Customer purchase power prediction method and related device

Abstract

The invention discloses a customer buying power prediction method and a related device, and relates to the field of risk control, and the method comprises the steps: obtaining first buying power data representing the risk level of a customer during product buying in a statistical time period, and second buying power data obtained based on a scoring questionnaire and a final transaction operation; splitting processing of a client side and a product side is carried out on the first purchasing power data, and a client side purchasing power matrix and a product side purchasing power matrix are obtained; and preprocessing the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix. And training a purchasing power prediction model based on the obtained data, and performing risk assessment on the purchasing power of the target customer for purchasing the financial product by using the trained purchasing power prediction model. According to the invention, the purchasing power prediction model can be converged more efficiently, the deviation of customer purchasing power prediction is reduced, and the prediction accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of risk control technology, and in particular to a method and related device for predicting customer purchasing power. Background Art

[0002] With the continuous increase in the number of financial products, the types of risks in financial products are also showing a trend of diversification. In order to prevent ambiguity in the assessment of customer risk tolerance and to more accurately match target customers and launch more accurate marketing services, each financial institution will establish its own risk tolerance assessment indicator system to predict relevant risks. At present, financial institutions generally use reputation questionnaire models to identify and judge customers' risk tolerance levels. This method relies on active questionnaires and a single risk type, resulting in the relevant sales services and marketing methods of bank financial products, which easily lead customers to purchase financial products that do not meet their expectations, resulting in a lose-lose situation. Summary of the invention

[0003] In view of the above problems, this application provides a customer purchasing power prediction method and related devices to achieve the purpose of reducing the deviation of customer purchasing power prediction and improving the accuracy of prediction. The specific scheme is as follows:

[0004] The first aspect of the present application provides a method for predicting customer purchasing power, comprising:

[0005] Obtaining first purchasing power data representing the risk level of customers when purchasing products during the statistical period, and second purchasing power data obtained based on the scoring questionnaire and the final transaction operation;

[0006] Splitting the first purchasing power data into a customer side and a product side to obtain a customer side purchasing power matrix and a product side purchasing power matrix;

[0007] Preprocessing the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix;

[0008] Training a purchasing power prediction model based on customer-side purchasing power prediction data and product-side purchasing power prediction data, wherein the customer-side purchasing power prediction data includes: the customer-side purchasing power matrix and the customer implicit factor query vector, and the product-side purchasing power prediction data includes: the product-side purchasing power matrix and the product implicit factor query vector;

[0009] Based on the trained purchasing power prediction model, the purchasing power of target customers for financial products is assessed for risk.

[0010] In a possible implementation, the purchasing power prediction model includes: a convolutional attention encoder, an implicit factor extractor, and a convolutional attention decoder. The purchasing power prediction model is trained based on the customer-side purchasing power prediction data and the product-side purchasing power prediction data, including:

[0011] Based on the convolutional attention encoder, feature extraction is performed on the customer-side purchasing power prediction data and the product-side purchasing power prediction data to obtain customer interest feature data and product attraction feature data;

[0012] Based on the implicit factor extractor, extracting customer implicit factors and product implicit factors from the customer interest feature data and the product attraction feature data respectively;

[0013] Based on the convolutional attention decoder, the customer interest feature data and the product attraction feature data are restored respectively.

[0014] In a possible implementation, the training of the purchasing power prediction model based on the customer-side purchasing power prediction data and the product-side purchasing power prediction data further includes:

[0015] Based on the optimizer aiming to minimize the value of the target loss function, the model parameters of the convolutional attention encoder, the implicit factor extractor and the convolutional attention decoder are updated.

[0016] In a possible implementation, the convolutional attention encoder includes: a convolutional encoder and a first attention encoding layer, the convolutional encoder includes: at least one convolutional encoding unit, the convolutional encoding unit includes: a convolutional layer and a pooling layer; based on the convolutional attention encoder, the customer-side purchasing power prediction data and the product-side purchasing power prediction data are respectively extracted to obtain customer interest feature data and product attraction feature data, including:

[0017] Based on the convolution encoder, feature extraction is performed on the customer-side purchasing power matrix and the product-side purchasing power matrix respectively to obtain customer-side purchasing power feature data and product-side purchasing power feature data;

[0018] Based on the first attention encoding layer, the first customer-side feature data and the first product-side feature data are processed respectively to obtain the customer interest feature data and the product attraction feature data, the first customer-side feature data including: the customer-side purchasing power feature data and the customer implicit factor query vector, the first product-side feature data including: the product-side purchasing power feature data and the product implicit factor query vector.

