Method, device, electronic equipment and computer program product for determining a financial product

Abstract

This application discloses a method, apparatus, electronic device, and computer program product for determining financial products. Relating to the field of fintech, the method includes: obtaining user profile data of the target user with the user's authorization, and obtaining financial text information of the target user, wherein the financial text information refers to the financial product demand text sent by the target user through a client; determining M candidate financial products based on the financial text information; inputting the M candidate financial products and the user profile data into a product screening model, processing to obtain a product screening result, wherein the product screening model filters the M candidate financial products based on the user profile data, and the product screening result refers to the target financial product selected from the M candidate financial products; and sending the target financial product associated with the product screening result to the client. This application solves the technical problem of low screening accuracy in related technologies when screening financial products.

Description

Technical Field

[0001] This application relates to the field of financial technology, and more specifically, to a method, apparatus, electronic device, and computer program product for determining a financial product. Background Technology

[0002] In recent years, with the rapid development of technologies such as big data and artificial intelligence, financial institutions have widely applied AI (Artificial Intelligence) technology in areas such as intelligent customer service, driving the evolution of financial services towards personalization and intelligence. However, in the scenario of personalized operation of financial products, existing recommendation models based on deep reinforcement learning still face multiple technical bottlenecks, making it difficult to meet the business requirements of high security, strong robustness, and real-time responsiveness.

[0003] Traditional Deep Q-Networks (DQNs), as a value function-based reinforcement learning method, have achieved some success in decision-making tasks in discrete action spaces, but they have significant limitations in financial operation scenarios. First, DQNs use a single network architecture to simultaneously complete action selection and target Q-value calculation, leading to a systematic overestimation of occasional high-return actions and ignoring their long-term volatility risks.

[0004] Furthermore, traditional deep Q-networks require a discretized action space, while financial product operations require a continuous decision space, meaning that demand needs to be continuously distributed. Forced discretization not only easily leads to coarse-grained financial product operations and frequent suboptimal solutions, but also causes the model parameter size to grow exponentially when refining the action space to improve accuracy, resulting in a sharp decline in training efficiency and making it difficult to deploy in real-world systems. In addition, financial user profile data encompasses multi-source heterogeneous features. Traditional deep Q-networks, relying on fully connected neural networks for feature extraction, struggle to effectively model temporal dependencies and nonlinear interactions, leading to the loss of key behavioral pattern information. Moreover, the model update cycle is on a daily or weekly basis, failing to capture real-time changes in user risk preferences under unexpected events, thus causing severe strategy misalignment and significantly reducing user trust and service satisfaction.

[0005] There is currently no effective solution to the technical problem of low accuracy in screening financial products using related technologies. Summary of the Invention

[0006] The main objective of this application is to provide a method, apparatus, electronic device, and computer program product for identifying financial products, so as to solve the technical problem of low screening accuracy in the screening of financial products in related technologies.

[0007] To achieve the above objectives, according to one aspect of this application, a method for determining a financial product is provided. The method includes: obtaining user profile data of the target user with the user's authorization, and obtaining financial text information of the target user, wherein the financial text information refers to a financial product demand text sent by the target user through a client; determining M candidate financial products based on the financial text information; inputting the M candidate financial products and the user profile data into a product screening model, and processing to obtain a product screening result, wherein the product screening model filters the M candidate financial products based on the user profile data to obtain a product screening result, the product screening result referring to the target financial product selected from the M candidate financial products, where M is a positive integer; and sending the target financial product associated with the product screening result to the client.

[0008] Optionally, obtaining user profile data of the target user includes: obtaining financial behavior information of the target user within a preset time period, and standardizing the financial behavior information to obtain financial behavior data; obtaining user information of the target user, converting the user information into a format to obtain user data; and combining the financial behavior data and user data to obtain user profile data.

[0009] Optionally, the M candidate financial products and user profile data are input into the product screening model, and the product screening results are obtained by processing them as follows: the feature processing layer of the product screening model integrates the M candidate financial products and user profile data to obtain M feature sequences; the fully connected layer of the product screening model performs nonlinear transformation on the M feature sequences to obtain M decision feature vectors; and the output layer of the product screening model processes the M decision feature vectors to output the product screening results.

[0010] Optionally, the feature processing layer includes an input layer and a long short-term memory layer. The feature processing layer of the product screening model integrates M candidate financial products and user profile data to obtain M feature sequences, including: the input layer performs vector transformation on user data in the user profile data to obtain user feature vectors; the input layer performs vector transformation on financial behavior data in the user profile data to obtain behavior feature vectors; the input layer performs vector transformation on each candidate financial product to obtain M product feature vectors; the user feature vectors and behavior feature vectors are aggregated to obtain processed feature vectors; for a candidate financial product, the sliding window parameters are obtained, and the long short-term memory layer combines the processed feature vectors and the corresponding product feature vectors of the candidate financial product according to the sliding window parameters to obtain combined feature vectors; the combined feature vectors are then transformed to obtain the feature sequence of the candidate financial products.

[0011] Optionally, the fully connected layer includes a first fully connected layer and a second fully connected layer. The fully connected layer of the product screening model performs nonlinear transformations on the M feature sequences to obtain M decision feature vectors, including: for a feature sequence, the first fully connected layer performs a first nonlinear transformation on the feature sequence to obtain an initial decision feature vector; the second fully connected layer performs a second nonlinear transformation on the initial decision feature vector to obtain the decision feature vector of the feature sequence.

