Financial product recommendation method and device, storage medium and computer program product

Abstract

The invention discloses a financial product recommendation method and device, a storage medium and a computer program product. The method comprises the steps of obtaining user information of a target user; obtaining information of financial products to be purchased; determining recommended financial management product information based on the user information and the to-be-purchased financial management product information; and based on the recommended financial product information, a financial product recommendation list is generated, and the financial product recommendation list is used for displaying the recommended financial products. Through the method and the device, the problem of low financing product recommendation accuracy in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and more specifically, to a financial product recommendation method, device, storage medium and computer program product. Background Art

[0002] In recent years, with the rapid development of my country's online financial market and the continuous improvement of China's financial industry reform, capital efficiency and transaction structure are changing and adjusting. Informatization of financial institutions has led to the explosive development of Internet financial management. In order to continue to deepen the achievements of smart banking and consolidate our bank's dominant position in the field of financial management, we must realize the interconnected online and offline financial management ecosystem, increase the number of users and transaction volume through a large number of high-quality financial products, and effectively promote the development of bank financial products.

[0003] At present, there are thousands of financial products on the Internet, and various financial platforms emerge in an endless stream. Faced with such a rich variety and amount of financial products, users may buy one or several of their favorite financial products on the Internet. However, due to the problems of too many types of financial products, uncertain investment returns, and inability to meet users' needs for financial products in a timely manner, users may not always be able to choose their preferred financial products. The financial product recommendation system can reasonably recommend financial products that users may like. The developers of financial product platforms have gradually noticed the psychological needs of users, and most platforms have added financial product recommendation functions.

[0004] As people's financial awareness gradually improves, it is necessary to fully combine the Internet platform and algorithm technology to complete the work of traditional financial advisors, so as to more accurately judge the user's risk tolerance level and complete asset allocation and other tasks, thereby improving user satisfaction with the financial interaction process.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] The main purpose of this application is to provide a financial product recommendation method, device, storage medium and computer program product to solve the problem of low accuracy of financial product recommendation in related technologies.

[0007] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a financial product recommendation method is provided. The method comprises: obtaining user information of a target user; obtaining information of financial products to be purchased; determining recommended financial product information based on the user information and the information of financial products to be purchased; generating a financial product recommendation list based on the recommended financial product information, wherein the financial product recommendation list is used to display recommended financial products.

[0008] Optionally, generating a wealth management product recommendation list based on the recommended wealth management product information includes: obtaining wealth management product scores in the recommended wealth management product information; and sorting the wealth management products based on the wealth management product scores to generate the wealth management product recommendation list.

[0009] Optionally, based on the user information, determining recommended financial product information includes: when it is determined that the browsing interest information in the user information is not browsing financial products, performing cluster analysis on the user information to obtain user classification information; based on the user classification information, determining first recommended financial product information in the recommended financial product information, wherein the first recommended financial product information is the recommended financial product information obtained according to the cluster analysis.

[0010] Optionally, based on the user information, determining recommended financial product information includes: when it is determined that the browsing interest information in the user information is browsing financial products, and / or the financial transaction information in the user information is a purchased financial product, based on the target neural network prediction model, analyzing the user information and the information of the financial product to be purchased to obtain second recommended financial product information in the financial product recommendation information, wherein the second recommended financial product information is the recommended financial product information obtained based on the target neural network prediction model.

[0011] Optionally, before analyzing the user information based on the target neural network prediction model, the process includes: constructing an initial target neural network model; obtaining the target user's preferred financial product information, which is the information about financial products that the target user has subscribed to, collected and purchased; inputting the user information and the rate of return and risk rate in the preferred financial product information based on the initial target neural network model to obtain a predicted score for the financial product; and training the initial target neural network model based on the preferred financial product score and the financial product prediction score in the preferred financial product information to obtain a target neural network prediction model.

[0012] Optionally, constructing the target neural network initial model includes: performing structured processing on user information to obtain a user information vector; constructing a convolution layer based on the user information vector; and constructing the target neural network initial model based on the convolution layer.

[0013] Optionally, based on the user information vector, a convolution layer is constructed, including: constructing a first convolution layer based on a first user feature information vector in the user information vector, wherein the first user feature information vector includes an age feature vector and a city feature vector; constructing a second convolution layer based on a second user feature information vector in the user information vector, wherein the second user feature information vector includes an investment ability information vector; constructing a third convolution layer based on a third user feature information vector and a financial management knowledge information vector in the user information vector; wherein the third user feature information vector includes a training degree vector, an experience value information vector and a financial management score information vector.

[0014] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a financial product recommendation device is provided. The device includes: a first acquisition module, the first acquisition module is used to acquire user information of a target user; a second acquisition module, the second acquisition module is used to acquire information of financial products to be purchased; a determination module, the determination module is used to determine recommended financial product information based on user information and information of financial products to be purchased; a generation module, the determination module is used to generate a financial product recommendation list based on the recommended financial product information, and the financial product recommendation list is used to display recommended financial products.

[0015] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a computer-readable storage medium is provided, and the computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the financial product recommendation method in the embodiment of the present application.

[0016] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a computer program product is provided, including computer instructions, which implement the steps of the method in the embodiment of the present application when the computer instructions are executed by a processor.