[0019] In a possible implementation, the implicit factor extractor includes: two convolutional layers and a fully connected layer, and the extraction of customer implicit factors and product implicit factors based on the customer interest feature data and the product attraction feature data by the implicit factor extractor includes:

[0020] Mapping the customer interest feature data to the customer implicit factor space based on the first convolutional layer, and mapping the product attraction feature data to the product implicit factor space;

[0021] Extracting the customer implicit factor from the customer implicit factor space based on the fully connected layer, and extracting the product implicit factor from the product implicit factor space;

[0022] Based on the second convolutional layer, the customer implicit factor space and the product implicit factor space are restored respectively to obtain the customer interest feature data and the product attraction feature data.

[0023] In a possible implementation, the convolutional attention decoder includes: a convolutional decoder and a second attention encoding layer, the convolutional decoder includes: convolutional decoding units having the same number as the convolutional encoding units, each of the convolutional encoding units includes: a deconvolution layer and an upsampling layer; based on the convolutional attention decoder, the customer interest feature data and the product attraction feature data are restored respectively, including:

[0024] Based on the second attention coding layer, the second customer-side feature data and the second product-side feature data are processed respectively to obtain the customer-side purchasing power feature data and the product-side purchasing power feature data, wherein the second customer-side feature data includes: the customer implicit factor space data output by the first convolution layer and the customer interest feature data, and the second product-side feature data includes: the product implicit factor space data output by the first convolution layer and the product attraction feature data;

[0025] The customer-side purchasing power feature data and the product-side purchasing power feature data are processed respectively based on the convolution decoder to obtain the customer-side purchasing power matrix and the product-side purchasing power matrix.

[0026] In a possible implementation, preprocessing the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix includes:

[0027] Based on a two-layer fully connected neural network with an activation function and a normalization function, the second purchasing power data is processed to obtain the customer implicit factor query vector and the product implicit factor query vector.

[0028] A second aspect of the present application provides a customer purchasing power prediction device, comprising:

[0029] A purchasing power data acquisition module is used to acquire first purchasing power data representing the risk level of customers when purchasing products during a statistical period, and second purchasing power data obtained based on a scoring questionnaire and a final transaction operation;

[0030] A first data preprocessing module, used for splitting the first purchasing power data into a customer side and a product side to obtain a customer side purchasing power matrix and a product side purchasing power matrix;

[0031] A second data preprocessing module, used for preprocessing the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix;

[0032] A prediction model updating module, used for training a purchasing power prediction model based on customer-side purchasing power prediction data and product-side purchasing power prediction data, wherein the customer-side purchasing power prediction data includes: the customer-side purchasing power matrix and the customer implicit factor query vector, and the product-side purchasing power prediction data includes: the product-side purchasing power matrix and the product implicit factor query vector; and

[0033] The purchasing power prediction module is used to conduct risk assessment on the purchasing power of target customers for purchasing financial products based on the trained purchasing power prediction model.

[0034] A 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 customer purchasing power prediction method of the first aspect or any implementation of the first aspect.

[0035] A fourth aspect of the present application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0036] The memory is used to store computer programs;

[0037] The processor is used to execute the computer program so that the electronic device can implement the customer purchasing power prediction method of the first aspect or any implementation of the first aspect.

[0038] A 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 implement the customer purchasing power prediction method of the first aspect or any implementation of the first aspect.

[0039] By means of the above technical solution, the customer purchasing power prediction method provided by the present application obtains the first purchasing power data that characterizes the risk level of the customer when purchasing the product during the statistical period, and the second purchasing power data obtained based on the scoring questionnaire and the final transaction operation. After that, the first purchasing power data is split into the customer side and the product side to obtain the customer side purchasing power matrix and the product side purchasing power matrix. The second purchasing power data is preprocessed into the customer implicit factor query vector corresponding to the customer side purchasing power matrix, and the product implicit factor query vector corresponding to the product side purchasing power matrix. Based on the above-obtained data, the purchasing power prediction model is trained, and the purchasing power of the target customer to purchase financial products is risk assessed using the trained purchasing power prediction model. By utilizing the secondary operation data of the customer risk assessment questionnaire and the customer transaction behavior, the products that the customer cares about are combined from the customer side and the product side purchasing power data, and the customer and product interaction data are dynamically learned, so that the purchasing power prediction model can converge more efficiently, reduce the deviation of the customer purchasing power prediction, and improve the accuracy of the customer purchasing power prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] 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.

[0041] Figure 1 A structural diagram of a customer purchasing power prediction system provided for this application;

[0042] Figure 2 A structural diagram of a terminal provided for this application;

[0043] Figure 3 A structural diagram of a server provided for this application;

[0044] Figure 4 A flowchart of a method for predicting customer purchasing power provided for this application;

[0045] Figure 5 A diagram of the purchasing power prediction model provided for this application;

[0046] Figure 6 A structural diagram of a customer purchasing power prediction device provided in this application;

[0047] Figure 7 A structural diagram of an electronic device provided in this application. DETAILED DESCRIPTION

[0048] 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.

[0049] 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.

[0050] 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.