[0012] Optionally, before the product selection model's output layer processes the M decision feature vectors and outputs the product selection result, the method further includes: for a decision feature vector, the dual network layers of the product selection model construct a value function based on the decision feature vector to obtain a set of value function values. This set of value function values ​​includes state value function values ​​and advantage function values. The state value function value refers to the financial value data when the target user uses the candidate financial product corresponding to the decision feature vector, and the advantage function value indicates the ratio of the financial value of the target user using the candidate financial product corresponding to the decision feature vector compared to the target user using other candidate financial products. The product selection model's output layer processes the M decision feature vectors and outputs the product selection result by: aggregating the state value function value and advantage function value of each decision feature vector by the output layer to obtain M product function values; sorting the M product function values ​​in descending order to obtain a function value sequence; and determining the product selection result based on the candidate financial product corresponding to the product function value at a preset position in the function value sequence.

[0013] Optionally, the product selection model is trained as follows: obtain the historical product selection results of N users within a historical time period, and obtain the user profile data of each user to obtain N sets of historical user profile data, where N is a positive integer; construct a model training set based on the historical financial products corresponding to the N historical product selection results and the N sets of historical user profile data; train the preset product selection model using the model training set to obtain the trained product selection model; if the loss function value corresponding to the trained product selection model is less than the preset data, the trained product selection model is determined as the product selection model.

[0014] To achieve the above objectives, according to another aspect of this application, a financial product determination apparatus is provided. The apparatus includes: an acquisition unit, configured to acquire user profile data of the target user and acquire financial text information of the target user, with the user's authorization, wherein the financial text information refers to a financial product demand text sent by the target user through a client; a determination unit, configured to determine M candidate financial products based on the financial text information, input the M candidate financial products and the user profile data into a product screening model, and process the data to obtain a product screening result, wherein the product screening model filters the M candidate financial products based on the user profile data to obtain a product screening result, the product screening result referring to the target financial product selected from the M candidate financial products, where M is a positive integer; and a sending unit, configured to send the target financial product associated with the product screening result to the client.

[0015] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, the device where the computer-readable storage medium is located executes any of the above-mentioned financial products.

[0016] According to another aspect of the present invention, an electronic device is also provided, including one or more processors and a memory, the memory storing an executable program, and the processor for running the program, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining any of the above-described financial products.

[0017] According to another aspect of the present invention, a computer program product is also provided, the computer program product including a computer program, wherein the computer program, when executed by a processor, implements the method for determining any of the above-mentioned financial products.

[0018] In this embodiment, a method for determining financial products is adopted. With the authorization of the target user, user profile data of the target user is obtained, along with the target user's financial text information, which refers to the financial product demand text sent by the target user through a client. Based on the financial text information, M candidate financial products are determined. These M candidate financial products and the user profile data are input into a product filtering model, and the product filtering result is obtained. The product filtering model filters the M candidate financial products based on the user profile data, and the product filtering result refers to the target financial product selected from the M candidate financial products, where M is a positive integer. The target financial product associated with the product filtering result is sent to the client. This solves the technical problem of low filtering accuracy in related technologies. By inputting the candidate financial products determined based on the financial text information and the user profile data into the product filtering model and processing them to obtain the product filtering result, the technical effect of improving the filtering accuracy of financial products is achieved. Attached Figure Description

[0019] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0020] Figure 1 It is a hardware structure block diagram of a computer terminal (or mobile device) used to implement a method for determining financial products;

[0021] Figure 2 This is a flowchart of a method for determining financial products according to embodiments of this application;

[0022] Figure 3 This is a schematic diagram of a product screening model provided according to an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of a device for determining financial products according to an embodiment of this application;

[0024] Figure 5 This is a structural block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] It should be noted that all information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) involved in this application are information and data authorized by the user or fully authorized by all parties. For example, this system has interfaces with relevant users or organizations to provide users with corresponding operation data for them to choose to agree to or refuse automated decision-making results. Before obtaining relevant information, a request for obtaining the information needs to be sent to the aforementioned user or organization through the interface, and the relevant information is obtained after receiving consent from the aforementioned user or organization; if the user chooses to refuse, the expert decision-making process is initiated. Users can view the purpose of data use in real time through authorization decoding and have the right to withdraw authorization or delete data at any time. After the authorization is withdrawn, the system will terminate the relevant data processing within 24 hours.

[0028] It should be noted that the information collected in this application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize use or refuse use.

[0029] Example 1

[0030] According to an embodiment of this application, a method embodiment for determining a financial product is also provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0031] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 This is a hardware structure block diagram of a computer terminal (or mobile device) used to implement a method for determining financial products, such as... Figure 1 As shown, computer terminal 10 (or mobile device) may include one or more ( Figure 1 The processor 102 (which may include, but is not limited to, a microprocessor MCU (Microcontroller Unit) or a programmable gate array (FPGA)) is shown as 102a, 102b, ..., 102n. It also includes a memory 104 for storing data and a transmission device 106 for communication functions. In addition, it may include: a display, an input / output interface, a Universal Serial Bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a keyboard, a cursor control device, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0032] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0033] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the financial product determination method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned financial product determination method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0034] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network interface controller (NIC) and a network interface, which can be connected to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a radio frequency (RF) module, used for wireless communication with the Internet.