[0017] In the embodiment of the present application, the recommended financial product information is determined through user information and information of the financial product to be purchased, and the complex investment preferences of the user and the multi-dimensional characteristics of the financial product are captured and understood, thereby realizing a more intelligent recommendation service, thereby solving the problem of low accuracy of financial product recommendation in the related art. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0019] Figure 1 A hardware structure block diagram of a computer terminal for implementing a financial product recommendation method is shown;

[0020] Figure 2 is a flow chart of a financial product recommendation method provided according to an embodiment of the present application;

[0021] Figure 3 is a flow chart of a financial product recommendation method provided according to an embodiment of the present application;

[0022] Figure 4 is a schematic diagram of a financial product recommendation device provided according to an embodiment of the present application;

[0023] Figure 5It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0025] It should be noted that 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 are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following explanations:

[0027] 1) Construct a dilated convolutional neural network model:

[0028] Convolutional neural networks are composed of input layers, convolutional layers, pooling layers, and fully connected layers. The input layer uses a descent gradient algorithm to standardize the features of its input, and performs normalization, dimensionality reduction, and other operations. The convolutional layer is the core of the neural network, which extracts features from the input data, reduces useless features, and enhances effective features. The pooling layer performs feature selection and information filtering, reduces the total number of parameters, and retains important information. The fully connected layer can merge features into the same one and output a value, which can reduce the impact of feature position changes on the segmentation results.

[0029] 2) Similarity calculation

[0030] The correlation coefficient is a measure of the degree of linear correlation between research variables. Definition:

[0031]

[0032] Among them, Cov(x,y) is the covariance of x and y, var[x] is the variance of x, and var[y] is the variance of y. Obtaining the correlation coefficient of each sub-category of funds is one of the ways to calculate the similarity between funds. By calculating the correlation between fund sub-categories, it can be applied to the content recommendation model. By describing the funds in a labeled way, the similarity of various funds is associated in terms of risk, return, volatility and other indicators.

[0033] 3) Activation function

[0034] The significance of the activation function is to determine whether the output of the previous neuron has reached the threshold. In neural networks, commonly used activation functions include sigmoid function, Relu function, and tanh function.

[0035]

[0036] Among them, the Sigmoid function can map an input real number to the value space of (0,1). It has a disadvantage that if the input value is too large or too small, the function value will infinitely asymptotically approach the two extremes, causing the gradient of the training results to disappear, which will ultimately affect the accuracy.

[0037] It should be noted that the collected information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this application are 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 relevant data are in compliance with relevant laws, regulations and standards, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation entrances are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions to provide users with corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0038] Example 1

[0039] According to an embodiment of the present application, an embodiment of a financial product recommendation method is also provided. It should be noted that the steps shown in the flowchart of 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 can be executed in an order different from that shown here.

[0040] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1FIG. 1 shows a hardware structure block diagram of a computer terminal (or mobile device) for implementing a financial product recommendation method. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0041] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0042] 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 recommendation method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the above-mentioned financial product recommendation method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0043] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0044] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0045] Under the above operating environment, this application provides Figure 2 The financial product recommendation method shown. Figure 2 It is a flow chart of the financial product recommendation method according to Example 1 of the present application.

[0046] Step S11, obtaining user information of the target user.

[0047] In step S11, the user information includes at least one of the following: risk assessment information, user old and new information, user characteristic information, financial knowledge information, browsing interest information and financial transaction information. Among them, the risk assessment information includes the user's risk preference level, which is usually evaluated by questionnaire surveys, historical investment behavior analysis, etc. The old and new information of the user is used to distinguish whether the user is a new user or an old user. New users may need a more basic introduction to financial products, while old users may prefer more complex products. User characteristic information covers the user's personal attributes, such as age, gender, occupation, city, investment ability information, etc., which are used to judge their investment ability and preferences. Financial knowledge information is the user's financial knowledge level, which can be judged by the user's behavior on the platform, the articles read, etc. The financial knowledge information includes training information, experience value information and financial score information, etc., wherein the training information represents the user's financial knowledge learning on the platform, including but not limited to the number of times the user has participated in financial knowledge training, the amount of courses completed, the time of online learning, etc. The training information reflects whether the user is willing to spend time learning financial knowledge and his financial education level. Experience value information is the user's actual financial management experience, including his investment history, transaction frequency, types and quantity of financial products held, etc. The experience value reflects whether the user is an active user with frequent transactions, and his familiarity with different types of financial products. Financial management score information is an indicator for comprehensively evaluating the user's financial management ability, which can be a comprehensive score based on the user's past investment performance, risk control ability and financial management knowledge level. The higher the financial management score, the stronger the user's comprehensive ability in financial management. Browsing interest information is the user's browsing history for specific financial products, which is used to capture their potential interests. Financial management transaction information is the user's past purchase and transaction records of financial products, which is an important basis for the recommendation system.

[0048] Step S12, obtaining information on financial products to be purchased.