[0051] See also Figure 1 , Figure 1 FIG. 1 shows a schematic diagram of the architecture of a customer purchasing power prediction system. 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.

[0052] Among them, an application for purchasing products can be installed on the terminal 100. The above application and web page can provide an interface. The terminal 100 can receive relevant parameters entered by the user on the product purchase interface and send the above parameters to the server 200. The server 200 can obtain processing results based on the received parameters and return the processing results to the terminal 100.

[0053] 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.

[0054] Next describe Figure 1 The product form of the mid-terminal 100;

[0055] 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.

[0056] Figure 2 An optional hardware structure diagram of the terminal 100 is shown.

[0057] 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 These are merely examples of terminals or multi-function devices and do not constitute limitations on the terminals or multi-function devices, which may include more or fewer components than those shown in the figures, or combinations of certain components, or different components.

[0058] 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 portable multifunctional device. 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.

[0059] Among them, the input device 132 can receive input data and the like.

[0060] 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 displays, and / or playback of any multimedia file. In an embodiment of the present application, the display unit 140 may be used to display an interface for product purchase, processing results, and the like.

[0061] 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.

[0062] 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.

[0063] 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.

[0064] In this embodiment of the present application, the radio frequency unit 110 can send data such as purchased products and user information to the server 200, and receive the prediction result of the purchase risk sent by the server 200.

[0065] 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.

[0066] 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.

[0067] 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 .

[0068] 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.

[0069] Next describe Figure 1 The product form of the server 200;

[0070] Figure 3 A schematic diagram of the structure 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.

[0071] 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.

[0072] 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).

[0073] 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).

[0074] The memory 204 may be used to store software codes related to the customer purchasing power prediction method, and the processor 202 may execute the steps of the customer purchasing power prediction method of the chip, and may also schedule other units to implement corresponding functions.

[0075] 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.

[0076] The collaborative filtering algorithm is considered to be one of the simple, effective and most commonly used rating prediction methods, and is widely used in rating prediction scenarios. However, when traditional collaborative filtering methods are faced with a large number of customer product purchase interaction matrices for matrix filling, they often cause a large number of null values ​​or extremely small data values ​​in the predicted data, which is called the data sparsity problem. Limited by the data sparsity problem, the algorithm cannot fully mine the limited interaction information of customer product matrices with different sparsities, the algorithm has low applicability to different data sets, and the stability of the algorithm prediction accuracy is poor; in addition, the traditional collaborative filtering algorithm cannot obtain the nonlinear deep representation of customers and products, resulting in low algorithm accuracy. Therefore, in view of the current needs of consumers and service providers in the industry for predicting customer purchasing power, there is an urgent need for a prediction method that can deeply mine the features of various sparse data of customer products, and then accurately predict customer purchasing power with high algorithm applicability and stability.

[0077] In order to solve the above problems, the embodiment of the present application provides a method for predicting customer purchasing power. The method for predicting customer purchasing power of the embodiment of the present application is described in detail below with reference to the accompanying drawings.

[0078] Reference Figure 4 , Figure 4 A schematic diagram of a method for predicting customer purchasing power provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, a customer purchasing power prediction method provided in an embodiment of the present application may include steps 401 to 405, and these steps are described in detail below.

[0079] 401. Obtain first purchasing power data representing the risk level of the customer when purchasing a product during a statistical period, and second purchasing power data obtained based on a scoring questionnaire and a final transaction operation.

[0080] Specifically, the first purchasing power data may include: customer purchase columns, product history data, purchase time, customer portraits and other related information for each customer within a statistical period of one day, one month, etc. Then, the interactive data of these purchasing powers are normalized into a level data representation, for example, it can be normalized into a 0-5 level data representation.

[0081] The second purchasing power data obtained based on the scoring questionnaire and the final transaction operation may include: customer financial status assessment data, investment experience assessment data, investment purpose assessment data, liquidity requirement assessment data, return expectation assessment data, risk preference assessment data, risk loss tolerance assessment data, and customer secondary operations (buy, sell, etc.) of corresponding products after holding positions. These data can be normalized into 0-1 level data representation according to different dimensions.

[0082] It is understandable that those skilled in the art may select and adjust the content of the above purchasing power data as needed, and no limitation is imposed here.

[0083] 402. Split the first purchasing power data into a customer side and a product side to obtain a customer side purchasing power matrix and a product side purchasing power matrix.

