Acquisition method of input parameters of open account risk prediction model and related device

Abstract

The invention discloses a method for obtaining input parameters of an account opening risk prediction model and a related device, relates to the field of artificial intelligence, and aims to obtain a target rule strategy attribute from rule strategy attributes corresponding to a plurality of clients in a preset blacklist. The proportion of customers with the target rule strategy attribute in the preset blacklist is high, so that the target rule strategy attribute is a parameter influencing the account opening risk; and determining the degree of the target rule policy attribute influencing the account opening risk, namely the policy risk level. And taking the target rule strategy attribute and the strategy risk level corresponding to the target rule strategy attribute as partial parameters in the input parameters of the account opening account risk prediction model. Therefore, the input parameters of the model are continuously enriched, so that the output accuracy of the model is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method for obtaining input parameters of an account opening risk prediction model and a related device. Background Art

[0002] With the rapid development and digital transformation of the financial industry, the complexity of banking business and transaction risks are increasing. In order to reduce transaction risks, it is necessary to prohibit customers with higher risks from opening accounts at the source, that is, when opening an account, which reduces transaction risks to a certain extent.

[0003] In the related art, the risk of opening an account can be predicted based on a machine learning model. However, the input parameters of the machine learning model are too dependent on expert experience, resulting in low accuracy of the machine learning model.

[0004] Based on this, how to obtain input parameters that make the machine learning model more accurate is a technical problem that urgently needs to be solved in this field. Summary of the invention

[0005] In view of the above problems, the present application provides a method and a related device for obtaining input parameters of an account opening risk prediction model, so as to achieve the purpose of obtaining input parameters that make the machine learning model more accurate. The specific scheme is as follows:

[0006] The first aspect of the present application provides a method for obtaining input parameters of an account opening risk prediction model, comprising:

[0007] From the rule policy attributes respectively corresponding to the multiple customers included in the preset blacklist, obtain the target rule policy attribute that does not belong to the risk attribute set and whose proportion is greater than or equal to the preset probability; the proportion of the rule policy attribute refers to the ratio of the number of customers with the rule policy attribute in the preset blacklist to the number of all customers included in the preset blacklist; the risk attribute set includes the rule policy attribute and the policy risk level corresponding to the rule policy attribute;

[0008] Determine, from a plurality of preset indicator layer attributes, a target indicator layer attribute associated with the target rule strategy attribute;

[0009] Determine the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute;

[0010] Acquire multiple candidate policy risk levels corresponding to the target rule policy attribute; the minimum value of the multiple candidate policy risk levels is the minimum policy risk level;

[0011] Acquire the policy risk level corresponding to the target rule policy attribute from the plurality of candidate policy risk levels;

[0012] The target rule policy attributes and the policy risk levels corresponding to the target rule policy attributes are divided into the risk attribute set; the risk attribute set is the input of a preset account opening risk prediction model.

[0013] A second aspect of the present application provides a device for obtaining input parameters of an account opening risk prediction model, comprising:

[0014] The first acquisition module is used to obtain a target rule policy attribute that does not belong to the risk attribute set and has a proportion greater than or equal to a preset probability from the rule policy attributes corresponding to multiple customers included in the preset blacklist; the proportion of the rule policy attribute refers to the ratio of the number of customers with the rule policy attribute in the preset blacklist to the number of all customers included in the preset blacklist; the risk attribute set includes the rule policy attribute and the policy risk level corresponding to the rule policy attribute;

[0015] A first determination module is used to determine a target indicator layer attribute having an associated relationship with the target rule strategy attribute from a plurality of preset indicator layer attributes;

[0016] A second determination module is used to determine the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute;

[0017] A second acquisition module is used to acquire a plurality of candidate policy risk levels corresponding to the target rule policy attribute; the minimum value of the plurality of candidate policy risk levels is the minimum policy risk level;

[0018] A third acquisition module, configured to acquire the policy risk level corresponding to the target rule policy attribute from the plurality of candidate policy risk levels;

[0019] A division module is used to divide the target rule policy attribute and the policy risk level corresponding to the target rule policy attribute into the risk attribute set; the risk attribute set is the input of the preset account opening risk prediction model.

[0020] The third aspect of the present application provides a computer program product, including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements the method for obtaining the input parameters of the account opening risk prediction model of the first aspect or any implementation of the first aspect.

[0021] A fourth aspect of the present application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0022] The memory is used to store computer programs;

[0023] The processor is used to execute the computer program so that the electronic device can implement the method for obtaining the input parameters of the account opening risk prediction model of the first aspect or any implementation of the first aspect.

[0024] The fifth aspect of the present application provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can obtain the input parameters of the account opening risk prediction model of the above-mentioned first aspect or any implementation of the first aspect.

[0025] Through the above technical scheme, the present application provides a method for obtaining the input parameters of an account opening risk prediction model. With the development of technology, the parameters that affect the risk of an account opening are constantly changing. If the input parameters of the account opening risk prediction model, that is, the number of parameters included in the risk attribute set, remain unchanged, the accuracy of the output result of the account opening risk prediction model will be reduced; based on this, the embodiment of the present application obtains the target rule policy attributes that do not belong to the risk attribute set and have a proportion greater than or equal to the preset probability from the rule policy attributes corresponding to multiple customers included in the preset blacklist. Since the proportion of customers with the target rule policy attribute in the preset blacklist is relatively high, it means that the target rule policy attribute is a parameter that affects the account opening risk; however, it is necessary to further determine the degree to which the target rule policy attribute affects the account opening risk, that is, the policy risk level. The specific process includes: determining the target indicator layer attribute that has an association relationship with the target rule policy attribute from the preset multiple indicator layer attributes; determining the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute; obtaining multiple candidate policy risk levels corresponding to the target rule policy attribute; the minimum value of the multiple candidate policy risk levels is the minimum policy risk level; obtaining the policy risk level corresponding to the target rule policy attribute from the multiple candidate policy risk levels. The target rule policy attribute and the policy risk level corresponding to the target rule policy attribute are divided into the risk attribute set. In this way, the input parameters of the account opening risk prediction model are continuously enriched, thereby improving the accuracy of the output of the account opening risk prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0027] Figure 1 A schematic diagram of a system architecture provided for this application;

[0028] Figure 2 A schematic diagram of an optional hardware structure of a terminal 100 provided in this application;

[0029] Figure 3 A schematic diagram of the structure of a server 200 provided in this application;

[0030] Figure 4 A flowchart of a method for obtaining input parameters of an account opening risk prediction model provided in an embodiment of the present application;

[0031] Figure 5 A schematic diagram of the distribution effect of the account opening risk prediction model provided in the embodiment of the present application under different candidate strategy risk levels;

[0032] Figure 6 A comparison chart of the recall rates of the account opening risk prediction models provided in the embodiments of the present application;

[0033] Figure 7 A schematic diagram of a hierarchical structure of multiple risk rule attributes constructed using AHP provided in an embodiment of the present application;

[0034] Figure 8 A schematic diagram of a first judgment matrix including multiple risk rule attributes provided in an embodiment of the present application;

[0035] Fig. 9 For the general Figure 8 A schematic diagram of a second judgment matrix obtained after normalizing the first judgment matrix and a consistency test result is shown;

[0036] Fig.10 Schematic diagram of three third judgment matrices provided in the embodiments of the present application;

[0037] Fig.11 The embodiments of the present application provide Fig.10 A schematic diagram of the normalized fourth judgment matrix of the three third judgment matrices in FIG.