[0049] In step S12, the information of the financial product to be purchased includes at least one of the following: product type information, product risk level information, historical performance information, fee structure information, transaction rule information and product score information. Product type information is the basic classification of financial products, including stock type, bond type, hybrid type, money market fund, index fund, QDII fund, etc. This information helps the system understand the investment field and risk attributes of the product, so as to match it with the risk preference and investment goals of the user. Product risk level information reflects the potential volatility and loss possibility of financial products, and is usually divided into low risk, medium risk and high risk levels. This helps the system recommend suitable products for users with different risk tolerance. Historical performance information includes the yield, volatility, etc. of financial products in the past period of time. By analyzing historical data, the system can evaluate the long-term performance and stability of the product and provide a basis for user investment decisions. The fee structure information covers cost information such as management fees, custody fees, sales service fees, and redemption fees. The system will consider the impact of these fees on the user's investment return to ensure that the recommended products are cost-effective. Trading rules include the product's purchase threshold, redemption conditions, trading time limits, etc. These rules directly affect the user's investment experience and operational convenience, and are important factors that need to be considered when the system makes recommendations. Product score information is an indicator for comprehensively evaluating financial products. It can be calculated by the system algorithm based on multi-dimensional information such as product type, risk level, historical performance, and fee structure. The higher the product score, the more outstanding its performance in one or more aspects. The system will give priority to recommending financial products with higher scores.

[0050] Step S13, determining recommended financial product information based on the user information and the information of the financial product to be purchased;

[0051] In step S13, the user information and financial product information are deeply analyzed using the convolutional neural network model to determine which financial products are most suitable for the user. The specific operations include: converting the user information and financial product information into vector or matrix form for input into the target neural network model; using the convolutional neural network model to extract and process features to match user preferences with financial product attributes; calculating the matching degree or similarity between financial products and user features through the output of the target neural network model; sorting the financial products according to the matching degree or similarity, and selecting the most suitable financial product as the recommended object.

[0052] Step S14: Generate a recommended financial product list based on the recommended financial product information, where the recommended financial product list is used to display recommended financial products.

[0053] In step S14, a wealth management product recommendation list is generated based on the product recommendation information determined in step S13. The specific process is as follows: the wealth management product information sorted in step S13 is integrated into a list; on the user interface of the wealth management platform, the recommendation list is displayed according to the user's personalized information (such as interface preference, display format, etc.); the recommendation list should be updated in real time according to the user's latest behavior and market dynamics to maintain the accuracy and timeliness of the recommendation.

[0054] Through the above steps, the recommended financial product information is determined based on user information and information on the financial products to be purchased, and the user's complex investment preferences and the multi-dimensional characteristics of financial products are captured and understood. At the same time, the user's complex investment preferences and the multi-dimensional characteristics of financial products can be captured and understood, thereby achieving more intelligent recommendation services, thereby solving the problem of low accuracy of financial product recommendations in related technologies, and achieving the effect of improving user satisfaction and investment decision-making efficiency.

[0055] Optionally, in the financial product recommendation method provided in the embodiment of the present application, generating a financial product recommendation list based on the recommended financial product information includes:

[0056] Step S201: Obtain the financial product score in the recommended financial product information.

[0057] In step S201, score information of each financial product is extracted from the recommended financial product information. The product score is usually a comprehensive indicator that reflects the performance of the financial product in multiple dimensions, such as risk, return, liquidity, fees, etc. Obtaining product score information may involve the following operations:

[0058] Query the score information of each recommended financial product from the database or data storage system. This may require accessing the database table that stores product feature information and product scores, and obtaining data through query statements or API calls; sorting the obtained product score information to ensure that the data format is consistent and there are no errors or missing values. This may include data preprocessing steps such as data type conversion, missing value filling, and outlier processing; extract the sorted product score information from the original data set to form a separate list or data set for subsequent sorting operations.

[0059] Step S202: Sorting the financial products based on their scores to generate a recommended list of financial products.

[0060] In step S202, after obtaining the score information of the financial products, the scores are sorted to generate a final recommendation list of financial products. The sorting process can be implemented by a variety of algorithms, but is usually based on the following steps: select a suitable sorting algorithm, such as quick sort, merge sort or heap sort, etc., with efficiency and stability as considerations; since financial product recommendations usually tend to recommend products with higher scores, the system will sort them in descending order from high to low according to the scores; after the sorting is completed, the system generates a list containing key information such as the name of the financial product, product type, and product score, i.e., the financial product recommendation list; during the sorting process, the system can also fine-tune the recommendation list according to the user's personalized preferences (such as risk preference, investment period preference, etc.) to provide recommendation results that better meet the user's needs; finally, the financial product recommendation list is displayed to the user, usually in the user interface of the financial platform, in the form of a list, card or grid, providing sufficient product information for the user to make a decision.

[0061] Through the above steps, the system can refine product scores from the recommended financial product information, and based on this, generate a recommended list of financial products sorted by scores, helping users quickly find financial products that best meet their needs and expectations, thereby improving user satisfaction and platform transaction activity.

[0062] Optionally, in the financial product recommendation method provided in the embodiment of the present application, determining recommended financial product information based on user information includes:

[0063] Step S203: When it is determined that the browsing interest information in the user information is not browsing financial products, cluster analysis is performed on the user information to obtain user classification information.

[0064] In step S203, when the system finds that there is no recent browsing behavior of financial products in the user's browsing interest information, it generates first recommended financial product information based on cluster analysis of user information to stimulate user interest and provide financial products that may meet their needs.