[0084] Specifically, the first purchasing power data may be a customer-product interaction matrix consisting of n customers and m products. Figure 5 Shown on the left:

[0085] U={u 1 ,u 2 , ..., u n} means there are n customers, that is, n rows of vectors, product V = {v 1 , v 2 , ..., v m} indicates that there are m products, i.e. m column vectors, and the value is the normalized risk rating r of the customer purchasing the product in the database. The position without data is filled with 0. Considering the need to extract local features based on convolution in the future and facilitate the subsequent perception interaction calculation of the double convolution autoencoder of the customer product, the customer-product interaction matrix is ​​processed into a square matrix. This interaction matrix is ​​decomposed into a purchasing power matrix based on the customer side and the product side:

[0086]

[0087] Among them, X i / Y j(i=1, 2, ..., n / j=1, 2, ..., m) represent the customer side purchasing power matrix and the product side purchasing power matrix respectively, ceil(·) represents returning the smallest integer greater than or equal to the specified expression, represents the side length of the square product-side purchasing power matrix output in the preprocessing stage, It represents the side length of the square customer-side purchasing power matrix output in the preprocessing stage, and i / j represents the batch numbers corresponding to the product-side purchasing power matrix / customer-side purchasing power matrix respectively.

[0088] 403. Preprocess the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix.

[0089] Specifically, the query input vectors can be customer-based and product-based A two-layer fully connected neural network with an activation function and a softmax normalization function is used to extract customer implicit factor query vectors and product implicit factor query vectors.

[0090]

[0091] in, represents the query input vector based on customer i and product j, softmax(·) is the column-normalized function, ζ(·) represents the activation function, and finally obtains the D×n / m-dimensional customer implicit factor query vector and product implicit factor query vector, respectively. D is the target dimension encoded by the double convolutional autoencoder. represents a solid matrix with d rows and i / j columns, Represent the model parameters of the two fully connected layers respectively.

[0092] 404. Based on the customer-side purchasing power prediction data and the product-side purchasing power prediction data, the purchasing power prediction model is trained. The customer-side purchasing power prediction data includes: a customer-side purchasing power matrix and a customer implicit factor query vector. The product-side purchasing power prediction data includes: a product-side purchasing power matrix and a product implicit factor query vector.

[0093] Specifically, the purchasing power prediction model may include: a convolutional attention encoder, an implicit factor extractor, and a convolutional attention decoder.

[0094] The convolutional attention encoder comprises: a convolutional encoder and a first attention encoding layer, the convolutional encoder comprises: at least one convolutional encoding unit, the convolutional encoding unit comprises: a convolutional layer and a pooling layer;

[0095] Based on the convolution encoder, feature extraction is performed on the customer-side purchasing power matrix and the product-side purchasing power matrix respectively to obtain customer-side purchasing power feature data and product-side purchasing power feature data.

[0096] Based on the first attention coding layer, the first customer-side feature data and the first product-side feature data are processed respectively to obtain customer interest feature data and product attraction feature data. The first customer-side feature data includes: customer-side purchasing power feature data and customer implicit factor query vector, and the first product-side feature data includes: product-side purchasing power feature data and product implicit factor query vector.

[0097] The implicit factor extractor includes two convolutional layers and a fully connected layer. Based on the first convolutional layer, the customer interest feature data can be mapped to the customer implicit factor space, and the product attraction feature data can be mapped to the product implicit factor space.

[0098] Based on the fully connected layer, customer implicit factors are extracted from the customer implicit factor space, and product implicit factors are extracted from the product implicit factor space;

[0099] Based on the second convolutional layer, the customer implicit factor space and the product implicit factor space are restored respectively to obtain customer interest feature data and product attraction feature data.

[0100] The convolutional attention decoder includes: a convolutional decoder and a second attention encoding layer, the convolutional decoder includes: convolutional decoding units with the same number as the convolutional encoding units, and each convolutional encoding unit includes: a deconvolution layer and an upsampling layer.

[0101] The second customer-side feature data and the second product-side feature data can be processed respectively based on the second attention coding layer to obtain customer-side purchasing power feature data and product-side purchasing power feature data. The second customer-side feature data includes: customer implicit factor space data and customer interest feature data output by the first convolution layer; the second product-side feature data includes: product implicit factor space data and product attraction feature data output by the first convolution layer.

[0102] Based on the convolution decoder, the customer-side purchasing power feature data and the product-side purchasing power feature data are processed respectively to obtain the customer-side purchasing power matrix and the product-side purchasing power matrix.

[0103] For details, please refer to Figure 5 As shown in FIG, the dual convolutional attention encoding, i.e., the convolutional attention encoder, includes: two convolutional layers, two pooling layers and one attention layer (i.e., the first attention encoding layer).