[0035] The display can be, for example, a touchscreen liquid crystal display (LCD), which allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0036] Under the aforementioned operating environment, this application provides the following: Figure 2 The method for determining the financial products shown. Figure 2 This is a flowchart of a method for determining financial products according to embodiments of this application, such as... Figure 2 As shown, the method includes the following steps:

[0037] Step S201: With the authorization of the target user, obtain the user profile data of the target user and obtain the financial text information of the target user, wherein the financial text information refers to the financial product demand text sent by the target user through the client.

[0038] It should be noted that in order to provide precise financial products to target users, their authorization is required beforehand. The target user can complete a privacy authorization signing through the client interface. After authorization, relevant user data is obtained, i.e., user profile data. This user profile data can include static attributes (age, credit score, etc.) and dynamic behaviors (such as transaction frequency). Financial text information is also obtained to indicate which financial products should be offered to the target user, thereby alleviating the rigidity in product screening caused by relying on historical information and assisting in product selection.

[0039] Step S202: Determine M candidate financial products based on financial text information, input the M candidate financial products and user profile data into the product screening model, and process them to obtain the product screening result. The product screening model filters the M candidate financial products based on the user profile data to obtain the product screening result. The product screening result refers to the target financial product selected from the M candidate financial products, where M is a positive integer.

[0040] It should be noted that the product selection model refers to an improved Deep Q-Network architecture, which can be a neural network system that integrates LSTM (Long Short-Term Memory) and Deep Q-Network (DQN). LSTM can capture user behavior trends and solve the model lag problem, while Deep Q-Network can accurately separate the user's state from the product's quality. After obtaining the target user's profile data and financial text information, keywords can first be extracted from the financial text information, and then matched with the financial products of financial institutions to output multiple candidate financial products.

[0041] Furthermore, the candidate financial products and user profile data selected above are input into the trained product selection model, which evaluates each candidate financial product and outputs the product selection results, including the optimal target financial product.

[0042] Step S203: Send the target financial products associated with the product screening results to the client.

[0043] Specifically, after obtaining the product screening results output by the product screening model, the product screening results and related information can be pushed to the target user's client through communication channels (such as HTTPS (Hypertext Transfer Protocol Secure) and static key authentication).

[0044] The method for determining financial products provided in this application, with the authorization of the target user, obtains the target user's user profile data and financial text information, wherein the financial text information refers to the financial product demand text sent by the target user through a client; M candidate financial products are determined based on the financial text information; the M candidate financial products and user profile data are input into a product screening model, and the product screening result is obtained; wherein the product screening model filters the M candidate financial products based on the user profile data, and the product screening result refers to the target financial product selected from the M candidate financial products, where M is a positive integer; the target financial product associated with the product screening result is sent to the client. This method solves the technical problem of low screening accuracy in related technologies. By inputting the candidate financial products determined based on the financial text information and user profile data into the product screening model and processing them to obtain the product screening result, the technical effect of improving the screening accuracy of financial products is achieved.

[0045] Optionally, in the method for determining financial products provided in this application embodiment, obtaining user profile data of the target user includes: obtaining financial behavior information of the target user within a preset time period, and standardizing the financial behavior information to obtain financial behavior data; obtaining user information of the target user, converting the user information into a format to obtain user data; and combining the financial behavior data and user data to obtain user profile data.

[0046] Specifically, user profile data for target users can include various types. When acquiring user profile data, the first step is to obtain the target user's financial behavior information within a preset time period. This financial behavior information can refer to the user's operational records related to financial products, such as transaction amounts and frequencies over the past 30 days. Then, the financial behavior information is standardized, meaning the raw financial behavior information is numerically normalized to have the same dimensions and distribution range, thus obtaining financial behavior data. Simultaneously, user information for the target user can also be acquired, such as the user's basic static attributes. This user information is then processed in the same way, for example, by format conversion, transforming unstructured or semi-structured information into numerical vectors or categorical codes, thus obtaining user data. Finally, the two types of data are combined, such as by concatenating them in a fixed order, to form a unified, fixed-length user profile data set.

[0047] This embodiment constructs user profile data by collecting, standardizing, formatting, and combining user static attributes and dynamic behavior data. This enables the subsequent product screening model based on deep Q-networks to accurately input state variables containing multi-dimensional behavioral features, effectively overcoming the problems of feature extraction distortion and model convergence difficulties caused by chaotic data formats and inconsistent dimensions in traditional recommendation systems. This provides a reliable data foundation for screening accurate, stable, and interpretable financial products.

[0048] Optionally, in the method for determining financial products provided in this application embodiment, inputting M candidate financial products and user profile data into a product screening model and processing them to obtain product screening results includes: integrating the M candidate financial products and user profile data by the feature processing layer of the product screening model to obtain M feature sequences; performing nonlinear transformation on the M feature sequences by the fully connected layer of the product screening model to obtain M decision feature vectors; and processing the M decision feature vectors by the output layer of the product screening model to output the product screening results.

[0049] It should be noted that due to the inefficiency of deep Q-network architecture, a long short-term memory network structure (i.e., long short-term memory layer) can be introduced. This structure includes three types of gates: forget gate, input gate, and output gate. When new information is encountered, the input gate of the long short-term memory network structure can be used to control the threshold of input information entering the cell, the forget gate can be used to control the degree of appearance of the input value in the cell, and the output gate can be used to use the stored input values ​​to calculate the output activation.

