be trained to obtain the initial sample features corresponding to each sample user, and the initial sample features are feature scaled to obtain the second sample features corresponding to each sample user. Then, the obtained first sample features and second sample features are fused to obtain fused sample features, and the fused sample features are input into the risk control model to be trained to obtain the prediction results output by the risk control model to be trained. Finally, the risk control model to be trained is trained based on the prediction results and the risk control annotations.

[0056] The risk control label can be determined based on the specific scenario and the specific services provided by the platform. For example, the risk control label can include risk control level, risk control intensity, risk probability, etc. Of course, in actual applications, different risk control levels, risk control intensity, or risk control probability should correspond to different risk control strategies. For example, if the platform provides payment transaction services, the risk control result can be the risk level of the user's payment transaction, such as high, medium, and low. The corresponding strategies for different levels can include limiting the user's transaction amount, number of times, and prohibiting the user from trading, etc.

[0057] Based on the risk control method described above, the embodiment of this specification also provides a schematic diagram of a device for risk control, as shown in FIG3 .

[0058] FIG3 is a schematic diagram of a risk control device provided in an embodiment of the present specification. The device includes: an acquisition module 300 for determining each user to be risk controlled and acquiring user data of each user to be risk controlled; a first input module 302 for grouping each user to be risk controlled according to a specified attribute of each user to be risk controlled to obtain user groups, and inputting the user data corresponding to each user group into a risk control model to obtain a first feature corresponding to each user group; a second input module 304 for inputting the user data of each user to be risk controlled into the risk control model to obtain an initial feature corresponding to each user to be risk controlled; a feature scaling module 306 for scaling the initial feature according to an attribute interval corresponding to the specified attribute to obtain a second feature corresponding to each user to be risk controlled; a feature fusion module 308 for fusing the obtained first features and the second features to obtain a fused feature; an output module 310 for inputting the fused feature into the risk control model to obtain a risk control result output by the risk control model; and a risk control module 312 for performing risk control on each user to be risk controlled based on the risk control result.

[0059] Optionally, the feature scaling module 306 is specifically used to determine the standard value of the initial feature corresponding to each user to be risk controlled based on the attribute interval and the initial feature; for each user to be risk controlled, the second feature corresponding to the user to be risk controlled is obtained based on the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled.

[0060] Optionally, the feature scaling module 306 is specifically configured to normalize the initial features based on a Min-Max normalization method to obtain standard values ​​of the initial features corresponding to the users to be risk controlled.

[0061] Optionally, the feature scaling module 306 is specifically configured to use a product of a standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled as the second feature corresponding to the user to be risk controlled.

[0062] Optionally, the feature fusion module 308 is specifically used to determine, for each user to be risk controlled, the user group to which the user to be risk controlled belongs, and determine the third feature corresponding to the user to be risk controlled based on the first feature corresponding to the user group to which the user to be risk controlled belongs; and fuse the third feature corresponding to the user to be risk controlled and the second feature corresponding to the user to be risk controlled to obtain the fused feature of the user to be risk controlled.

[0063] Optionally, the device also includes a training module 314; the training module 314 is specifically used to obtain sample user data of each sample user and risk control labels corresponding to the sample user data; group the sample users to obtain sample user groups, input the sample user data corresponding to each sample user group into the risk control model to be trained, and obtain the first sample features corresponding to each sample user group output by the risk control model to be trained; input the sample user data of each sample user into the risk control model to be trained to obtain the initial sample features corresponding to each sample user, perform feature scaling on the initial sample features to obtain the second sample features corresponding to each sample user; fuse the obtained first sample features and the second sample features to obtain fused sample features; input the fused sample features into the risk control model to be trained to obtain the prediction results output by the risk control model to be trained; and train the risk control model to be trained based on the prediction results and the risk control labels.

[0064] The embodiments of this specification also provide a computer-readable storage medium, which stores a computer program. The computer program can be used to execute the risk control method described above.

[0065] Based on the risk control methods described above, the embodiments of this specification also provide a schematic diagram of the electronic device shown in Figure 4. As shown in Figure 4, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, memory, and non-volatile storage, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile storage into the memory and then runs it to implement the risk control methods described above.

[0066] Of course, in addition to software implementation, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0067] In the 1990s, technological improvements could be clearly distinguished as either hardware improvements (for example, improvements to circuit structures like diodes, transistors, and switches) or software improvements (improvements to process flows). However, with the advancement of technology, many process flow improvements today can now be considered direct improvements to hardware circuit structures. Designers almost always create the corresponding hardware circuit structure by programming the improved process flow into the hardware circuit. Therefore, it cannot be said that a process flow improvement cannot be implemented using hardware modules. For example, a programmable logic device (PLD), such as a field programmable gate array (FPGA), is an integrated circuit whose logical function is determined by user programming. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to hire a chip manufacturer to design and manufacture a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly performed using "logic compiler" software. This is similar to the software compiler used during program development. Before compilation, the original code must also be written in a specific programming language, called a hardware description language (HDL). There is not just one HDL, but many, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art will also understand that by simply programming the method flow in one of these hardware description languages ​​and then programming it into an integrated circuit, a hardware circuit that implements the logic method flow can be easily obtained.

[0068] The controller can be implemented in any suitable manner. For example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also know that in addition to implementing the controller in a purely computer-readable program code format, the controller can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be considered as structures within the hardware component. Or even, the devices for implementing various functions can be considered as both software modules that implement the method and structures within the hardware component.