[0038] Fig.12 A schematic diagram of the structure of a device for obtaining input parameters of an account opening risk prediction model provided in an embodiment of the present application;

[0039] Fig.13 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0040] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0041] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0042] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0043] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0044] See also Figure 1 , Figure 1 1 shows a schematic diagram of a system architecture. The system may include a terminal 100 and a server 200. The server 200 may include one or more servers ( Figure 1 In the example, a server is included, and the server 200 can provide the method provided in the embodiment of the present application for one or more terminals.

[0045] Among them, an application can be installed on the terminal 100, and the above application can provide an interface. The terminal 100 can receive relevant parameters entered by the user on the interface, such as setting the interval time for executing the "method for obtaining the input parameters of the account opening risk prediction model", and sending the above parameters to the server 200. The server 200 can obtain the processing results based on the received parameters and return the processing results to the terminal 100.

[0046] It should be understood that in some optional implementations, the terminal 100 can also complete the action of obtaining the processing result based on the received parameters by itself without the cooperation of the server, and the embodiments of the present application are not limited to this.

[0047] Next describe Figure 1 The product form of the mid-terminal 100;

[0048] The terminal 100 in the embodiment of the present application can be a mobile phone, a tablet computer, a wearable device, a vehicle-mounted device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), etc., and the embodiment of the present application does not impose any limitation on this.

[0049] Figure 2 An optional hardware structure diagram of the terminal 100 is shown.

[0050] refer to Figure 2 As shown, the terminal 100 may include a radio frequency unit 110, a memory 120, an input unit 130, a display unit 140, a camera 150 (optional), an audio circuit 160 (optional), a speaker 161 (optional), a microphone 162 (optional), an earphone jack 163 (optional), a processor 170, an external interface 180, a power supply 190 and other components. Those skilled in the art will appreciate that Figure 2 This is merely an example of a terminal and does not constitute a limitation on the terminal. The terminal may include more or fewer components than shown in the figure, or a combination of certain components, or different components.

[0051] The input unit 130 can be used to receive input digital or character information, and generate key signal input related to the user settings and function control of the terminal. Specifically, the input unit 130 may include a touch screen 131 (optional) and / or other input devices 132. The touch screen 131 can collect the user's touch operations on or near it (such as the user's operation on or near the touch screen using any suitable object such as fingers, joints, stylus, etc.), and drive the corresponding connection device according to a pre-set program. The touch screen can detect the user's touch action on the touch screen, convert the touch action into a touch signal and send it to the processor 170, and can receive and execute the command sent by the processor 170; the touch signal at least includes the touch point coordinate information. The touch screen 131 can provide an input interface and an output interface between the terminal 100 and the user. In addition, the touch screen can be implemented using multiple types such as resistive, capacitive, infrared and surface acoustic wave. In addition to the touch screen 131, the input unit 130 can also include other input devices. Specifically, other input devices 132 may include, but are not limited to, one or more of a physical keyboard, function keys (such as a volume control key, a switch key, etc.), a trackball, a mouse, a joystick, and the like.

[0052] Among them, the input device 132 can receive input data and the like.

[0053] The display unit 140 may be used to display information input by the user or provided to the user, various menus of the terminal 100, interactive interfaces, file display and / or playback of any multimedia file. In the embodiment of the present application, the display unit 140 may be used to display interfaces, processing results, etc.

[0054] The memory 120 can be used to store instructions and data. The memory 120 can mainly include an instruction storage area and a data storage area. The data storage area can store various data, such as multimedia files, texts, etc.; the instruction storage area can store software units such as operating systems, applications, instructions required for at least one function, or their subsets and extensions. It can also include a non-volatile random access memory; provide the processor 170 with hardware, software and data resources including management of computing and processing equipment, and support control software and applications. It is also used for the storage of multimedia files, and the storage of running programs and applications.

[0055] The processor 170 is the control center of the terminal 100. It uses various interfaces and lines to connect various parts of the entire terminal 100. By running or executing instructions stored in the memory 120 and calling data stored in the memory 120, it executes various functions of the terminal 100 and processes data, thereby controlling the terminal device as a whole. Optionally, the processor 170 may include one or more processing units; preferably, the processor 170 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application program, and the modem processor mainly processes wireless communication. It is understandable that the above-mentioned modem processor may not be integrated into the processor 170. In some embodiments, the processor and the memory may be implemented on a single chip, and in some embodiments, they may also be implemented separately on separate chips. The processor 170 may also be used to generate corresponding operation control signals, send them to corresponding components of the computing and processing device, read and process data in the software, especially read and process data and programs in the memory 120, so that each functional module therein performs corresponding functions, thereby controlling the corresponding components to act according to the requirements of the instructions.

[0056] Among them, the memory 120 can be used to store software codes related to the method for obtaining the input parameters of the account opening risk prediction model, the processor 170 can execute the steps of the method for obtaining the input parameters of the account opening risk prediction model, and can also schedule other units (such as the above-mentioned input unit 130 and the display unit 140) to realize corresponding functions.

[0057] The radio frequency unit 110 (optional) can be used for receiving and sending information or receiving and sending signals during a call, for example, after receiving the downlink information of the base station, it is sent to the processor 170 for processing; in addition, the designed uplink data is sent to the base station. Generally, the RF circuit includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (Low Noise Amplifier, LNA), a duplexer, etc. In addition, the radio frequency unit 110 can also communicate with network devices and other devices through wireless communication. The wireless communication can use any communication standard or protocol, including but not limited to Global System of Mobile communication (Global System of Mobile communication, GSM), General Packet Radio Service (General Packet Radio Service, GPRS), Code Division Multiple Access (Code Division Multiple Access, CDMA), Wideband Code Division Multiple Access (Wideband Code Division Multiple Access, WCDMA), Long Term Evolution (Long Term Evolution, LTE), email, Short Messaging Service (SMS), etc.

[0058] In this embodiment of the present application, the RF unit 110 can send data to the server 200 and receive processing results sent by the server 200.

[0059] It should be understood that the radio frequency unit 110 is optional and can be replaced by other communication interfaces, such as a network port.

[0060] The terminal 100 also includes a power supply 190 (such as a battery) for supplying power to various components. Preferably, the power supply can be logically connected to the processor 170 through a power management system, so that the power management system can manage functions such as charging, discharging, and power consumption.

[0061] The terminal 100 further includes an external interface 180 , which may be a standard Micro USB interface or a multi-pin connector, and may be used to connect the terminal 100 to communicate with other devices, or to connect a charger to charge the terminal 100 .

[0062] Although not shown, the terminal 100 may also include a flashlight, a wireless fidelity (WiFi) module, a Bluetooth module, sensors with different functions, etc., which are not described in detail here. Some or all of the methods described below may be applied in the following embodiments. Figure 2 In the terminal 100 shown.

[0063] Next describe Figure 1 The product form of the server 200;

[0064] Figure 3 A structural diagram of a server 200 is provided, such as Figure 3 As shown, the server 200 includes a bus 201, a processor 202, a communication interface 203, and a memory 204. The processor 202, the memory 204, and the communication interface 203 communicate with each other via the bus 201.