[0050] Specifically, after obtaining candidate financial products and user profile data, Figure 3 This is a schematic diagram of the product screening model provided according to the embodiments of this application, such as... Figure 3 As shown, the first step involves integrating candidate financial products and user profile data using a feature processing layer (including an input layer and an LSTM layer) that incorporates the aforementioned long short-term memory layer. This means using user profile data as a common context and concatenating it with the features of each candidate financial product to form multiple sets of user-product paired feature sequences. Then, a fully connected layer used for feature abstraction performs a non-linear transformation on each of these feature sequences. This fully connected layer (FC layer) can consist of multiple neurons. The non-linear transformation refers to mapping the input sequence layer by layer through multiple fully connected layers to extract higher-order relationships, thereby outputting corresponding decision feature vectors. Each vector corresponds to an abstract representation of a candidate financial product, with a lower dimensionality than the input but higher information density, providing effective input for subsequent Q-value calculation.

[0051] Furthermore, the output layer of the product selection model calculates the Q-value of each candidate financial product through the decision feature vector, and then determines the product selection result based on the Q-value, realizing a decision-making closed loop from the candidate list to the optimal recommendation. The output layer can be based on the Dueling Double DQN structure and can contain two sub-networks: state value stream and action advantage stream. The Q-value reflects the expected comprehensive return of choosing the product in the current user state.

[0052] This embodiment achieves a systematic evaluation of multiple candidate financial products through the collaborative processing of the feature processing layer, the fully connected layer, and the output layer. This enables the model to dynamically determine the matching degree of each product based on the user profile, avoiding the crude decision-making method that relies solely on the product's historical performance or a single-dimensional user label. Thus, while ensuring the accuracy of recommendations, it also enhances the adaptability to user behavior trends.

[0053] Optionally, in the method for determining financial products provided in this application embodiment, the feature processing layer includes an input layer and a long short-term memory layer. The feature processing layer of the product screening model integrates M candidate financial products and user profile data to obtain M feature sequences, including: the input layer performing vector transformation on user data in the user profile data to obtain user feature vectors; performing vector transformation on financial behavior data in the user profile data to obtain behavior feature vectors; performing vector transformation on each candidate financial product to obtain M product feature vectors; aggregating the user feature vectors and behavior feature vectors to obtain processed feature vectors; for a candidate financial product, obtaining sliding window parameters; the long short-term memory layer combining the processed feature vectors and the corresponding product feature vectors of the candidate financial product according to the sliding window parameters to obtain combined feature vectors; and transforming the combined feature vectors to obtain the feature sequence of the candidate financial product.

[0054] Specifically, the feature processing layer can include an input layer and a long short-term memory layer. When processing data in the feature processing layer, the input layer can first perform vector transformation on the user data and financial behavior data in the user profile data to obtain a fixed-dimensional user feature vector and a behavior feature vector for each candidate financial product. Then, each candidate financial product is vector transformed to obtain a product feature vector. Vector transformation refers to mapping discrete or scalar values ​​into dense vectors of fixed length, thereby preserving the time-series features of user behavior.

[0055] Furthermore, to construct the feature sequence, the user feature vector and behavior feature vector need to be concatenated or added sequentially to form a unified feature vector, i.e., the processed feature vector. Then, the time window length and step size for extracting the user's historical behavior sequence are obtained, i.e., the sliding window parameters are acquired. The Long Short-Term Memory (LSTM) layer uses these parameters to combine the processed feature vector with the product feature vector corresponding to each candidate financial product, resulting in a combined feature vector. The LSM layer is a recurrent neural network structure used to model time-series dependencies; its input is the sequence data, and the combined feature vector serves as the input for each time step. Finally, the LSM layer transforms the combined feature vector to obtain the feature sequence of the candidate financial products, i.e., time-step processing of the combined feature vector, outputting the hidden state of the last time step (i.e., a condensed representation of the sequence), thus obtaining the feature sequence.

[0056] This embodiment uses a collaborative design of the input layer and the long short-term memory layer to perform time-series fusion of user static attributes, dynamic behavior sequences and candidate financial product features, thereby effectively capturing the complex relationships between users, solving the problem that traditional models cannot handle the linkage between time-series behavior and product attributes, and improving the model's perception capabilities.

[0057] Optionally, in the method for determining financial products provided in the embodiments of this application, the fully connected layer includes a first fully connected layer and a second fully connected layer. The fully connected layer of the product screening model performs nonlinear transformation on M feature sequences to obtain M decision feature vectors, including: for a feature sequence, the first fully connected layer performs a first nonlinear transformation on the feature sequence to obtain an initial decision feature vector; the second fully connected layer performs a second nonlinear transformation on the initial decision feature vector to obtain a decision feature vector of the feature sequence.

[0058] It should be noted that a fully connected layer can include a first fully connected layer and a second fully connected layer. These are structural layers in a neural network where each neuron is connected to all neurons in the previous layer. The first and second fully connected layers are connected in series, both consisting of a linear transformation followed by an activation function. Specifically, after obtaining the feature sequence output by the feature processing layer, the first fully connected layer first performs a nonlinear transformation on the feature sequence to obtain the initial decision feature vector. This involves linearly mapping the feature sequence (e.g., matrix multiplication) and then introducing nonlinearity through the ReLU activation function, outputting an intermediate vector, which is the initial decision feature vector.

[0059] Furthermore, the second fully connected layer performs a linear transformation and ReLU activation on the initial decision feature vector output by the first layer again, outputting the final decision feature vector, thereby further enhancing the feature discrimination capability and making the output vector more focused on useful semantic information.