[0104] In specific use, two standard convolutional layers and two maximum pooling operations can be used. Through the translation equilaterality, parameter sharing and sparse interaction properties of the convolution operation, the number of parameters is greatly reduced and better local perception is obtained. The nonlinear activation function can be added after the convolution layer to learn better data representation and accurate deep local features. After the activation function output, in order to enhance the robustness of the algorithm, retain the main features after convolution and reduce the size of the feature map to avoid overfitting, the maximum pooling layer is used to realize feature extraction. The encoding formula of the corresponding convolution encoder is:

[0105]

[0106] Among them, * represents the convolution operation, ζ(·) represents the nonlinear activation function, ψ(·) represents the maximum pooling operation, and Represent the customer and product convolution kernel weights respectively, and Respectively and The number of convolution kernels increases continuously in the two encodings. 1 <k 2 , which indicates that the implicit factor features of customers and products encoded by the convolutional encoder are increasing, and more deep nonlinear representations can be learned. represents the customer side purchasing power matrix, represents the product side purchasing power matrix, It represents the customer purchasing power feature data obtained after two convolution and pooling operations. Represents the product-side purchasing power feature data obtained after two convolution and pooling operations.

[0107] Use the attention layer to calculate the customer's attention distribution function α for financial products and product features, and use the scoring function to calculate the customer's multi-peak interest and the product's attractiveness to different types of customers based on the output of the convolutional encoder in the previous step:

[0108]

[0109] Among them, w u , w v Represent the customer-based and product-based attention weight matrices respectively, Represents customer interest feature data, Indicates product attractive feature data, Indicates customer side purchasing power characteristic data, Represents the purchasing power characteristic data of the product side, Q i represents the customer implicit factor query vector, Q j Represents the product implicit factor query vector.

[0110] The attention distribution function is calculated using a bilinear model with softmax normalization as the attention scoring function. The use of a bilinear model introduces the asymmetric characteristics of customer-product interaction, making feature extraction richer and reducing the underfitting phenomenon caused by the high dimension D of the implicit factor model. The model training moves towards a more accurate direction, and the training can obtain a more accurate implicit factor model, thereby improving the accuracy of model prediction.

[0111] The implicit factor extractor first uses k3 small kernel multi-layer convolution to obtain the input mapped to the implicit factor feature space; then, a fully connected layer with nonlinear activation is used to obtain the customer and product implicit factor models; finally, convolution is used instead of deconvolution operation to map the implicit factor space back to the customer product feature encoding space, laying the foundation for the next step of the double convolution attention decoder to restore the encoding matrix. The specific calculation process is as follows:

[0112]

[0113] in, They represent the fully connected model parameters during mapping and restoration, respectively. The number of convolution kernels k3>k2, the depth increases and the number of encodings and extractions in each layer increases, making the customer-product implicit factor model It can obtain different kinds of deep and critical nonlinear representations between customers and products; FC(·) represents a single-layer feedforward neural network with a nonlinear activation function; ReShape(·) means using the same square processing mechanism of the data preprocessor layer to convert the implicit factor space model into a square convolutional layer input specification, Represents customer interest feature data and product attraction feature data, represents the bias of the model parameters on the customer side and the product side in the customer-product implicit factor model, Represents the bias of the model parameters on the client side and product side during the processing of two convolutional layers.

[0114] Dual convolutional attention decoding is a convolutional attention decoder, which includes two deconvolution layers, two upsampling layers and an attention layer (i.e., the second attention encoding layer).

[0115] As the decoding process of the attention encoding layer, the output of the customer-product implicit factor extractor is used as the self-attention scorer.

[0116]

[0117] The query vector is the output of the first convolutional layer of the customer implicit factor extractor. This allows the input encoded feature to be used as a query vector, which promotes the decoder to focus on key features and obtain the customer product purchasing power prediction matrix. Indicates the purchasing power characteristic data on the customer side and the purchasing power characteristic data on the product side, represents the attention distribution on the customer side and the product side, Represents customer interest feature data and product attraction feature data, They represent the model parameters of the attention layer on the client side and the model parameters of the attention layer on the product side respectively.

[0118] The dual convolution decoder is the corresponding process of the dual convolution encoder, that is, it is necessary to decode the deep nonlinear implicit factor model after attention encoding into a similar input matrix, so that the decoder can verify the importance of the deep local features extracted by the encoder. The design uses two convolutional layers as deconvolution layers and uses a bilinear interpolation filling method to complete the decoding matrix. This design is the reverse process of the encoder convolution pooling, avoiding the decoding error caused by the artifacts caused by the deconvolution operation on the reconstructed matrix, and enhancing the robustness and accuracy of the purchasing power prediction method:

[0119]

[0120] where θ(·) represents the bilinear interpolation upsampling operation. 2 >k 1 Corresponding to the number of convolution kernels in the encoder stage, the number of convolution kernels in the decoder stage gradually decreases, which makes it possible to decode and restore the input matrix representation. Finally, a fully connected layer is used to output the restored decoding matrix and represents the customer side purchasing power matrix, represents the product side purchasing power matrix, Represents the model parameters of the two deconvolutions on the client side, represents the model parameters of the two deconvolutions on the product side, Indicates customer side purchasing power characteristic data, Indicates product-side purchasing power characteristic data, represents the bias of the client-side model parameters, Represents the bias of the product-side model parameters.