[0065] The bus 201 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0066] The processor 202 may be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0067] The memory 204 may include a volatile memory, such as a random access memory (RAM). The memory 204 may also include a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0068] Among them, the memory 204 can be used to store software codes related to the method for obtaining the input parameters of the account opening risk prediction model, the processor 202 can execute the steps of the method for obtaining the input parameters of the account opening risk prediction model of the chip, and can also schedule other units to implement corresponding functions.

[0069] It should be understood that the above-mentioned terminal 100 and server 200 can be centralized or distributed devices, and the processors in the above-mentioned terminal 100 and server 200 (such as processor 170 and processor 202) can be hardware circuits (such as application specific integrated circuit (ASIC), field-programmable gate array (FPGA), general-purpose processor, digital signal processor (DSP), microprocessor or microcontroller, etc.), or a combination of these hardware circuits. For example, the processor can be a hardware system with an instruction execution function, such as a CPU, DSP, etc., or a hardware system without an instruction execution function, such as an ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without an instruction execution function and hardware systems with an instruction execution function.

[0070] Reference Figure 4 , Figure 4 A flow chart of a method for obtaining input parameters of an account opening risk prediction model provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, a method for obtaining input parameters of an account opening risk prediction model provided in an embodiment of the present application may include steps S401 to S406, and these steps are described in detail below.

[0071] Step S401: From the rule policy attributes corresponding to the plurality of customers included in the preset blacklist, a target rule policy attribute that does not belong to the risk attribute set and has a proportion greater than or equal to a preset probability is obtained.

[0072] The proportion of the rule policy attribute refers to the ratio of the number of customers with the rule policy attribute in the preset blacklist to the number of all customers included in the preset blacklist; the risk attribute set includes the rule policy attribute and the policy risk level corresponding to the rule policy attribute.

[0073] Exemplarily, the preset probability may be determined based on actual conditions, which is not limited here.

[0074] Exemplarily, the customers included in the preset blacklist are all account opening risk customers with account opening risks.

[0075] Exemplarily, the preset blacklist may be a blacklist obtained within a set time period close to the current time, or the preset blacklist may be a blacklist obtained from the initial time of establishing the blacklist to the current time.

[0076] The following is an explanation of the rule policy attributes. Rule policy attributes refer to the attributes that affect account opening risks, such as the mobile phone number unreal-name policy attribute and the mobile phone number out-of-region policy attribute.

[0077] Exemplarily, the rule policy attributes included in the risk attribute set and the policy risk levels corresponding to the rule policy attributes are obtained through steps S401 to S406. When steps S401 to S406 are performed for the first time, the risk attribute set includes 0 rule policy attributes.

[0078] Exemplarily, the rule policy attributes included in the risk attribute set are obtained by experienced personnel based on their own experience.

[0079] Exemplarily, since the policy risk levels of the rule policy attributes included in the risk attribute set are known, the obtained target rule policy attribute does not belong to the risk attribute set, because there is no need to obtain the policy risk level of the known rule attribute information again.

[0080] It can be understood that the larger the proportion corresponding to the rule strategy attribute is, the greater the probability that the rule strategy attribute affects the customer to become a risky customer for opening an account, and the more likely the rule strategy attribute is to become an input parameter for predicting the risk of opening an account.

[0081] Step S402: Determine a target indicator layer attribute associated with the target rule policy attribute from a plurality of preset indicator layer attributes.

[0082] Exemplarily, the indicator layer attributes are obtained by experts with rich experience based on experience. Exemplarily, the indicator layer attributes are attributes that affect account opening risks. Exemplarily, the indicator layer attributes include but are not limited to: abnormal transaction behavior, abnormal login behavior, abnormal binding account, legal record verification, abnormal transfer of funds compliance, real-name system compliance, high-risk areas, frequency of use in different places, and frequent address changes. The indicator layer attributes are described in detail below.

[0083] Abnormal trading behavior refers to abnormal behavior that violates the principles of fairness and justice during the trading process and affects the market order. For example, abnormal trading behavior includes but is not limited to: the number of confirmed abnormal transactions.

[0084] Abnormal login behavior refers to some abnormal login patterns that occur when users log in to the system or application. These login patterns may indicate that the account has been stolen, there are security risks or other problems. Exemplary, abnormal login behavior includes but is not limited to the following 7 abnormal contents: 1. Logging in from an uncommon login location. It is understandable that the user logs in to the account from a new or uncommon location, which may indicate that the account information has been leaked or accessed by an unauthorized third party; 2. Abnormal login frequency. It is understandable that the user frequently attempts to log in within a short period of time, or the login frequency suddenly increases. This may be an automated tool or a malicious user trying to crack the password; 3. Too many failed logins. It is understandable that multiple failed logins in a short period of time may indicate that someone is trying to guess the password; 4. Using malicious IP (Internet protocol) 1. Login using a different IP address (Internet Protocol Address). It is understandable that login attempts are made using an IP address identified as malicious, which may indicate a hacker attack or other security threat. 2. Login at abnormal login times. It is understandable that logins are made at unusual times (such as early morning), which may not be normal behavior for the account owner. 3. Account sharing behavior. It is understandable that the same account logs in from multiple different IP addresses in a short period of time, which may indicate that the account is shared or used by unauthorized users. 4. Abnormal access to sensitive data. It is understandable that there is a sudden increase in user access to sensitive data, which may be due to internal personnel abusing their authority or external attackers trying to obtain sensitive information.

[0085] Abnormalities in account binding include but are not limited to the following four types: 1. Incorrect account information, which means that there are errors in the bank card number, ID number, cardholder name, mobile phone number, expiration date and CVN information entered during the binding process; 2. Account status problem, which means that the bank card may be in a lost, frozen or cancelled state and the binding cannot be completed; 3. Off-site, batch or agency account opening, which means that the regulatory authorities particularly emphasize preventing abnormal behaviors such as off-site account opening, batch account opening, agency account opening, etc. through Internet channels; 4. False identity information, which means that personal information is untrue or incomplete.

[0086] Legal record verification refers to the review and verification of legal activities of individuals or organizations to test their legality and compliance. This may include supervision and inspection of administrative licenses and verification of other legal matters.

[0087] Compliance with abnormal transfer of funds includes but is not limited to the following five anomalies: 1. False transactions; 2. Abnormal account fund flow, which refers to abnormal fund flow in the customer's existing account, including payment of large integer amounts, transaction frequency and amount inconsistent with business activities, etc.; 3. Cross-border business compliance.

[0088] Real-name compliance means that in various financial business scenarios, users are required to provide real personal identity information and undergo verification when registering, using services or conducting transactions. Real-name compliance includes but is not limited to: mobile phone number real-name system.

[0089] High-risk areas refer to whether the account is opened in an area where abnormal transactions often occur.

[0090] The frequency of out-of-town usage includes but is not limited to: the frequency of using an account in a city other than the city where the bank is located, and the frequency of using an out-of-town mobile phone number to open an account.

[0091] Frequent address changes include, but are not limited to: frequent changes in the home address of the customer to whom the account belongs.

[0092] Exemplary methods for determining which indicator layer attributes are associated with target rule policy attributes include but are not limited to the following two methods. The first method is to pre-set the corresponding relationship between the rule policy attributes and the indicator layer attributes, so that the target indicator layer attribute corresponding to the target rule policy attribute can be found from the corresponding relationship; the second method is to manually determine the target indicator layer attribute corresponding to the target rule policy attribute.