[0060] This embodiment sets up two cascaded fully connected layers to perform step-by-step nonlinear transformations on the feature sequence, enabling the model to extract and strengthen the deep correlation features between users and products layer by layer. This avoids the problem of insufficient expressive power of single-layer networks, improves the characterization accuracy of decision feature vectors, and provides a stable and effective intermediate representation for the subsequent output layer to accurately evaluate the Q value of each candidate financial product.

[0061] Optionally, in the method for determining financial products provided in this application embodiment, before the output layer of the product screening model processes the M decision feature vectors and outputs the product screening result, the method further includes: for a decision feature vector, the dual network layer of the product screening model constructs a value function based on the decision feature vector to obtain a set of value function values, wherein the set of value function values ​​includes state value function values ​​and advantage function values. The state value function value refers to the financial value data when the target user uses the candidate financial product corresponding to the decision feature vector, and the advantage function value is used to indicate the ratio of the financial value of the target user using the candidate financial product corresponding to the decision feature vector to the target user using other candidate financial products. The processing of the M decision feature vectors by the output layer of the product screening model to output the product screening result includes: the output layer aggregating the state value function value and advantage function value of each decision feature vector to obtain M product function values; sorting the M product function values ​​in descending order to obtain a function value sequence; and determining the product screening result based on the candidate financial product corresponding to the product function value at a preset position in the function value sequence.

[0062] like Figure 3 As shown, the product selection model can also include a dual network layer (A&V layer), which can consist of two parallel sub-networks that simultaneously calculate state value and action advantage. Before data processing by the output layer, the decision feature vector of each candidate financial product needs to be processed by the aforementioned dual network layer. Specifically, since the dual network layer uses two identical Q-networks, one for selecting actions (i.e., selecting candidate financial products) and the other for evaluating the value of the current action, the dual network layer can construct a value function based on the decision feature vector to obtain the state value function value and the advantage function value. Both are input from the same decision feature vector and output through two independent fully connected sub-networks, thus decomposing into two independent dimensions: the quality of the environmental state and the superiority or inferiority of the action relative to other actions.

[0063] It should be noted that the state value function value Used to evaluate the overall value of a user's current state, allowing the model to distinguish whether a user's state itself is advantageous. Advantage function value It measures the relative merits of a candidate financial product compared to other candidate financial products in the current user state, enabling the model to identify which product is more suitable for the current user, rather than just looking at the product's historical performance, thereby achieving truly personalized recommendations.

[0064] Furthermore, after obtaining the above set of value function values, these function values ​​can be aggregated to obtain the product function value (i.e., Q value) for each candidate financial product, which can be calculated using the following formula:

[0065]

[0066] in, Let be the state function over time. For actions (i.e., each candidate financial product). For network parameters, For the next step, For state value function value, The advantage function value is based on the state. Finally, the product function values ​​of each candidate financial product are arranged from high to low to form a function value sequence, and the product with the highest ranking (e.g., the one at the preset position) is selected as the final product screening result.

[0067] This embodiment decomposes the Q-value into state value and action advantage through a dual network layer, and calculates the product function value of each candidate financial product based on the aggregation formula. This makes the screening results not only reflect the return potential of the product itself, but also more accurately match the user's current preferences and behavioral trends, significantly improving the rationality, stability and interpretability of the screening.

[0068] Optionally, in the method for determining financial products provided in this application embodiment, the product screening model is trained in the following way: obtaining the historical product screening results of N users within a historical time period, and obtaining the user profile data of each user to obtain N sets of historical user profile data, where N is a positive integer; constructing a model training set based on the historical financial products corresponding to the N historical product screening results and the N sets of historical user profile data; training the preset product screening model using the model training set to obtain the trained product screening model; if the loss function value corresponding to the trained product screening model is less than the preset data, the trained product screening model is determined as the product screening model.

[0069] Specifically, before using a product screening model to screen candidate financial products, the model needs to be trained. First, historical product screening results and user profile data for multiple users can be obtained. Historical product screening results refer to the records of products that financial institutions have screened for users within a historical time period. User profile data is a standardized feature vector generated by combining user data and financial behavior data, containing static attributes and dynamic behaviors. Then, the historical financial products corresponding to the historical product screening results and the historical user profile data for each user constitute the model training set. This model is then used to train the preset product screening model. The backpropagation algorithm, using mean squared error as the loss function, can be used to minimize the deviation between the model's output Q-value and the actual true value.

[0070] Each time the model is input, it takes a set of historical user profile data and corresponding historical candidate financial products as input, outputs the Q-value of the product, compares it with the expected reward (i.e., the historical financial product corresponding to the historical product selection result), calculates the error, and updates the network parameters θ, so that the model gradually approximates the real behavior. If the loss function value corresponding to the trained product selection model is less than the preset data, the trained product selection model is determined as the product selection model. The loss function value is the average mean squared error calculated in each iteration during training, reflecting the degree of deviation between the model's predicted Q-value and the actual result. If the loss function value is lower than 0.05 or the decrease is less than 0.001 for 5 consecutive iterations, it indicates that the model has learned stably. When this condition is met, training stops, the current model parameters are saved, and it is used as the final product selection model for online use.

[0071] This embodiment utilizes real data to construct a supervised training set, uses user profiles and product features as inputs to train the product selection model end-to-end, and uses the convergence of the loss value as the stopping condition. This avoids the subjectivity and lag of manually set rules, and improves the applicability and generalization ability of the model.

[0072] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0073] Example 2

[0074] This application also provides a device for determining financial products. It should be noted that this device can be used to execute the method for determining financial products provided in this application. The following describes the device for determining financial products provided in this application.