[0065] First, cluster analysis is performed on user information to identify user groups with similar characteristics and lay the foundation for personalized recommendations. The cluster analysis process includes: data preprocessing, feature selection, clustering algorithm selection and clustering result evaluation. Among them, data preprocessing includes standardizing and encoding user information to ensure that all features are on the same scale to facilitate subsequent clustering algorithm processing. Feature selection includes selecting features from user information that are highly related to financial management preferences, such as user age, gender, occupation, investment ability, past investment behavior, etc. Clustering algorithm selection includes using appropriate clustering algorithms, such as K-means, DBSCAN, hierarchical clustering, etc., to divide users into multiple groups with similar characteristics according to their characteristics. Clustering result evaluation includes evaluating the clustering effect through indicators such as silhouette coefficient and Calinski-Harabasz index (variance ratio criterion) to ensure the similarity of user characteristics within each group and the differences between groups.

[0066] Step S204: determining first recommended financial product information in the recommended financial product information based on the user classification information, wherein the first recommended financial product information is the recommended financial product information obtained according to cluster analysis.

[0067] In step S204, when the cluster analysis is completed, the system will determine the first recommended financial product information for each user group based on the user classification information. This process takes into account the common needs and preferences of users within the group, and the specific steps are as follows: Group financial needs analysis: Analyze the commonalities of each user group and identify their demand points for financial products, such as risk preference, return expectations, investment period, etc. Financial product optimization: Filter out financial products that meet the needs of a specific user group from the financial product pool. These products may match the financial preferences of the group in terms of risk, return, liquidity, etc. Generate first recommended financial product information: Generate a list of first recommended financial product information for each user group, including key information such as product name, type, risk level, expected rate of return, etc. Personalized recommendation display: Finally, the system will display the first recommended financial product information to the user based on the group to which the user belongs, in order to re-attract the user's attention to financial products and meet their potential financial needs.

[0068] Through the above steps, even when the user does not directly express interest in browsing financial products, the system can capture the user's potential needs through cluster analysis and provide corresponding financial product recommendations, thereby improving user satisfaction and the platform's interaction rate.

[0069] Optionally, in the financial product recommendation method provided in the embodiment of the present application, determining recommended financial product information based on user information includes:

[0070] Step S205: When it is determined that the browsing interest information in the user information is browsing financial products, and / or the financial transaction information in the user information is purchased financial products, based on the target neural network prediction model, the user information and the information of the financial products to be purchased are analyzed to obtain the second recommended financial product information in the financial product recommendation information, wherein the second recommended financial product information is the recommended financial product information obtained based on the target neural network prediction model.

[0071] In step S205, the user's browsing interest information and financial transaction information are checked to determine whether the user has a clear interest in and investment experience in financial products. If the user has recently browsed financial products or has a record of purchasing financial products, this indicates that the user has a continuous or recent interest in financial management, and the next step is entered. After confirming that the user has browsing or purchasing behavior, the previously trained target neural network prediction model is used to deeply analyze the user information and the information of the financial products to be purchased, thereby generating the second recommended financial product information. The specific steps are as follows: collect information on financial products that users have recently browsed, including browsing frequency, browsing time, browsed product types, etc.; at the same time, obtain information on financial products that users have purchased and their related transaction data; perform feature extraction and engineering processing on the collected data, for example, convert user browsing and transaction behaviors into numerical features, and add user basic information (such as age, gender, occupation, etc.) and financial knowledge level, etc., to form user and product feature vectors; input the processed user feature vector and the feature vector of the financial product to be purchased into the target neural network prediction model, which can effectively learn and extract the complex relationship between users and financial products through multi-layer convolution, pooling and full connection operations; the model will output the predicted score or recommendation index for each financial product to be purchased for the user, based on which the system will generate the second recommended financial product information, usually a list of financial products that are most attractive to users or best meet their investment needs.

[0072] Based on step S205, through the accurate prediction of the deep learning model, the second recommended financial product information can more closely fit the user's specific needs and preferences, thereby improving the personalization and effectiveness of financial product recommendations and enhancing the user's satisfaction and trust in the platform.

[0073] Optionally, in the financial product recommendation method provided in the embodiment of the present application, before analyzing the user information based on the target neural network prediction model, the method includes:

[0074] Step S206: constructing an initial model of the target neural network.

[0075] In step S206, a neural network model (target neural network initial model) is first created as a starting point. This model can be a convolutional neural network (CNN) or other deep learning architecture suitable for processing time series and user preferences. Model construction includes defining the network structure (such as the number of layers, the number of neurons, activation functions, etc.), initializing parameters (such as weights and biases), and selecting loss functions and optimizers.

[0076] Step S207: Obtain the target user's preferred financial product information, where the preferred financial product information is the financial product information that the target user has subscribed to, collected, and purchased.

[0077] In step S207, the target user's specific financial product preference information is collected, including financial products that the user has subscribed to, collected or purchased. This information can reflect the user's historical investment behavior and potential financial preferences, and is the cornerstone of personalized recommendations.

[0078] Step S208: Based on the target neural network initial model, input the user information, the rate of return and the risk rate in the preferred financial product information to obtain the financial product prediction score.