[0093] The following example illustrates that, assuming that the target rule policy attribute is the mobile phone number unreal-name policy attribute, the target indicator layer attributes associated with the mobile phone number unreal-name policy attribute include but are not limited to: real-name compliance, binding account abnormalities; assuming that the target rule policy attribute is the mobile phone number out-of-region policy attribute, the target indicator layer attributes associated with the mobile phone number out-of-region policy attribute include but are not limited to: out-of-region usage frequency.

[0094] Step S403: Determine the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute.

[0095] Exemplarily, the first preset weights corresponding to the attributes of each indicator layer may be set manually, or obtained through AHP (Analytic Hierarchy Process) technology.

[0096] Exemplarily, the lowest policy risk level=the first preset weight of the target indicator layer attribute*the corresponding proportion of the target rule policy attribute*100.

[0097] Step S404: obtaining a plurality of candidate policy risk levels corresponding to the target rule policy attribute; the minimum value of the plurality of candidate policy risk levels is the minimum policy risk level.

[0098] Exemplarily, the multiple candidate policy risk levels corresponding to the target rule policy attributes can be artificially obtained by experts with rich experience. Exemplarily, assuming that the target rule policy attribute is the mobile phone number unreal-name policy and the minimum policy risk level is 36, then the multiple candidate policy risk levels corresponding to the target rule policy attributes can be: 36, 40, 50, 70.

[0099] Step S405: Acquire the policy risk level corresponding to the target rule policy attribute from the plurality of candidate policy risk levels.

[0100] Step S406: classify the target rule policy attributes and the policy risk levels corresponding to the target rule policy attributes into the risk attribute set; the risk attribute set is the input of the preset account opening risk prediction model.

[0101] An embodiment of the present application provides a method for obtaining the input parameters of an account opening risk prediction model. With the development of technology, the parameters that affect the risk of an account opening are constantly changing. If the input parameters of the account opening risk prediction model, that is, the number of parameters included in the risk attribute set, remain unchanged, the accuracy of the output result of the account opening risk prediction model will be reduced. Based on this, an embodiment of the present application obtains a target rule policy attribute that does not belong to the risk attribute set and has a proportion greater than or equal to a preset probability from the rule policy attributes corresponding to multiple customers included in a preset blacklist. Since the proportion of customers with the target rule policy attribute in the preset blacklist is relatively high, it means that the target rule policy attribute is a parameter that affects the account opening risk; however, it is necessary to further determine the degree to which the target rule policy attribute affects the account opening risk, that is, the policy risk level. The specific process includes: determining the target indicator layer attribute that has an association relationship with the target rule policy attribute from the preset multiple indicator layer attributes; determining the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute; obtaining multiple candidate policy risk levels corresponding to the target rule policy attribute; the minimum value of the multiple candidate policy risk levels is the minimum policy risk level; obtaining the policy risk level corresponding to the target rule policy attribute from the multiple candidate policy risk levels. The target rule policy attribute and the policy risk level corresponding to the target rule policy attribute are divided into the risk attribute set. In this way, the input parameters of the account opening risk prediction model are continuously enriched, thereby improving the accuracy of the output of the account opening risk prediction model.

[0102] It is understandable that there are multiple ways to implement step S405, and the embodiments of the present application provide but are not limited to the following two.

[0103] The first implementation of step S405 includes the following steps A11 to A14.

[0104] Step A11: for each of the candidate strategy risk levels, the account opening risk prediction model is trained using training samples to obtain the account opening risk prediction models corresponding to a plurality of the candidate strategy risk levels.

[0105] The training sample includes a risk attribute set corresponding to the training sample user, a target rule policy attribute corresponding to the training sample user, the candidate policy risk corresponding to the target rule policy attribute, and an account opening risk score of the training sample user.

[0106] Among them, the risk attribute set corresponding to the training sample users, the target rule policy attributes corresponding to the training sample users, and the candidate policy risks corresponding to the target rule policy attributes are taken as inputs, and the account opening risk scores of the training sample users are taken as training targets to train and obtain an account opening account risk prediction model.

[0107] It is understandable that the number of target rule policy attributes can be one or more; if the number of target rule policy attributes is multiple, it is necessary to train the account opening risk prediction model for each target rule policy attribute, that is, the training sample includes a risk attribute set and a target rule policy attribute with an undetermined policy risk level. After obtaining the policy risk level of the target rule policy attribute through step A14, the target rule policy attribute and the policy risk level of the target rule policy attribute are divided into a risk attribute set, and steps A11 to A14 are performed again. At this time, the training sample includes a risk attribute set and a target rule policy attribute with an undetermined policy risk level. Until the policy risk levels of the target rule policy attributes are all determined.

[0108] Exemplarily, the account opening risk prediction model may be a machine learning model or a large language model.

[0109] Exemplarily, it is assumed that the risk attribute set includes: mobile phone number unreal-name policy attribute, mobile phone number out-of-region policy attribute, and login at an uncommon login location; assuming that the sample user's mobile phone number is real-named and the mobile phone number has been used in a different region and the mobile banking has been logged in at an uncommon login location, then the risk attribute set corresponding to the sample user is: mobile phone number out-of-region policy attribute, and login at an uncommon login location; assuming that the sample user's mobile phone number is not real-named and the mobile phone number has been used in a different region and the mobile banking has been logged in at an uncommon login location, then the risk attribute set corresponding to the sample user is: mobile phone number unreal-name policy attribute, mobile phone number out-of-region policy attribute, and login at an uncommon login location.

[0110] Step A12: for each of the account opening risk prediction models corresponding to the risk level of the candidate strategies, the precision and recall of the account opening risk prediction model are obtained through test samples.

[0111] The test sample includes a risk attribute set corresponding to the test sample user, a target rule policy attribute corresponding to the test sample user, the candidate policy risk corresponding to the target rule policy attribute, and an account opening risk score of the test sample user.

[0112] Step A13: for each of the account opening risk prediction models corresponding to the candidate strategy risk level, determine the F1 score of the account opening risk prediction model based on the precision and recall of the account opening risk prediction model through a confusion matrix.

[0113] The confusion matrix is ​​a table commonly used in machine learning and statistical classification to show the performance of classification models. It intuitively shows the comparison between the model threshold results and the true categories, which is crucial for evaluating and optimizing classification models. The confusion matrix can be used to calculate the F1 score of the account opening risk prediction model corresponding to the candidate strategy risk level.

[0114] The confusion matrix is ​​described below, as shown in Table 1.

[0115] Table 1

[0116]

[0117] True Positive (TP): The true class refers to the true result of the test sample user is that there is an account opening risk (positive class), and the result identified by the account opening account risk prediction model is also that there is an account opening risk.

[0118] False Negative (FN): False negative class refers to the test sample user's true result is that there is account opening risk, but the account opening account risk prediction model identifies it as not having account opening risk.

[0119] False Positive (FP): False positive refers to the fact that the true result of the test sample user does not have account opening risk, but the account opening account risk prediction model identifies it as having account opening risk.

[0120] True Negative (TN): True negative class refers to the true result of the test sample user is that there is no account opening risk (negative class), and the account opening account risk prediction model identifies it as having no account opening risk.

[0121] Accuracy is the most commonly used classification performance indicator. It can be used to indicate the accuracy of the account opening risk prediction model, that is, the number of test sample users correctly identified by the account opening risk prediction model / the total number of test sample users. Generally speaking, the higher the accuracy of the account opening risk prediction model, the better the effect of the account opening risk prediction model.