[0075] According to an embodiment of this application, an apparatus for implementing the above-described method for determining financial products is also provided. Figure 4 This is a schematic diagram of a financial product determination device provided according to an embodiment of this application, such as... Figure 4 As shown, the device includes: an acquisition unit 40, a determination unit 41, and a transmission unit 42.

[0076] The acquisition unit 40 is used to acquire user profile data of the target user and financial text information of the target user when authorized by the target user. The financial text information refers to the text of financial product demand sent by the target user through the client.

[0077] The determining unit 41 is used to determine M candidate financial products based on financial text information, input the M candidate financial products and user profile data into the product screening model, and process them to obtain the product screening result. The product screening model filters the M candidate financial products based on the user profile data to obtain the product screening result. The product screening result refers to the target financial product selected from the M candidate financial products, where M is a positive integer.

[0078] Sending unit 42 is used to send the target financial products associated with the product screening results to the client.

[0079] The financial product determination device provided in this application embodiment obtains user profile data and financial text information of the target user with the authorization of the target user by the acquisition unit 40. The financial text information refers to the financial product demand text sent by the target user through the client. The determination unit 41 determines M candidate financial products based on the financial text information, inputs the M candidate financial products and user profile data into the product screening model, and processes them to obtain the product screening result. The product screening model filters the M candidate financial products based on the user profile data to obtain the product screening result, which refers to the target financial product selected from the M candidate financial products, where M is a positive integer. The sending unit 42 sends the target financial product associated with the product screening result to the client. This solves the technical problem of low screening accuracy when screening financial products in related technologies. By inputting the candidate financial products determined based on the financial text information and user profile data into the product screening model and processing them to obtain the product screening result, the technical effect of improving the screening accuracy of financial products is achieved.

[0080] Optionally, in the financial product determination device provided in the embodiments of this application, the acquisition unit 40 includes: a first acquisition module, used to acquire financial behavior information of the target user in a preset time period, and to standardize the financial behavior information to obtain financial behavior data; a second acquisition module, used to acquire user information of the target user, and to convert the user information into a format to obtain user data; and a combination module, used to combine the financial behavior data and the user data to obtain user profile data.

[0081] Optionally, in the financial product determination device provided in this application embodiment, the determination unit 41 includes: an integration module, used to integrate M candidate financial products and user profile data by the feature processing layer of the product screening model to obtain M feature sequences; a first transformation module, used to perform nonlinear transformation on the M feature sequences by the fully connected layer of the product screening model to obtain M decision feature vectors; and a processing module, used to process the M decision feature vectors by the output layer of the product screening model to output the product screening result.

[0082] Optionally, in the financial product determination device provided in this application embodiment, the determination unit 41 includes: a conversion module, used to perform vector conversion on user data in user profile data by the input layer to obtain user feature vectors, perform vector conversion on financial behavior data in user profile data to obtain behavior feature vectors, and perform vector conversion on each candidate financial product to obtain M product feature vectors; a first aggregation module, used to aggregate user feature vectors and behavior feature vectors to obtain processed feature vectors; and a third acquisition module, used to acquire sliding window parameters for a candidate financial product, and have the long short-term memory layer combine the processed feature vectors and the product feature vectors corresponding to the candidate financial product according to the sliding window parameters to obtain combined feature vectors, and convert the combined feature vectors to obtain the feature sequence of the candidate financial product.

[0083] Optionally, in the financial product determination device provided in the embodiments of this application, the determination unit 41 includes: a second transformation module, used to perform a first nonlinear transformation on a feature sequence by a first fully connected layer to obtain an initial decision feature vector; and a third transformation module, used to perform a second nonlinear transformation on the initial decision feature vector by a second fully connected layer to obtain a decision feature vector of the feature sequence.

[0084] Optionally, in the financial product determination device provided in this application embodiment, the device further includes: a construction unit, used to, before processing the M decision feature vectors by the output layer of the product screening model and outputting the product screening result, construct a value function for a decision feature vector by the dual network layer of the product screening model based on the decision feature vector, to obtain a set of value function values, wherein the set of value function values ​​includes state value function values ​​and advantage function values, the state value function value refers to the financial value data when the target user uses the candidate financial product corresponding to the decision feature vector, and the advantage function value is used to indicate the ratio of the financial value of the target user using the candidate financial product corresponding to the decision feature vector to the target user using other candidate financial products; the device further includes: a second aggregation module, used to aggregate the state value function value and advantage function value of each decision feature vector by the output layer to obtain M product function values; a sorting module, used to sort the M product function values ​​in descending order to obtain a function value sequence; and a determination module, used to determine the product screening result based on the candidate financial product corresponding to the product function value at a preset position in the function value sequence.

[0085] Optionally, in the financial product determination device provided in this application embodiment, the determination unit 41 includes: a fourth acquisition module, used to acquire the historical product screening results of N users within a historical time period, and acquire user profile data of each user to obtain N sets of historical user profile data, where N is a positive integer; a construction module, used to construct a model training set based on the historical financial products corresponding to the N historical product screening results and the N sets of historical user profile data; and a training module, used to train a preset product screening model using the model training set to obtain a trained product screening model, and if the loss function value corresponding to the trained product screening model is less than the preset data, the trained product screening model is determined as the product screening model.