[0121] Finally, based on the optimizer, the model parameters of the convolutional attention encoder, implicit factor extractor, and convolutional attention decoder are updated with the goal of minimizing the value of the target loss function. Specifically, the target loss function of the dual convolutional attention autoencoder is a square error function with an L2 regularization term, and the Adam optimizer with a degraded learning rate is used for optimization learning until the model converges or reaches a predefined number of iterations:

[0122]

[0123] Among them, λ is the training parameter vector, X i is the purchasing power matrix of the customer side, X′ i is the customer side purchasing power matrix predicted by the purchasing power prediction model, Y j is the purchasing power matrix on the product side, Y′ j It is the purchasing power matrix of the product side predicted by the purchasing power prediction model.

[0124] 405. Based on the trained purchasing power prediction model, conduct risk assessment on the purchasing power of target customers for purchasing financial products.

[0125] Specifically, the model-based collaborative filtering matrix decomposition method is finally used to predict customer i's purchasing power rating for product j as follows:

[0126]

[0127] in, and They represent the customer latent factor model and product latent factor model after training convergence respectively.

[0128] This customer purchasing power prediction method uses a dual convolutional autoencoder with an attention mechanism to capture more important deep nonlinear implicit factors, which helps the training of the dual convolutional autoencoder, avoids model overfitting, and extracts more accurate features. The dual convolutional autoencoder can be mapped from the feature capture space to the implicit factor space; then the predictor is used to directly perform matrix decomposition model training from the customer implicit factor model based on the customer perspective and the product implicit factor model based on the product perspective to obtain the customer purchasing power prediction matrix. The square matrix is ​​used to reduce the number of kernels in the encoder layer and the decoder layer while increasing the number of convolution kernels, which enhances the robustness of the model and the stability of the prediction results. The deviation of the customer purchasing power prediction is reduced, and the accuracy of the customer purchasing power score prediction is improved.

[0129] A method for predicting customer purchasing power provided by an embodiment of the present application is introduced above, and a device for executing the above method for predicting customer purchasing power is introduced below.

[0130] See also Figure 6 , Figure 6 This is a schematic diagram of the structure of a customer purchasing power prediction device provided in an embodiment of the present application. Figure 7 As shown, the customer purchasing power prediction device comprises:

[0131] Purchasing power data acquisition module 601, used to acquire first purchasing power data representing the risk level of the customer when purchasing the product during the statistical period, and second purchasing power data obtained based on the scoring questionnaire and the final transaction operation;

[0132] A first data preprocessing module 602 is used to split the first purchasing power data into a customer side and a product side to obtain a customer side purchasing power matrix and a product side purchasing power matrix;

[0133] A second data preprocessing module 603, used to preprocess the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix;

[0134] A prediction model updating module 604 is used to train a purchasing power prediction model based on customer-side purchasing power prediction data and product-side purchasing power prediction data, wherein the customer-side purchasing power prediction data includes: the customer-side purchasing power matrix and the customer implicit factor query vector, and the product-side purchasing power prediction data includes: the product-side purchasing power matrix and the product implicit factor query vector; and

[0135] The purchasing power prediction module 605 is used to conduct risk assessment on the purchasing power of target customers for purchasing financial products based on the trained purchasing power prediction model.

[0136] In a possible implementation, the purchasing power prediction model includes: a convolutional attention encoder, an implicit factor extractor, and a convolutional attention decoder. The prediction model updating module 604 trains the purchasing power prediction model based on the customer-side purchasing power prediction data and the product-side purchasing power prediction data, including:

[0137] Based on the convolutional attention encoder, feature extraction is performed on the customer-side purchasing power prediction data and the product-side purchasing power prediction data to obtain customer interest feature data and product attraction feature data;

[0138] Based on the implicit factor extractor, extracting customer implicit factors and product implicit factors from the customer interest feature data and the product attraction feature data respectively;

[0139] Based on the convolutional attention decoder, the customer interest feature data and the product attraction feature data are restored respectively.

[0140] In a possible implementation, the process of training the purchasing power prediction model by the prediction model updating module 604 based on the customer-side purchasing power prediction data and the product-side purchasing power prediction data further includes:

[0141] Based on the optimizer aiming to minimize the value of the target loss function, the model parameters of the convolutional attention encoder, the implicit factor extractor and the convolutional attention decoder are updated.

[0142] In a possible implementation, the convolutional attention encoder includes: a convolutional encoder and a first attention encoding layer, the convolutional encoder includes: at least one convolutional encoding unit, the convolutional encoding unit includes: a convolutional layer and a pooling layer; the prediction model update module 604 extracts features from the customer-side purchasing power prediction data and the product-side purchasing power prediction data based on the convolutional attention encoder to obtain customer interest feature data and product attraction feature data, including:

[0143] Based on the convolution encoder, feature extraction is performed on the customer-side purchasing power matrix and the product-side purchasing power matrix respectively to obtain customer-side purchasing power feature data and product-side purchasing power feature data;

[0144] Based on the first attention encoding layer, the first customer-side feature data and the first product-side feature data are processed respectively to obtain the customer interest feature data and the product attraction feature data, the first customer-side feature data including: the customer-side purchasing power feature data and the customer implicit factor query vector, the first product-side feature data including: the product-side purchasing power feature data and the product implicit factor query vector.