[0122] The accuracy rate or precision is also called the precision rate, which indicates the proportion of test sample users that are truly positive among the test sample users identified as positive by the account opening risk prediction model. Generally speaking, the higher the precision rate, the better the effect of the account opening risk prediction model.

[0123] Recall, also known as the recall rate, refers to the ratio of the number of test sample users correctly identified as positive by the account opening risk prediction model to the total number of positive test sample users. Generally speaking, the higher the Recall, the more positive test sample users are correctly predicted by the account opening risk prediction model, and the better the effect of the account opening risk prediction model.

[0124] Step A14: Determine the candidate policy risk level corresponding to the highest F1 score as the policy risk level of the target rule policy attribute.

[0125] In order for those skilled in the art to better understand the embodiments of the present application, it is assumed that the target rule policy attribute is the mobile phone number unreal-name policy attribute, and the multiple candidate policy risk levels corresponding to the target rule policy attribute are 36, 40, 50, and 70. The applicant trains the account opening risk prediction model through a large number of training sample users, and tests the trained account opening risk prediction model through a large number of test sample users, thereby obtaining the F1 scores corresponding to the multiple candidate policy risk levels.

[0126] pass Figure 5 It can be seen that the distribution effect of the opening account risk prediction model under different risk levels of candidate strategies.

[0127] Figure 5 The horizontal axis is the account opening risk score output by the account opening risk prediction model, and the vertical axis is the F1 score. For example, if the account opening risk score is higher than the preset score, it is confirmed that there is an account opening risk, otherwise it is confirmed that there is no account opening risk.

[0128] For example, Figure 5 The F1 score on the ordinate corresponding to the horizontal coordinate 35+ is the sum of the F1 score corresponding to the account opening risk score of [35,40), the F1 score corresponding to the account opening risk score of [40,45), ..., and the F1 score corresponding to the account opening risk score of [100,+∞). The F1 score on the ordinate corresponding to the horizontal coordinate 100+ is the F1 score corresponding to the account opening risk score of [100,+∞).

[0129] pass Figure 5 The effect comparison table shown in Table 2 can be obtained.

[0130] Table 2

[0131]

[0132] In Table 2, the first number refers to the total number of test sample users used to test the account opening risk prediction model; the second number refers to the total number of test sample users with account opening risk output by the account opening risk prediction model; the fourth number refers to the number of test sample users with account opening risk output by the account opening risk prediction model that actually have account opening risk. The third number is the number of test sample users with actual account opening risk among the test sample users used to test the account opening risk prediction model.

[0133] It can be seen from Table 2 that when the account opening risk score is [100, +∞), the F1 score corresponding to the candidate strategy risk level 50 is the largest, so the strategy risk level of the mobile phone number unrealized name strategy attribute is determined to be 50. Compared with the mobile phone number unrealized name strategy attribute based on AHP hierarchical analysis, which is 36 points, the second number increases by 70 people, and the F1 score increases by 1.55%.

[0134] In order for those skilled in the art to better understand that adding the mobile phone number unreal-name policy attribute to the risk attribute set improves the accuracy of the account opening risk prediction model, an example is given below to illustrate.

[0135] like Figure 6 , which is a comparison chart of the recall rates of the account opening risk prediction models provided in the embodiments of the present application.

[0136] Figure 6 The horizontal axis refers to the account opening risk score output by the account opening risk prediction model, and the vertical axis refers to the recall rate. Figure 6 It can be seen that after adding the mobile phone number unreal-name policy attribute to the risk attribute set, the recall rate of the account opening risk prediction model is greatly improved.

[0137] It is understandable that there are multiple ways to implement step S401. The embodiment of the present application provides but is not limited to the following method, which includes but is not limited to the following steps B1 to B5 during implementation.

[0138] Step B1: Acquire candidate rule policy attributes that do not belong to the risk attribute set from the rule policy attributes corresponding to the plurality of customers included in the preset blacklist.

[0139] Step B2: for each of the candidate rule policy attributes, the customers in the preset blacklist having the same candidate rule policy attribute are divided into the same customer set.

[0140] Step B3: obtaining the proportion of the customer set corresponding to each of the plurality of candidate rule policy attributes, wherein the proportion of the customer set refers to the ratio of the number of customers included in the customer set to the number of customers in the preset blacklist.

[0141] Step B4: searching for the possibility levels corresponding to the proportions of the attributes of the plurality of candidate rule policies from the preset risk rule policies, wherein the risk rule policies include the corresponding relationship between the proportions and the possibility levels.

[0142] The following is an explanation of the risk rule strategy. For example, the risk rule strategy is determined using the Words to Estimate Probability (WEP). The Words to Estimate Probability uses probability words to determine the probability of a rule strategy attribute occurring, such as using extremely likely, highly likely, possible, unlikely, rare, etc. to describe it.

[0143] Exemplarily, different or the same scale subdivisions, ie, the frequency descriptions in Table 3, may be set for different rule policy attributes.

[0144] The following example illustrates the risk rule strategy. Assuming that the rule strategy attribute is the mobile phone number unreal-name policy attribute, the risk rule strategy is shown in Table 3.

[0145] Table 3

[0146]

[0147] Exemplarily, the frequency descriptions in Table 3 may be determined by experts with extensive experience.

[0148] Step B5: Acquire the target rule policy attribute whose possibility level is higher than or equal to the preset level from the plurality of candidate rule policy attributes.

[0149] Exemplarily, in combination with Table 3, the preset level is 3.

[0150] In an optional implementation, step S401 may not include step B4 and step B5.

[0151] It is understandable that there are multiple ways to implement step S402, and the embodiments of the present application provide but are not limited to the following two.

[0152] The first implementation of step S402 includes the following steps C11 to C12.

[0153] Step C11: Determine a target risk rule attribute associated with the target rule policy attribute from the preset risk rule attributes.

[0154] Step C12: Determine a target indicator layer attribute associated with the target rule strategy attribute from a plurality of indicator layer attributes included in the target risk rule attribute.

[0155] The risk rule attributes include one or more indicator layer attributes. An example is given below.

[0156] Exemplarily, risk rule attributes include, but are not limited to: legal compliance, behavioral risk, and regional risk; among them, legal compliance includes multiple indicator-level attributes, namely: legal record verification, compliance with abnormal fund transfers, and compliance with the real-name system; behavioral risk includes multiple indicator-level attributes, namely: abnormal transaction behavior, abnormal login behavior, and abnormal binding account; regional risk includes multiple indicator-level attributes, namely: high-risk areas, frequency of use in other places, and frequent changes of addresses.

[0157] Exemplarily, for each of the target indicator layer attributes, a product of a second preset weight corresponding to the target indicator layer attribute and a third preset weight corresponding to the target indicator layer attribute is determined as a first preset weight of the target indicator layer attribute.

[0158] The second implementation of step S402 includes: acquiring a target indicator layer attribute associated with the target rule policy attribute from a plurality of preset indicator layer attributes.

[0159] Exemplarily, the first preset weight of the target indicator layer attribute may be obtained based on an AHP algorithm.

[0160] In an optional implementation, if step S401 includes step B4 and step B5, and step S402 includes step C11 to step C12, then step S403 includes steps D11 to step D13.

[0161] Step D11: Obtain the candidate risk level of the target risk rule attribute, which is the sum of the products of the first preset weight corresponding to the target indicator layer attribute contained in the target risk rule attribute and the possibility level corresponding to the target rule strategy attribute.