[0079] In step S208, during the model training process, the system inputs the user information and key features such as the rate of return and risk rate of the preferred financial products into the model, and calculates the model's possible predicted scores or recommendation indexes for these financial products through forward propagation. These predicted scores are the model's prediction of the attractiveness of the financial products based on the current parameter settings.

[0080] Step S209: Based on the preferred financial product scores and the financial product prediction scores in the preferred financial product information, the target neural network initial model is trained to obtain a target neural network prediction model.

[0081] In step S209, the step of training the model is the core of the iterative training process, which specifically includes: calculating losses, back propagation, parameter updating, and model evaluation and selection. The system compares the historical scores of the preferred financial products with the predicted scores of the model, and calculates the losses (such as mean square error, cross entropy, etc.), which reflects the accuracy of the model prediction. Based on the calculated losses, the back propagation algorithm is used to calculate the gradient, that is, the degree of influence of the model parameters on the losses. The parameters of the model are updated using the gradient and a pre-set learning rate to minimize the loss function and improve the prediction accuracy of the model. The performance of the model is regularly evaluated during the training process, and a validation set may be used to avoid overfitting. The model version that performs best on the validation set is selected as the final training model, that is, the target neural network prediction model.

[0082] Based on steps S206 to S209, the neural network model can be continuously adjusted and trained according to the historical preference data of the target user, and finally a target neural network prediction model that can more accurately predict and recommend financial products is obtained. This not only improves the personalization level of the recommendation, but also enhances the generalization ability of the model, so that it can better adapt to future user behavior and market changes.

[0083] Optionally, in the financial product recommendation method provided in the embodiment of the present application, constructing a target neural network initial model includes:

[0084] Step S210: Structural processing is performed on the user information to obtain a user information vector.

[0085] In step S210, before starting to build the model, the user's unstructured information (such as age, gender, occupation, historical investment records, etc.) needs to be converted into a structured numerical vector. This process usually involves steps such as data cleaning, encoding, quantization and normalization to ensure that the format and dimension of the data are friendly to model training. It should be noted that data cleaning includes removing or filling missing values ​​and processing abnormal data. Encoding includes converting categorical data (such as gender, occupation) into numerical form, possibly using one-hot encoding. Quantization includes converting non-numerical data (such as investment preference descriptions) into quantifiable numerical indicators. Normalization includes scaling numerical data so that it falls into the same numerical range, such as between 0 and 1, to avoid the numerical range of certain features being too large and dominating the model learning process.

[0086] Step S211: construct a convolutional layer based on the user information vector.

[0087] In step S211, the system designs the parameters of the convolution layer according to the structured user information vector, such as the convolution kernel size, step size, padding method, etc. The convolution layer can capture the patterns and regularities in the information, laying the foundation for subsequent more complex feature combination and learning.

[0088] Step S212: Based on the convolutional layer, construct an initial model of the target neural network.

[0089] In step S212, a fully connected layer, a pooling layer, a normalization layer, etc. are added after the convolution layer to complete the entire neural network architecture. The fully connected layer is used for the final decision or scoring output, the pooling layer is used to reduce the feature dimension, and the normalization layer is used to train the gradient update. All parameters in the network (including the weights and biases of the convolution layer) are initialized. Common initialization methods include random initialization or specific initialization strategies. Select a suitable loss function, such as mean square error, cross entropy loss, etc., to evaluate the gap between the model prediction value and the actual value, and guide the training direction of the model training. Set the optimizer and learning rate to adjust the parameters during the model training process to minimize the loss function.

[0090] Integrate the model: Integrate all built layers and defined loss functions and optimizers into a complete model, ready for training.

[0091] Based on step S211 to step S212, starting from the user information vector, an initial neural network model that can automatically learn and extract features from the data is constructed. This model will be used for subsequent training and optimization to provide more accurate financial product recommendations. Optionally, in the financial product recommendation method provided in the embodiment of the present application, a convolutional layer is constructed based on the user information vector, including:

[0092] Step S213: Based on the first user feature information vector in the user information vector, the first convolution layer is constructed, wherein the first user feature information vector includes an age feature vector and a city feature vector; in step S213, the feature information of the user is focused on, including but not limited to basic information such as age, gender, city, and possible risk preferences, investment styles, etc. These feature information are encoded into vector form and input to the first convolution layer of CNN. The convolution layer slides on these vectors through the convolution kernel (i.e., filter) to perform feature extraction and local correlation detection, providing a preliminary feature representation for subsequent analysis.

[0093] Step S214: constructing a second convolutional layer based on a second user feature information vector in the user information vector, wherein the second user feature information vector includes an investment capability information vector;

[0094] In step S214, the investment ability information in the second user feature information vector reflects his economic ability and has a direct impact on the choice of financial products. These feature information are encoded into vector form, and the investment ability information vector is also processed by the second convolutional layer, which may focus more on the periodicity and stability of investment and the correlation between investment ability and financial management goals. The convolutional layer can identify key features in the user's investment pattern, such as seasonal changes in investment, long-term growth trends, etc., so as to better understand the user's financial situation and investment ability.

[0095] Step S215: constructing a third convolutional layer based on the third user feature information vector and the financial management knowledge information vector in the user information vector; wherein the third user feature information vector includes a training degree vector, an experience value information vector and a financial management score information vector.