[0145] In a possible implementation, the implicit factor extractor includes: two convolutional layers and a fully connected layer, and the prediction model updating module 604 extracts the customer implicit factors and the product implicit factors based on the customer interest feature data and the product attraction feature data by the implicit factor extractor, respectively, including:

[0146] Mapping the customer interest feature data to the customer implicit factor space based on the first convolutional layer, and mapping the product attraction feature data to the product implicit factor space;

[0147] Extracting the customer implicit factor from the customer implicit factor space based on the fully connected layer, and extracting the product implicit factor from the product implicit factor space;

[0148] Based on the second convolutional layer, the customer implicit factor space and the product implicit factor space are restored respectively to obtain the customer interest feature data and the product attraction feature data.

[0149] In a possible implementation, the convolutional attention decoder includes: a convolutional decoder and a second attention encoding layer, the convolutional decoder includes: convolutional decoding units having the same number as the convolutional encoding units, each of the convolutional encoding units includes: a deconvolution layer and an upsampling layer; the prediction model updating module 604 performs a process of restoring the customer interest feature data and the product attraction feature data based on the convolutional attention decoder, respectively, including:

[0150] Based on the second attention coding layer, the second customer-side feature data and the second product-side feature data are processed respectively to obtain the customer-side purchasing power feature data and the product-side purchasing power feature data, wherein the second customer-side feature data includes: the customer implicit factor space data output by the first convolution layer and the customer interest feature data, and the second product-side feature data includes: the product implicit factor space data output by the first convolution layer and the product attraction feature data;

[0151] The customer-side purchasing power feature data and the product-side purchasing power feature data are processed respectively based on the convolution decoder to obtain the customer-side purchasing power matrix and the product-side purchasing power matrix.

[0152] In a possible implementation, the second data preprocessing module 603 preprocesses the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix, including:

[0153] Based on a two-layer fully connected neural network with an activation function and a normalization function, the second purchasing power data is processed to obtain the customer implicit factor query vector and the product implicit factor query vector.

[0154] The present application also provides an electronic device in an embodiment. Figure 7 As shown, it shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present application. The electronic device in the embodiments of the present application may include but is not limited to fixed terminals such as mobile phones, notebook computers, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 7 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.

[0155] like Figure 7As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 into a random access memory (RAM) 703. 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 703. The processing device 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0156] Typically, the following devices may be connected to the I / O interface 705: an input device 706 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 707 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 708 including, for example, a memory card, a hard disk, etc.; and a communication device 709. The communication device 709 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 7 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.

[0157] An embodiment of the present application also provides a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any one of the customer purchasing power prediction methods provided in the embodiments of the present application.

[0158] 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 the one or more computer programs are executed by an electronic device, the electronic device can implement any customer purchasing power prediction method provided in the embodiment of the present application.

[0159] It is understandable that before using the technical solutions disclosed in the embodiments of this disclosure, the types, scope of use, usage scenarios, etc. of personal information involved in this disclosure should be informed to customers and their authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0160] 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.

[0161] 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.

[0162] 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.

[0163] 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 predicting customer purchasing power, characterized in that: include: Obtaining first purchasing power data representing the risk level of customers when purchasing products during the statistical period, and second purchasing power data obtained based on the scoring questionnaire and the final transaction operation; Splitting the first purchasing power data into a customer side and a product side to obtain a customer side purchasing power matrix and a product side purchasing power matrix; Preprocessing the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix; Training a purchasing power prediction model based on customer-side purchasing power prediction data and product-side purchasing power prediction data, wherein the customer-side purchasing power prediction data includes: the customer-side purchasing power matrix and the customer implicit factor query vector, and the product-side purchasing power prediction data includes: the product-side purchasing power matrix and the product implicit factor query vector; Based on the trained purchasing power prediction model, the purchasing power of target customers for financial products is assessed for risk.

2. The method for predicting customer purchasing power according to claim 1, characterized in that: The purchasing power prediction model includes: a convolutional attention encoder, an implicit factor extractor and a convolutional attention decoder. The purchasing power prediction model is trained based on the customer-side purchasing power prediction data and the product-side purchasing power prediction data, including: Based on the convolutional attention encoder, feature extraction is performed on the customer-side purchasing power prediction data and the product-side purchasing power prediction data to obtain customer interest feature data and product attraction feature data; Based on the implicit factor extractor, extracting customer implicit factors and product implicit factors from the customer interest feature data and the product attraction feature data respectively; Based on the convolutional attention decoder, the customer interest feature data and the product attraction feature data are restored respectively.