[0086] It should be noted that the acquisition unit 40, determination unit 41, and sending unit 42 mentioned above correspond to steps S201 to S203 in Embodiment 1. The instances and application scenarios implemented by the above units and corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above modules or units can be hardware or software components stored in memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n). The above units can also be part of a device and can run in the computer terminal 10 provided in Embodiment 1.

[0087] Example 3

[0088] Embodiments of this application may provide a computer terminal, which may be any computer terminal device in a group of computer terminals. Optionally, in this embodiment, the aforementioned computer terminal may also be replaced with a mobile terminal or an electronic device, etc.

[0089] Optionally, in this embodiment, the computer terminal may be located in at least one of a plurality of network devices in a computer network.

[0090] In this embodiment, the computer terminal described above can execute the program code for the following steps in the method for determining financial products: With the authorization of the target user, obtain the target user's user profile data and obtain the target user's financial text information, wherein the financial text information refers to the financial product demand text sent by the target user through the client; determine M candidate financial products based on the financial text information; input the M candidate financial products and the user profile data into a product screening model, and process to obtain a product screening result, wherein the product screening model filters the M candidate financial products based on the user profile data to obtain a product screening result, where the product screening result refers to the target financial product selected from the M candidate financial products, and M is a positive integer; send the target financial product associated with the product screening result to the client.

[0091] Optionally, the aforementioned computer terminal may execute program code for the following steps in the method for determining financial products: obtaining financial behavior information of the target user within a preset time period, and standardizing the financial behavior information to obtain financial behavior data; obtaining user information of the target user, converting the user information into a format to obtain user data; and combining the financial behavior data and user data to obtain user profile data.

[0092] Optionally, the aforementioned computer terminal can execute the program code for the following steps in the method for determining financial products: the feature processing layer of the product screening model integrates M candidate financial products and user profile data to obtain M feature sequences; the fully connected layer of the product screening model performs nonlinear transformation on the M feature sequences to obtain M decision feature vectors; and the output layer of the product screening model processes the M decision feature vectors to output the product screening result.

[0093] Optionally, the aforementioned computer terminal can execute the program code for the following steps in the method for determining financial products: The input layer performs vector transformation on user data in the user profile data to obtain user feature vectors; performs vector transformation on financial behavior data in the user profile data to obtain behavioral feature vectors; performs vector transformation on each candidate financial product to obtain M product feature vectors; aggregates the user feature vectors and behavioral feature vectors to obtain processed feature vectors; for a candidate financial product, obtains sliding window parameters; the long short-term memory layer combines the processed feature vectors and the corresponding product feature vectors of the candidate financial product according to the sliding window parameters to obtain a combined feature vector; and transforms the combined feature vectors to obtain the feature sequence of the candidate financial product.

[0094] Optionally, the computer terminal described above can execute program code for the following steps in the method for determining financial products: for a feature sequence, a first nonlinear transformation is performed on the feature sequence by a first fully connected layer to obtain an initial decision feature vector; a second nonlinear transformation is performed on the initial decision feature vector by a second fully connected layer to obtain a decision feature vector of the feature sequence.

[0095] Optionally, the aforementioned computer terminal can execute the program code for the following steps in the method for determining financial products: For a decision feature vector, the dual network layers of the product screening model construct a value function based on the decision feature vector to obtain a set of value function values. This set of value function values ​​includes state value function values ​​and advantage function values. The state value function value refers to the financial value data when the target user uses the candidate financial product corresponding to the decision feature vector. The advantage function value indicates the ratio of the financial value of the target user using the candidate financial product corresponding to the decision feature vector compared to the target user using other candidate financial products. The output layer of the product screening model processes the M decision feature vectors, outputting the product screening result, including: aggregating the state value function value and advantage function value of each decision feature vector by the output layer to obtain M product function values; sorting the M product function values ​​in descending order to obtain a function value sequence; and determining the product screening result based on the candidate financial product corresponding to the product function value at a preset position in the function value sequence.

[0096] Optionally, the aforementioned computer terminal can execute the program code for the following steps in the method for determining financial products: obtaining the historical product screening results of N users within a historical time period, and obtaining user profile data for each user to obtain N sets of historical user profile data, where N is a positive integer; constructing a model training set based on the historical financial products corresponding to the N historical product screening results and the N sets of historical user profile data; training a preset product screening model using the model training set to obtain a trained product screening model; and determining the trained product screening model as the product screening model if the loss function value corresponding to the trained product screening model is less than the preset data.

[0097] Optionally, Figure 5 This is a structural block diagram of an electronic device according to an embodiment of this application. Figure 5 As shown, the electronic device may include: one or more ( Figure 5 (Only one is shown) processor 502, memory 504, memory controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0098] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the financial product determination method and apparatus in this application embodiment. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the aforementioned financial product determination method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0099] The processor can access the information and application programs stored in the memory via the transmission device to execute the steps described above in the method for determining the aforementioned financial product.

[0100] Those skilled in the art will understand that Figure 5 The structure shown is for illustrative purposes only. Electronic devices can also be smartphones, tablets, handheld computers, mobile internet devices (MIDs), PADs, and other terminal devices. Figure 5 This does not limit the structure of the aforementioned electronic device. For example, electronic devices may also include components that are more... Figure 5 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 5 The different configurations shown.

[0101] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0102] Example 4

[0103] Embodiments of this application also provide a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the method for determining financial products provided in Embodiment 1.