[0096] In step S215, the third user feature information vector is a financial knowledge information vector, which covers the user's understanding of financial products, investment strategy proficiency, etc., and is an important basis for personalized recommendations. The third user feature information includes but is not limited to training information, experience value information, and financial score information. These feature information are encoded into vector form. Through the third convolutional layer, the system can identify the characteristics of the user's investment decision-making model, such as preference for low-risk products, good at long-term investment, and in-depth understanding of specific industries or markets. This helps the model more accurately predict the user's demand and response to different types of financial products.

[0097] Based on step S213 to step S214, in the process of constructing these three convolutional layers, the system will perform special feature extraction and processing for the first user feature information vector, the second user feature information vector, and the third user feature information vector. Each convolutional layer has its own specific filter size, step size, padding method and other parameters, and the selection of these parameters is based on the understanding of the potential structure and pattern in the user information vector. Through multi-layer convolution operations, the model can comprehensively analyze user information from different dimensions and details, thereby constructing a more comprehensive and personalized user portrait, providing strong data support for subsequent financial product recommendations.

[0098] Furthermore, after completing the construction of the three convolutional layers, the feature dimension is further reduced through the pooling layer, and all features are integrated through the fully connected layer to make the final prediction or recommendation. This series of operations work together to ensure that the recommendation results can fully reflect the characteristics and needs of individual users, and improve the satisfaction and conversion rate of recommendations.

[0099] Optionally, combined Figure 3 As shown, in one embodiment of the present application, the financial product recommendation method includes the following steps:

[0100] Step (1) determines whether the user has done risk assessment. If not, prompt the user to do risk assessment, otherwise jump to step (2).

[0101] Step (2) determines whether the user is a new user. If the user is a new user, an initialized financial recommendation list is recommended based on user collaborative filtering, and then jumps to step (3).

[0102] Step (3) determines whether the customer clicks on favorites, subscribes to topics, or browses detailed information of financial products in the logged-in session. If the customer does not click on financial products, the financial recommendation list obtained in steps (1), (2), and (3) (the financial recommendation list generated based on the first recommended financial product information) is the final financial recommendation list. Otherwise, the process proceeds to step (4).

[0103] Step (4), based on the user's financial transaction behavior sequence data in the past period of time, through the void model based on the convolutional neural network (target neural network prediction model), finally output the financial product recommendation list (the financial product recommendation list generated based on the second recommended financial product information).

[0104] Step (5): Display the list of financial products on the home page of the financial management platform in a personalized recommendation manner.

[0105] After making personalized recommendations, the system will track the user's feedback results, and measure the recommendation results based on whether the user clicks on the recommended financial products and whether the user purchases the recommended financial products. And based on the user's feedback results, the user's preference information and the weights of each algorithm are updated. It is a safe solution to consider stable and commonly used models when building a personalized recommendation system in the early stage. In the later stage, you can learn from the experience in the recommendation process and continuously iterate and develop new recommendation algorithms to meet the new needs of customers.

[0106] Furthermore, the convolution kernel consists of the first convolution layer, the second convolution layer and the third convolution layer, and the interval number of the convolution kernel is 2, which is the dilation rate of the void.

[0107] First convolutional layer: B = ω 1 sex+ω 2 age+ω 3 city;

[0108] Second convolutional layer: S = σ 1 basic+σ 2 bonus+σ 3 allowance;

[0109] The third convolutional layer: F = β 1 train+β 2 experience+β 3 scope;

[0110] Among them, ω is the weight vector of the first user feature information, sex is the age feature vector, age is the gender feature vector, city is the city feature vector, σ is the weight vector of the second user feature information, basic is the first investment ability feature vector, bonus is the second investment ability feature vector, allwoance is the third investment ability feature vector, β is the third user feature information weight vector, trainwei is the training degree information feature vector, experience is the experience value information feature vector, and scopewei is the financial management score information feature vector. Among them, the first investment ability feature vector, the second investment ability feature vector, and the third investment ability feature vector are feature vectors determined based on different types of investment capabilities.

[0111] Building a model involves data migration. Assume that the total data set is randomly divided into training set, test set and validation set, training set: test set: validation set = 40%: 50%: 10%.

[0112] Take the transaction data of the fund (financial product to be purchased) for one week, divide the closing price of each day of the fund by the closing price of the previous day and multiply by 100 to get the rate of return x; use the opening price, lowest price, highest price, and closing price to calculate the average, and then calculate the standard deviation of the fund as the risk rate y; the user information vector is used as z; then use x, y, z to construct a spatial vector (x, y, z), and input it into the hollow convolutional network structure through the embedding layer. The convolution layer performs convolution operation on the input data, and the number of intervals of the convolution kernel is used to expand the perception field of view and capture more global trading behavior characteristics. The feature calculation is as follows:

[0113] T i =f(w*vet i + b);

[0114] Among them, w is the weight ratio, and b is the bias term in the convolution operation.

[0115] The neuron in row i and column j of feature map k in the second convolutional layer is connected to the neuron in convolutional layer 1 located at i×S w +f w -1 row, j×S h +f k -1 column of neuron outputs.