[0162] Exemplarily, a method for obtaining the first preset weight of the target indicator layer attribute includes: for each target indicator layer attribute, determining the product of the second preset weight corresponding to the target indicator layer attribute and the third preset weight corresponding to the target risk rule attribute to which the target indicator layer attribute belongs as the first preset weight.

[0163] Step D12: If there are multiple target risk rule attributes associated with the target rule policy attribute, determine the sum of the candidate risk levels respectively corresponding to the multiple target risk rule attributes as the minimum policy risk level.

[0164] Step D13: If there is one target risk rule attribute associated with the target rule policy attribute, determine that the candidate risk level corresponding to the target risk rule attribute is the lowest policy risk level.

[0165] In order for those skilled in the art to better understand the following technical solution, namely, “step S401 includes step B4 and step B5, step S402 includes step C11 to step C12, and step S403 includes step D11 to step D13”, the above technical solution is described below by taking an example.

[0166] Assume that step S401 obtains two rule policy attributes, which are: the mobile phone number unrealized policy attribute and the mobile phone number out-of-region policy attribute; analyze the blacklisted customers who opened accounts in April 2024, and obtain that the mobile phone number unrealized policy attribute accounts for 33.7%, and the mobile phone number out-of-region policy attribute accounts for 26.6%. Combined with Table 3, the mobile phone number unrealized policy attribute accounts for 33.7%, which belongs to [25%-50%), so the possibility level corresponding to the mobile phone number unrealized policy attribute is 3; the mobile phone number out-of-region policy attribute accounts for 26.6%, which belongs to [25%-50%), so the possibility level corresponding to the mobile phone number unrealized policy attribute is 3. Assuming the preset level is 3, the mobile phone number unrealized policy attribute and the mobile phone number out-of-region policy attribute are both target rule policy attributes.

[0167] Assume that the preset multiple risk rule attributes are: legal compliance, behavioral risk and regional risk; the target risk rule attributes that are associated with the mobile phone number's unreal-name policy attribute include: legal compliance and behavioral risk; the target risk rule attributes that are associated with the mobile phone number's remote policy attribute include: regional risk. Among the multiple indicator layer attributes included in legal compliance, the target indicator layer attribute that is associated with the mobile phone number's unreal-name policy attribute is real-name compliance; among the multiple indicator layer attributes included in behavioral risk, the target indicator layer attribute that is associated with the mobile phone number's unreal-name policy attribute is binding account anomaly; among the multiple indicator layer attributes included in regional risk, the target indicator layer attribute that is associated with the mobile phone number's remote policy attribute is remote usage frequency. In order for those skilled in the art to be more clear about the implementation process of steps D11 to D13, the following is an explanation in conjunction with Table 4.

[0168] In Table 4, the places without correlation are indicated by “None”.

[0169] It can be seen from Table 4 that the first preset weight of the target indicator layer attribute C=M×S; the candidate risk level of the target risk rule attribute R=L×C. The mobile phone number unreal-name policy attribute is associated with legal compliance and behavioral risk, so the minimum policy risk level of the mobile phone number unreal-name policy attribute = 0.21×100+0.15×100=36. The mobile phone number out-of-region policy attribute is associated with regional risk, so the minimum policy risk level of the mobile phone number out-of-region policy attribute = 0.09×100=9.

[0170] Table 4

[0171]

[0172] In an optional implementation, the third preset weight of the risk rule attribute and the second preset weight of the indicator layer attribute are manually obtained by an expert with rich experience.

[0173] In an optional implementation, the third preset weight of the risk rule attribute and the second preset weight of the indicator layer attribute are obtained by an AHP algorithm. The process includes but is not limited to the following steps E1 to E5.

[0174] Step E1: Acquire multiple risk rule attributes and indicator layer attributes respectively included in the multiple risk rule attributes.

[0175] The following table provides examples of risk rule attributes and indicator level attributes, as shown in Table 5.

[0176] Table 5

[0177]

[0178] AHP is used to construct a hierarchical structure of multiple risk rule attributes, such as Figure 7 , which is a schematic diagram of a hierarchical structure of multiple risk rule attributes constructed using AHP provided in an embodiment of the present application.

[0179] like Figure 7 As shown, in the embodiment of the present application, multiple risk rule attributes constitute the criterion layer; multiple indicator layer attributes constitute the indicator layer. In the embodiment of the present application, multiple risk rule attributes and multiple indicator layer attributes respectively included in multiple risk rule attributes all belong to account opening risk rules.

[0180] AHP is used to calculate the third preset weight of risk rule attributes. The key step is to build a judgment matrix and perform consistency test. When making pairwise comparisons, relative scales are used to describe the importance of indicators, so as to minimize the difficulty of comparing risk rule attributes of different natures and improve the accuracy of judgment. The scales used are shown in Table 6.

[0181] Table 6 Scale scale of judgment matrix

[0182]

[0183] Step E2: Obtain the scale importance scores of the business teachers and experts on the multiple risk rule attributes, and construct a first judgment matrix including the multiple risk rule attributes.

[0184] If the risk rule attributes are taken as Table 5, a first judgment matrix including multiple risk rule attributes can be obtained. Figure 8 A schematic diagram of a first judgment matrix including multiple risk rule attributes provided in an embodiment of the present application.

[0185] Combining Table 6 and Figure 8 , it can be seen that the scale importance scores of business teachers and experts for multiple risk rule attributes are as follows:

[0186] The scale importance of legal compliance relative to legal compliance is 1, the scale importance of legal compliance relative to behavioral risk is 3, and the scale importance of legal compliance relative to regional risk is 5; the scale importance of behavioral risk relative to legal compliance is 1 / 3, the scale importance of behavioral risk relative to behavioral risk is 1, and the scale importance of behavioral risk relative to regional risk is 4; the scale importance of regional risk relative to legal compliance is 1 / 5, the scale importance of regional risk relative to behavioral risk is 1 / 4, and the scale importance of regional risk relative to regional risk is 1. Thus, the first judgment matrix can be obtained.

[0187] Step E3: normalizing the first judgment matrix to obtain a second judgment matrix.

[0188] Fig. 9 For the general Figure 8 The diagram shows a second judgment matrix obtained after normalizing the first judgment matrix and a schematic diagram of a consistency test result.

[0189] Here is an example of Fig. 9 The normalization process is explained in the following example. For example, the importance of the scale of legal compliance is 1, 1 / 3, and 1 / 5, respectively. The normalized value of the importance of legal compliance relative to legal compliance is 1 / (1+1 / 3+1 / 5)=0.65; the normalized value of the importance of legal compliance relative to behavioral risk is (1 / 3) / (1+1 / 3+1 / 5)=0.22. The subsequent steps are similar and no further examples are given.

[0190] Step E4: Determine the third preset weight of each risk rule attribute in the second judgment matrix as the average value of the row where the risk rule attribute is located.

[0191] Still Fig. 9For example, the third preset weight of legal compliance is 0.62=(0.65+0.70+0.50) / 3; the third preset weight of behavioral risk is 0.28=(0.22+0.24+0.40) / 3; and the third preset weight of regional risk is 0.10=(0.13+0.06+0.10) / 3.

[0192] Step E5: Perform a consistency check on the second judgment matrix.

[0193] like Fig. 9 Shown is the result of the consistency test on the second judgment matrix.