[0069] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0070] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0071] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0073] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0075] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0076] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0077] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0078] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0079] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Thus, this specification may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0080] This specification may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0081] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0082] The foregoing is merely an example of the present invention and is not intended to limit the present invention. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A risk control method, the method comprising: Determine each user to be risk-controlled, and obtain user data of each user to be risk-controlled; According to the specified attributes of the users to be risk controlled, the users to be risk controlled are grouped to obtain user groups, and user data corresponding to the user groups are respectively input into the risk control model to obtain the first features corresponding to the user groups; Inputting the user data of each of the users to be risk controlled into the risk control model to obtain initial features corresponding to each of the users to be risk controlled; According to the attribute interval corresponding to the specified attribute, the initial feature is scaled to obtain a second feature corresponding to each of the users to be risk controlled; Fusing the obtained first features and the second features to obtain a fused feature; Inputting the fusion feature into the risk control model to obtain the risk control result output by the risk control model; Based on the risk control result, risk control is performed on each of the users to be risk controlled.

2. The method according to claim 1, performing feature scaling on the initial feature to obtain the second feature corresponding to each of the users to be risk controlled, specifically comprising: Determine, according to the attribute interval and the initial feature, a standard value of the initial feature corresponding to each of the users to be risk controlled; For each user to be risk controlled, a second feature corresponding to the user to be risk controlled is obtained according to a standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled.

3. The method according to claim 2, determining the standard value of the initial feature corresponding to each of the users to be risk controlled, specifically comprising: Based on the Min-Max standardization method, the initial features are normalized to obtain the standard values ​​of the initial features corresponding to the users to be risk controlled.

4. The method according to claim 2, obtaining the second feature corresponding to the user to be risk controlled according to the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled, specifically comprising: The product of the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled is used as the second feature corresponding to the user to be risk controlled.

5. The method according to claim 1, fusing the obtained first features and the second features to obtain a fused feature, specifically comprising: For each user to be risk controlled, determine the user group to which the user to be risk controlled belongs, and determine the third feature corresponding to the user to be risk controlled according to the first feature corresponding to the user group to which the user to be risk controlled belongs; The third feature corresponding to the user to be risk controlled and the second feature corresponding to the user to be risk controlled are combined to obtain The integrated characteristics of the user to be risk controlled.

6. The method according to claim 1, wherein the risk control model is trained by the following method: Obtaining sample user data of each sample user and risk control annotations corresponding to the sample user data; The sample users are grouped to obtain sample user groups, and the sample user data corresponding to the sample user groups are respectively input into the risk control model to be trained to obtain the first sample features corresponding to the sample user groups output by the risk control model to be trained; Inputting the sample user data of each sample user into the risk control model to be trained to obtain initial sample features corresponding to each sample user, and performing feature scaling on the initial sample features to obtain second sample features corresponding to each sample user; Fusing the obtained first sample features and the second sample features to obtain a fused sample feature; Inputting the fused sample features into the risk control model to be trained to obtain a prediction result output by the risk control model to be trained; Based on the prediction result and the risk control annotation, the risk control model to be trained is trained.

7. A wind control device, the device specifically comprising: An acquisition module is used to determine each user to be risk-controlled and obtain user data of each user to be risk-controlled; A first input module, configured to group the users to be risk controlled according to the specified attributes of the users to be risk controlled to obtain user groups, and input the user data corresponding to the user groups into the risk control model to obtain the first features corresponding to the user groups; A second input module is used to input the user data of each user to be risk controlled into the risk control model to obtain the initial features corresponding to each user to be risk controlled; A feature scaling module, used to perform feature scaling on the initial feature according to the attribute interval corresponding to the specified attribute, to obtain a second feature corresponding to each of the users to be risk controlled; A feature fusion module, used for fusing the obtained first features and the second features to obtain a fused feature; An output module, used to input the fusion feature into the risk control model to obtain the risk control result output by the risk control model; The risk control module is used to perform risk control on each of the users to be risk controlled based on the risk control results.

8. In the device as described in claim 7, the feature scaling module is specifically used to determine the standard value of the initial feature corresponding to each user to be risk controlled according to the attribute interval and the initial feature; for each user to be risk controlled, the second feature corresponding to the user to be risk controlled is obtained according to the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled.

9. In the device as described in claim 8, the feature scaling module is specifically used to normalize the initial features based on the Min-Max normalization method to obtain the standard values ​​of the initial features corresponding to each of the users to be risk controlled.

10. The device according to claim 8, wherein the feature scaling module is specifically used to use the product of the standard value of the initial feature corresponding to the user to be risk controlled and the initial feature corresponding to the user to be risk controlled as the second feature corresponding to the user to be risk controlled.

11. In the device as described in claim 7, the feature fusion module is specifically used to determine, for each user to be risk controlled, the user group to which the user to be risk controlled belongs, and determine the third feature corresponding to the user to be risk controlled based on the first feature corresponding to the user group to which the user to be risk controlled belongs; and fuse the third feature corresponding to the user to be risk controlled and the second feature corresponding to the user to be risk controlled to obtain the fused feature of the user to be risk controlled.

12. The apparatus of claim 7, further comprising a training module; The training module is specifically used to obtain sample user data of each sample user and risk control annotations corresponding to the sample user data; group the sample users to obtain sample user groups, input the sample user data corresponding to the sample user groups into the risk control model to be trained, and obtain first sample features corresponding to the sample user groups output by the risk control model to be trained; input the sample user data of each sample user into the risk control model to be trained to obtain initial sample features corresponding to the sample users, perform feature scaling on the initial sample features to obtain second sample features corresponding to the sample users; fuse the obtained first sample features and the second sample features to obtain fused sample features; input the fused sample features into the risk control model to be trained to obtain prediction results output by the risk control model to be trained; and train the risk control model to be trained based on the prediction results and the risk control annotations.

13. A computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

14. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 6 when executing the program.

Images