[0116] After the convolution operation, the financial vector data matrix will be smaller than the original input matrix, and the new feature matrix is ​​obtained. In this embodiment, the sigmoid activation function is used to adjust the weight ratio and the bias term. The activation function is as follows:

[0117]

[0118] The pooling layer samples the financial features for the second time, where each neuron is connected to the output of a limited number of neurons in the previous layer to reduce the computational burden, memory usage, and number of parameters. This embodiment uses the maximum pooling method to process the data after the convolution operation, and the pooling result is shown in the following formula:

[0119] P = max(F) = max{f 1 ,f 2 .......f n};

[0120] The input data is input into the convolutional neural network to implement the convolution operation and obtain the predicted score A of the financial product. The predicted score of the financial product is compared with the actual score of the financial product (the score of the preferred financial product).

[0121] After calculating the gradient of each weight using the chain rule of composite function derivation, the model is then trained for parameter tuning and the learning rate parameter tuning model is set:

[0122] w = w - (w * lr * decay);

[0123] Among them, w is the gradient, the left side of the equal sign is the new round of gradient, the right side of the equal sign is the previous round of gradient, Lr is the learning rate, and decay parameter. The decay parameter can be a fixed decimal, such as 1 / 2. Each round of learning rate is multiplied by the decay parameter to achieve the effect of large changes at the beginning and smaller changes later.

[0124] The model training process is divided into the following four steps:

[0125] 1. Forward propagation: the model accepts input and predicts results;

[0126] 2. Calculate Loss: Calculate the loss value between the predicted value and the true value;

[0127] 3. Back propagation: calculate the gradient of each weight;

[0128] 4. Correction parameters: Update the parameters according to the gradient, learning rate, and decay parameters.

[0129] In this embodiment, the return function (root mean square error) RMSE is used to measure the inference accuracy of the model. The root mean square error can also explain the error between the predicted value and the true value. The definition is as follows: the smaller the RMSE value, the higher the score of the recommended fund, and the better the model recommendation effect.

[0130]

[0131] Among them, h(x i ) is the score of the fund predicted by the convolutional neural network, yi Provide users with real ratings of funds.

[0132] Predict the scores of the recommended funds, and finally sort the scores of the recommended financial products output by the scoring system according to the scoring rules; generate a recommendation list based on the scores.

[0133] The financial product recommendation method provided in the embodiment of the present application determines the recommended financial product information through user information and information of the financial products to be purchased, captures and understands the user's complex investment preferences and the multi-dimensional characteristics of financial products, and realizes more intelligent recommendation services, thereby solving the problem of low accuracy of financial product recommendations in related technologies, thereby achieving the effect of improving user satisfaction and investment decision-making efficiency.

[0134] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0135] Example 2

[0136] The embodiment of the present application also provides a financial product recommendation device. It should be noted that the financial product recommendation device of the embodiment of the present application can be used to execute the financial product recommendation method provided in the embodiment of the present application. The financial product recommendation device provided in the embodiment of the present application is introduced below.

[0137] According to an embodiment of the present application, a device for implementing the above-mentioned financial product recommendation method is also provided, such as Figure 4 As shown, the device comprises:

[0138] A first acquisition module 301, the first acquisition module 301 is used to acquire user information of a target user;

[0139] A second acquisition module 302, the second acquisition module 302 is used to acquire information of financial products to be purchased;

[0140] Determination module 303, determination module 303 is used to determine recommended financial product information based on user information and information about the financial product to be purchased;

[0141] The generating module 304 is used to generate a recommended financial product list based on the recommended financial product information, and the recommended financial product list is used to display the recommended financial products.

[0142] The financial product recommendation device provided in the embodiment of the present application determines recommended financial product information through user information and information of the financial products to be purchased, captures and understands the user's complex investment preferences and the multi-dimensional characteristics of financial products, and realizes more intelligent recommendation services, thereby solving the problem of low accuracy of financial product recommendations in related technologies, thereby achieving the effect of improving user satisfaction and investment decision-making efficiency.

[0143] Optionally, in the financial product recommendation device provided in the embodiment of the present application, the second acquisition module 302 is further used to obtain the financial product score in the recommended financial product information; the generation module 304 is further used to sort the financial product based on the score to generate a financial product recommendation list.

[0144] Optionally, in the financial product recommendation device provided in the embodiment of the present application, the generation module 304 is also used to perform cluster analysis on the user information to obtain user classification information when it is determined that the browsing interest information in the user information is not browsing financial products; based on the user classification information, determine the first recommended financial product information in the recommended financial product information, wherein the first recommended financial product information is the recommended financial product information obtained according to the cluster analysis.

[0145] Optionally, in the financial product recommendation device provided in the embodiment of the present application, the generation module 304 is also used to analyze the user information and the financial product information to be purchased based on the target neural network prediction model when it is determined that the browsing interest information in the user information is browsing financial products, and / or the financial transaction information in the user information is a purchased financial product, to obtain second recommended financial product information in the financial product recommendation information, wherein the second recommended financial product information is the recommended financial product information obtained based on the target neural network prediction model.

[0146] Optionally, in the financial product recommendation device provided in the embodiment of the present application, the generation module 304 is also used to construct an initial model of the target neural network; obtain the target user's preferred financial product information, which is the information of financial products subscribed, collected and purchased by the target user; based on the initial model of the target neural network, input the user information, the rate of return and the risk rate in the preferred financial product information to obtain the financial product prediction score; based on the preferred financial product score and the financial product prediction score in the preferred financial product information, train the initial model of the target neural network to obtain the target neural network prediction model.