[0194] Generally, if the CR value is less than 0.1, the second judgment matrix satisfies the consistency test; if the CR value is greater than 0.1, it indicates that there is no consistency. Fig. 9 The CR value is less than 0.1, so it passes the consistency test.

[0195] In an optional implementation, the second preset weights respectively corresponding to the multiple indicator layer attributes may also be obtained through AHP, which is described below. The process of obtaining the second preset weights respectively corresponding to the multiple indicator layer attributes includes steps F1 to F4.

[0196] Step F1: Obtain the scale importance scores of the business teachers and experts on the multiple indicator layer attributes belonging to the same risk rule attribute, and construct a third judgment matrix including the multiple indicator layer attributes belonging to the same risk rule attribute.

[0197] The following is an example.

[0198] like Fig.10 As shown, it is a schematic diagram of three third judgment matrices provided in an embodiment of the present application.

[0199] Fig.10 The three third judgment matrices are the third judgment matrix corresponding to behavioral risk, the third judgment matrix corresponding to legal compliance, and the third judgment matrix corresponding to regional risk.

[0200] Step F2: normalize the third judgment matrix to obtain a fourth judgment matrix.

[0201] like Fig.11 As shown in the embodiment of the present application, Fig.10 Schematic diagram of the normalized fourth judgment matrix of the three third judgment matrices.

[0202] The process of normalizing the third judgment matrix is ​​the same as the process of normalizing the first judgment matrix, which will not be described in detail here.

[0203] Step F3: Determine the second preset weight of each indicator layer attribute in the fourth judgment matrix as the average value of the row where the indicator layer attribute is located.

[0204] See also Fig.11 The second preset weights of the attributes of each indicator layer are not described here.

[0205] Step F4: Perform consistency check on the fourth judgment matrix.

[0206] Generally, if the CR value is less than 0.1, the fourth judgment matrix satisfies the consistency test; if the CR value is greater than 0.1, it indicates that there is no consistency. Fig.11 The CR value is less than 0.1, so it passes the consistency test.

[0207] The above introduces a method for obtaining input parameters of an account opening risk prediction model provided in an embodiment of the present application. The following will introduce a device for executing the above method for obtaining input parameters of the account opening risk prediction model.

[0208] See also Fig.12 , Fig.12 A schematic diagram of a device for obtaining input parameters of an account risk prediction model provided in an embodiment of the present application. Fig.12 As shown, including:

[0209] The first acquisition module 1201 is used to obtain a target rule policy attribute that does not belong to the risk attribute set and has a proportion greater than or equal to a preset probability from the rule policy attributes corresponding to multiple customers included in the preset blacklist; the proportion of the rule policy attribute refers to the ratio of the number of customers with the rule policy attribute in the preset blacklist to the number of all customers included in the preset blacklist; the risk attribute set includes the rule policy attribute and the policy risk level corresponding to the rule policy attribute;

[0210] The first determination module 1202 is used to determine a target indicator layer attribute associated with the target rule policy attribute from a plurality of preset indicator layer attributes;

[0211] The second determination module 1203 is used to determine the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute;

[0212] The second acquisition module 1204 is used to acquire multiple candidate policy risk levels corresponding to the target rule policy attribute; the minimum value of the multiple candidate policy risk levels is the minimum policy risk level;

[0213] The third acquisition module 1205 is used to acquire the policy risk level corresponding to the target rule policy attribute from the plurality of candidate policy risk levels;

[0214] The classification module 1206 is used to classify the target rule policy attribute and the policy risk level corresponding to the target rule policy attribute into the risk attribute set; the risk attribute set is the input of the preset account opening risk prediction model.

[0215] In an optional implementation, the third acquisition module includes:

[0216] A training unit, configured to train an account opening risk prediction model for each candidate policy risk level through the risk attribute set, the target rule policy attribute, and the candidate policy risk level corresponding to the target rule policy attribute;

[0217] A testing unit, configured to obtain, for each of the candidate strategy risk levels, an F1 score of the account opening risk prediction model corresponding to the candidate strategy risk level through a confusion matrix;

[0218] The first determining unit is configured to determine the candidate policy risk level corresponding to the highest F1 score as the policy risk level of the target rule policy attribute.

[0219] In an optional implementation, the first acquisition module includes:

[0220] A first acquisition unit, configured to acquire a candidate rule policy attribute that does not belong to the risk attribute set from the rule policy attributes respectively corresponding to the plurality of customers included in the preset blacklist;

[0221] A division unit, for dividing the customers in the preset blacklist having the same candidate rule policy attribute into the same customer set for each candidate rule policy attribute;

[0222] A second acquisition unit is used to acquire the proportion of the customer set corresponding to each of the plurality of candidate rule policy attributes, wherein the proportion of the customer set refers to the ratio of the number of customers included in the customer set to the number of customers in the preset blacklist;

[0223] A second determining unit is used to find the possibility level corresponding to the proportion of the attributes of the plurality of candidate rule strategies from a preset risk rule strategy, wherein the risk rule strategy includes a correspondence between the proportion and the possibility level;

[0224] The third acquisition unit is used to acquire the target rule policy attribute whose possibility level is higher than or equal to the preset level from the multiple candidate rule policy attributes.

[0225] In an optional implementation, the first determining module includes:

[0226] A third determining unit is used to determine a target risk rule attribute having an associated relationship with the target rule policy attribute from preset risk rule attributes;

[0227] The fourth determining unit is configured to determine a target indicator layer attribute associated with the target rule strategy attribute from a plurality of indicator layer attributes included in the target risk rule attribute.

[0228] In an optional implementation, the method further includes:

[0229] The third determination module is used to determine, for each target indicator layer attribute, the product of the second preset weight corresponding to the target indicator layer attribute and the third preset weight corresponding to the target risk rule attribute to which the target indicator layer attribute belongs as the first preset weight.

[0230] In an optional implementation, the second determining module includes:

[0231] a fourth acquisition unit, configured to acquire a candidate risk level of the target risk rule attribute, wherein the candidate risk level of the target risk rule attribute is the sum of the products of the first preset weight corresponding to the target indicator layer attribute contained in the target risk rule attribute and the possibility level corresponding to the target rule strategy attribute;

[0232] a fifth determining unit, configured to determine, if there are multiple target risk rule attributes associated with the target rule policy attribute, that the sum of the candidate risk levels respectively corresponding to the multiple target risk rule attributes is the minimum policy risk level;

[0233] The sixth determining unit is configured to determine that the candidate risk level corresponding to the target risk rule attribute is the lowest policy risk level if there is one target risk rule attribute associated with the target rule policy attribute.

[0234] The present application also provides an electronic device in an embodiment. Fig.13 As shown, it shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present application. The electronic device in the embodiment of the present application may include but is not limited to fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Fig.13 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0235] like Fig.13As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 1301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1302 or a program loaded from a storage device 1308 to a random access memory (RAM) 1303. When the electronic device is powered on, various programs and data required for the operation of the electronic device are also stored in the RAM 1303. The processing device 1301, the ROM 1302, and the RAM 1303 are connected to each other via a bus 1304. An input / output (I / O) interface 1305 is also connected to the bus 1304.

[0236] Typically, the following devices may be connected to the I / O interface 1305: an input device 1306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 1307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1308 including, for example, a memory card, a hard disk, etc.; and a communication device 1309. The communication device 1309 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 6 An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0237] Also provided in an embodiment of the present application is a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements a method for obtaining input parameters of any account opening risk prediction model provided in an embodiment of the present application.