[0104] Optionally, in this embodiment, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0105] Optionally, in this embodiment, the storage medium is configured to store program code for performing the following steps: with the authorization of the target user, obtaining the target user's user profile data and obtaining the target user's financial text information, wherein the financial text information refers to the financial product demand text sent by the target user through the client; determining M candidate financial products based on the financial text information, inputting the M candidate financial products and the user profile data into a product screening model, and processing to obtain a product screening result, wherein the product screening model filters the M candidate financial products based on the user profile data to obtain a product screening result, wherein the product screening result refers to the target financial product selected from the M candidate financial products, where M is a positive integer; and sending the target financial product associated with the product screening result to the client.

[0106] This application also provides a computer program product that, when executed on a data processing device, is suitable for performing the steps of a method for determining a financial product.

[0107] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0108] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0109] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0110] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0111] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0112] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0113] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for determining a financial product, characterized in that, include: With the authorization of the target user, the user profile data of the target user is obtained, and the financial text information of the target user is obtained, wherein the financial text information refers to the financial product demand text sent by the target user through the client. Based on the financial text information, M candidate financial products are determined. The M candidate financial products and the user profile data are input into a product screening model and processed to obtain a product screening result. The product screening model filters the M candidate financial products based on the user profile data to obtain the product screening result. The product screening result refers to the target financial product selected from the M candidate financial products, where M is a positive integer. The target financial products associated with the product screening results are sent to the client.

2. The method according to claim 1, characterized in that, Obtaining the user profile data of the target user includes: The financial behavior information of the target user within a preset time period is obtained, and the financial behavior information is standardized to obtain financial behavior data. Obtain the user information of the target user, and convert the user information into a format to obtain user data; The financial behavior data and the user data are combined to obtain the user profile data.

3. The method according to claim 1, characterized in that, The M candidate financial products and the user profile data are input into the product screening model, and the resulting product screening results include: The feature processing layer of the product screening model integrates the M candidate financial products and the user profile data to obtain M feature sequences; The fully connected layer of the product screening model performs a nonlinear transformation on the M feature sequences to obtain M decision feature vectors; The output layer of the product screening model processes the M decision feature vectors and outputs the product screening result.

4. The method according to claim 3, characterized in that, The feature processing layer includes an input layer and a long short-term memory layer. The feature processing layer of the product selection model integrates the M candidate financial products and the user profile data to obtain M feature sequences, including: The input layer performs vector transformation on the user data in the user profile data to obtain user feature vectors, performs vector transformation on the financial behavior data in the user profile data to obtain behavior feature vectors, and performs vector transformation on each candidate financial product to obtain M product feature vectors. The user feature vector and the behavior feature vector are aggregated to obtain the processed feature vector; For a candidate financial product, the sliding window parameters are obtained. The long short-term memory layer combines the processed feature vector and the product feature vector corresponding to the candidate financial product according to the sliding window parameters to obtain the combined feature vector. The combined feature vector is then transformed to obtain the feature sequence of the candidate financial product.

5. The method according to claim 3, characterized in that, The fully connected layer includes a first fully connected layer and a second fully connected layer. The fully connected layer of the product screening model performs a nonlinear transformation on the M feature sequences to obtain M decision feature vectors, including: For a feature sequence, the first fully connected layer performs a first nonlinear transformation on the feature sequence to obtain an initial decision feature vector; The second fully connected layer performs a second nonlinear transformation on the initial decision feature vector to obtain the decision feature vector of the feature sequence.

6. The method according to claim 3, characterized in that, Before processing the M decision feature vectors by the output layer of the product screening model and outputting the product screening result, the method further includes: For a decision feature vector, the dual network layers of the product screening model construct a value function based on the decision feature vector to obtain a set of value function values. The set of value function values ​​includes state value function values ​​and advantage function values. The state value function value refers to the financial value data when the target user uses the candidate financial product corresponding to the decision feature vector. The advantage function value is used to indicate the ratio of the financial value of the target user using the candidate financial product corresponding to the decision feature vector to the financial value of the target user using other candidate financial products. The product selection model's output layer processes the M decision feature vectors to output the product selection results, including: The output layer aggregates the state value function value and advantage function value of each decision feature vector to obtain M product function values; Sort the M product function values ​​in descending order to obtain a sequence of function values; The product screening result is determined based on the candidate financial products corresponding to the product function values ​​at preset positions in the function value sequence.

7. The method according to claim 1, characterized in that, The product selection model was trained in the following way: Obtain the historical product filtering results of N users within a historical time period, and obtain the user profile data of each user to obtain N sets of historical user profile data, where N is a positive integer; The model training set is constructed based on the historical financial products corresponding to the N historical product screening results and the N sets of historical user profile data. The preset product selection model is trained using the model training set to obtain the trained product selection model. If the loss function value corresponding to the trained product selection model is less than the preset data, the trained product selection model is determined as the product selection model.

8. A device for determining a financial product, characterized in that, include: The acquisition unit is used to acquire user profile data of the target user and financial text information of the target user when authorized by the target user, wherein the financial text information refers to the financial product demand text sent by the target user through the client. The determining unit is configured to determine M candidate financial products based on the financial text information, input the M candidate financial products and the user profile data into a product screening model, and process them to obtain a product screening result. The product screening model filters the M candidate financial products based on the user profile data to obtain the product screening result. The product screening result refers to the target financial product selected from the M candidate financial products, where M is a positive integer. The sending unit is used to send the target financial products associated with the product screening results to the client.

9. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method for determining a financial product according to any one of claims 1 to 7.

10. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the method for determining the financial product according to any one of claims 1 to 7.

Images