3. The method for predicting customer purchasing power according to claim 2, characterized in that: The training of the purchasing power prediction model based on the customer-side purchasing power prediction data and the product-side purchasing power prediction data also includes: Based on the optimizer aiming to minimize the value of the target loss function, the model parameters of the convolutional attention encoder, the implicit factor extractor and the convolutional attention decoder are updated.

4. The method for predicting customer purchasing power according to claim 2, characterized in that: The convolutional attention encoder comprises: a convolutional encoder and a first attention encoding layer, the convolutional encoder comprises: at least one convolutional encoding unit, the convolutional encoding unit comprises: a convolutional layer and a pooling layer; based on the convolutional attention encoder, the customer-side purchasing power prediction data and the product-side purchasing power prediction data are respectively subjected to feature extraction to obtain customer interest feature data and product attraction feature data, including: Based on the convolution encoder, feature extraction is performed on the customer-side purchasing power matrix and the product-side purchasing power matrix respectively to obtain customer-side purchasing power feature data and product-side purchasing power feature data; Based on the first attention encoding layer, the first customer-side feature data and the first product-side feature data are processed respectively to obtain the customer interest feature data and the product attraction feature data, the first customer-side feature data including: the customer-side purchasing power feature data and the customer implicit factor query vector, the first product-side feature data including: the product-side purchasing power feature data and the product implicit factor query vector.

5. The method for predicting customer purchasing power according to claim 4, characterized in that: The implicit factor extractor includes: two convolutional layers and a fully connected layer. The implicit factor extractor extracts the customer interest feature data and the product attraction feature data based on the customer interest feature data and the product attraction feature data, respectively, and includes: Mapping the customer interest feature data to the customer implicit factor space based on the first convolutional layer, and mapping the product attraction feature data to the product implicit factor space; Extracting the customer implicit factor from the customer implicit factor space based on the fully connected layer, and extracting the product implicit factor from the product implicit factor space; Based on the second convolutional layer, the customer implicit factor space and the product implicit factor space are restored respectively to obtain the customer interest feature data and the product attraction feature data.

6. The method for predicting customer purchasing power according to claim 5, characterized in that: The convolutional attention decoder includes: a convolutional decoder and a second attention coding layer, the convolutional decoder includes: convolutional decoding units having the same number as the convolutional coding units, each of the convolutional coding units includes: a deconvolution layer and an upsampling layer; based on the convolutional attention decoder, the customer interest feature data and the product attraction feature data are restored respectively, including: Based on the second attention coding layer, the second customer-side feature data and the second product-side feature data are processed respectively to obtain the customer-side purchasing power feature data and the product-side purchasing power feature data, wherein the second customer-side feature data includes: the customer implicit factor space data output by the first convolution layer and the customer interest feature data, and the second product-side feature data includes: the product implicit factor space data output by the first convolution layer and the product attraction feature data; The customer-side purchasing power feature data and the product-side purchasing power feature data are processed respectively based on the convolution decoder to obtain the customer-side purchasing power matrix and the product-side purchasing power matrix.

7. The method for predicting customer purchasing power according to any one of claims 1 to 6, characterized in that: Preprocessing the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix includes: Based on a two-layer fully connected neural network with an activation function and a normalization function, the second purchasing power data is processed to obtain the customer implicit factor query vector and the product implicit factor query vector.

8. A customer purchasing power prediction device, characterized in that: include: A purchasing power data acquisition module is used to acquire first purchasing power data representing the risk level of customers when purchasing products during a statistical period, and second purchasing power data obtained based on a scoring questionnaire and a final transaction operation; A first data preprocessing module, used for splitting the first purchasing power data into a customer side and a product side to obtain a customer side purchasing power matrix and a product side purchasing power matrix; A second data preprocessing module, used for preprocessing the second purchasing power data into a customer implicit factor query vector corresponding to the customer-side purchasing power matrix and a product implicit factor query vector corresponding to the product-side purchasing power matrix; A prediction model updating module, used for training a purchasing power prediction model based on customer-side purchasing power prediction data and product-side purchasing power prediction data, wherein the customer-side purchasing power prediction data includes: the customer-side purchasing power matrix and the customer implicit factor query vector, and the product-side purchasing power prediction data includes: the product-side purchasing power matrix and the product implicit factor query vector; and The purchasing power prediction module is used to conduct risk assessment on the purchasing power of target customers for purchasing financial products based on the trained purchasing power prediction model.

9. An electronic device, 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 household purchasing power prediction method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: The storage medium carries one or more computer programs, and when the one or more computer programs are executed by an electronic device, the electronic device can implement the user purchasing power prediction method as described in any one of claims 1 to 7.

Images