[0238] A computer-readable storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When one or more computer programs are executed by an electronic device, the electronic device can implement a method for obtaining input parameters of any account opening risk prediction model provided in an embodiment of the present application.

[0239] It should also be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0240] Through the description of the above implementation mode, the technicians in the field can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present application, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0241] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0242] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, a computer, a training device, or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, training device, or data center. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

Claims

1. A method for obtaining input parameters of an account opening risk prediction model, It is characterized in that include: From the rule policy attributes respectively corresponding to the multiple customers included in the preset blacklist, obtain the target rule policy attribute that does not belong to the risk attribute set and whose proportion is greater than or equal to the preset probability; the proportion of the rule policy attribute refers to the ratio of the number of customers with the rule policy attribute in the preset blacklist to the number of all customers included in the preset blacklist; the risk attribute set includes the rule policy attribute and the policy risk level corresponding to the rule policy attribute; Determine, from a plurality of preset indicator layer attributes, a target indicator layer attribute associated with the target rule strategy attribute; Determine the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute; Obtaining multiple candidate policy risk levels corresponding to the target rule policy attribute; The minimum value of the plurality of candidate policy risk levels is the minimum policy risk level; Acquire the policy risk level corresponding to the target rule policy attribute from the plurality of candidate policy risk levels; The target rule policy attributes and the policy risk levels corresponding to the target rule policy attributes are divided into the risk attribute set; the risk attribute set is the input of a preset account opening risk prediction model.

2. The method for obtaining input parameters of the account opening risk prediction model according to claim 1, It is characterized in that The step of acquiring the policy risk level corresponding to the target rule policy attribute from the plurality of candidate policy risk levels comprises: For each of the candidate strategy risk levels, the account opening risk prediction model is trained through training samples to obtain the account opening risk prediction models corresponding to a plurality of the candidate strategy risk levels, wherein the training samples include a risk attribute set corresponding to a training sample user, a target rule strategy attribute corresponding to the training sample user, the candidate strategy risk corresponding to the target rule strategy attribute, and an account opening risk score of the training sample user; For the account opening risk prediction model corresponding to each candidate strategy risk level, the precision and recall of the account opening risk prediction model are obtained through a test sample; the test sample includes a risk attribute set corresponding to the test sample user, a target rule strategy attribute corresponding to the test sample user, the candidate strategy risk corresponding to the target rule strategy attribute, and an account opening risk score of the test sample user; For each of the open account risk prediction models corresponding to the candidate strategy risk level, determine the F1 score of the open account risk prediction model based on the precision and recall of the open account risk prediction model through a confusion matrix; The candidate policy risk level corresponding to the highest F1 score is determined as the policy risk level of the target rule policy attribute.

3. A method for obtaining input parameters of an account opening risk prediction model according to any one of claims 1 or 2, It is characterized in that The step of obtaining a target rule strategy attribute that does not belong to the risk attribute set and has a proportion greater than or equal to a preset probability from the rule strategy attributes corresponding to the multiple customers included in the preset blacklist includes: Acquire a candidate rule policy attribute that does not belong to the risk attribute set from the rule policy attributes corresponding to the plurality of customers included in the preset blacklist; For each of the candidate rule policy attributes, the customers in the preset blacklist having the same candidate rule policy attribute are divided into the same customer set; Obtaining the proportion of the customer set corresponding to each of the plurality of candidate rule policy attributes, wherein the proportion of the customer set refers to the ratio of the number of customers included in the customer set to the number of customers in the preset blacklist; From the preset risk rule strategy, searching for the possibility level corresponding to the proportion of the attributes of the plurality of candidate rule strategies, wherein the risk rule strategy includes the corresponding relationship between the proportion and the possibility level; The target rule policy attribute whose possibility level is higher than or equal to the preset level is obtained from the plurality of candidate rule policy attributes.

4. The method for obtaining input parameters of the account opening risk prediction model according to claim 3, It is characterized in that The step of determining, from among a plurality of preset indicator layer attributes, a target indicator layer attribute associated with the target rule policy attribute comprises: Determine, from the preset risk rule attributes, a target risk rule attribute associated with the target rule policy attribute; From the plurality of indicator-layer attributes included in the target risk rule attribute, a target indicator-layer attribute associated with the target rule policy attribute is determined.

5. The method for obtaining input parameters of the account opening risk prediction model according to claim 4, It is characterized in that The first preset weight method for obtaining the target indicator layer attribute includes: For each of the target indicator layer attributes, the product of the second preset weight corresponding to the target indicator layer attribute and the third preset weight corresponding to the target risk rule attribute to which the target indicator layer attribute belongs is determined as the first preset weight.

6. The method for obtaining input parameters of the account opening risk prediction model according to claim 4, It is characterized in that The step of determining the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute includes: Acquire a candidate risk level of the target risk rule attribute, where the candidate risk level of the target risk rule attribute is the sum of the products of the first preset weight corresponding to the target indicator layer attribute contained in the target risk rule attribute and the possibility level corresponding to the target rule strategy attribute; If there are multiple target risk rule attributes associated with the target rule policy attribute, determine the sum of the candidate risk levels respectively corresponding to the multiple target risk rule attributes as the minimum policy risk level; If there is one target risk rule attribute associated with the target rule policy attribute, the candidate risk level corresponding to the target risk rule attribute is determined to be the lowest policy risk level.

7. A device for obtaining input parameters of an account opening risk prediction model, It is characterized in that include: The first acquisition module is used to obtain a target rule policy attribute that does not belong to the risk attribute set and has a proportion greater than or equal to a preset probability from the rule policy attributes corresponding to multiple customers included in the preset blacklist; the proportion of the rule policy attribute refers to the ratio of the number of customers with the rule policy attribute in the preset blacklist to the number of all customers included in the preset blacklist; the risk attribute set includes the rule policy attribute and the policy risk level corresponding to the rule policy attribute; A first determination module is used to determine a target indicator layer attribute having an associated relationship with the target rule strategy attribute from a plurality of preset indicator layer attributes; A second determination module is used to determine the minimum policy risk level of the target rule policy attribute based on the first preset weight of the target indicator layer attribute and the proportion corresponding to the target rule policy attribute; A second acquisition module is used to acquire a plurality of candidate policy risk levels corresponding to the target rule policy attribute; The minimum value of the plurality of candidate policy risk levels is the minimum policy risk level; A third acquisition module, configured to acquire the policy risk level corresponding to the target rule policy attribute from the plurality of candidate policy risk levels; A division module is used to divide the target rule policy attribute and the policy risk level corresponding to the target rule policy attribute into the risk attribute set; the risk attribute set is the input of the preset account opening risk prediction model.

8. A computer program product, It is characterized in that It includes computer-readable instructions, and when the computer-readable instructions are executed on an electronic device, the electronic device implements the method for obtaining input parameters of the account opening risk prediction model as described in any one of claims 1 to 6.

9. An electronic device, It is characterized in that The method comprises at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the electronic device can implement the method for obtaining input parameters of the account opening risk prediction model as described in any one of claims 1 to 6.

10. A computer storage medium, It is characterized in that The storage medium carries one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the method for obtaining input parameters of the account opening risk prediction model as described in any one of claims 1 to 6.

Images