[0147] Optionally, in the financial product recommendation device provided in the embodiment of the present application, the generation module 304 is also used to perform structured processing on the user information to obtain a user information vector; construct a convolution layer based on the user information vector; input the user information, the rate of return of the preferred financial product and the risk rate of the preferred financial product into the convolution layer to obtain a predicted score for the financial product.

[0148] Optionally, in the financial product recommendation device provided in the embodiment of the present application, the generation module 304 is also used to construct a first convolutional layer and a second convolutional layer based on the user feature information vector in the user information vector; and to construct a third convolutional layer based on the financial knowledge information vector in the user information vector.

[0149] It should be noted that the second acquisition module 302 and the generation module 304 correspond to steps S201 to S202 in Example 1, and the examples and application scenarios implemented by the two modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned Example 1. It should be noted that the above-mentioned modules or units may be hardware components or software components stored in a memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n), and the above-mentioned modules may also be part of the device and may be run in the computer terminal 10 provided in Example 1.

[0150] Example 3

[0151] An embodiment of the present application may provide an electronic device, Figure 5 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 5 As shown, the electronic device may include: one or more ( Figure 5 (only one is shown) processor 1002, memory 1004, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0152] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method. The memory may include a high-speed random access memory, and may also include a 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 a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0153] The processor may call the information and application program stored in the memory through the transmission device to perform the following steps:

[0154] Step S11, obtaining user information of the target user.

[0155] Step S12, obtaining information on financial products to be purchased.

[0156] Step S13, determining recommended financial product information based on the user information and the information of the financial product to be purchased;

[0157] Step S14: Generate a recommended financial product list based on the recommended financial product information, where the recommended financial product list is used to display recommended financial products.

[0158] The processor may also call the information and application program stored in the memory through the transmission device to perform the following steps:

[0159] Step S201: Obtain the financial product score in the recommended financial product information.

[0160] Step S202: Sorting the financial products based on their scores to generate a recommended list of financial products.

[0161] The processor may also call the information and application program stored in the memory through the transmission device to perform the following steps:

[0162] Step S203: When it is determined that the browsing interest information in the user information is not browsing financial products, cluster analysis is performed on the user information to obtain user classification information.

[0163] Step S204: determining first recommended financial product information in the recommended financial product information based on the user classification information, wherein the first recommended financial product information is the recommended financial product information obtained according to cluster analysis.

[0164] The processor may also call the information and application program stored in the memory through the transmission device to perform the following steps:

[0165] Step S205: When it is determined that the browsing interest information in the user information is browsing financial products, and / or the financial transaction information in the user information is purchased financial products, based on the target neural network prediction model, the user information and the information of the financial products to be purchased are analyzed to obtain the second recommended financial product information in the financial product recommendation information, wherein the second recommended financial product information is the recommended financial product information obtained based on the target neural network prediction model.

[0166] The processor may also call the information and application program stored in the memory through the transmission device to perform the following steps:

[0167] Step S206: constructing an initial model of the target neural network.

[0168] Step S207: Obtain the target user's preferred financial product information, where the preferred financial product information is the financial product information that the target user has subscribed to, collected, and purchased.

[0169] Step S208: Based on the target neural network initial model, input the user information, the rate of return and the risk rate in the preferred financial product information to obtain the financial product prediction score.

[0170] Step S209: Based on the preferred financial product scores and the financial product prediction scores in the preferred financial product information, the target neural network initial model is trained to obtain a target neural network prediction model.

[0171] The processor may also call the information and application program stored in the memory through the transmission device to perform the following steps:

[0172] Step S210: Structural processing is performed on the user information to obtain a user information vector.

[0173] Step S211: construct a convolutional layer based on the user information vector.

[0174] Step S212: Based on the convolutional layer, construct an initial model of the target neural network.

[0175] Step S213: constructing a first convolution layer and a second convolution layer based on the user feature information vector in the user information vector.

[0176] Step S214: constructing a third convolutional layer based on the financial management knowledge information vector in the user information vector.

[0177] The embodiment of the present application provides a solution for recommending financial products. The recommended financial product information is determined through user information and information about the financial products to be purchased, and the complex investment preferences of users and the multi-dimensional characteristics of financial products are captured and understood to achieve a more intelligent recommendation service, thereby solving the problem of low accuracy of financial product recommendations in related technologies, thereby achieving the effect of improving user satisfaction and investment decision-making efficiency.

[0178] It can be understood by those skilled in the art that Figure 5 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 5 The structure of the electronic device is not limited. Figure 5 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 5 Different configurations are shown.

[0179] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0180] Example 4

[0181] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the financial product recommendation method provided in the first embodiment.

[0182] Optionally, in this embodiment, the above storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0183] Example 5

[0184] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing the steps of the financial product recommendation method in the above-mentioned embodiment 1.

[0185] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0186] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0187] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0188] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0189] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0190] If the integrated unit is implemented in the form of 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 the present application can be essentially or partly or all or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc., which can store program code.

[0191] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for recommending financial products, characterized in that: include: Get the user information of the target user; Obtain information on financial products to be purchased; Determining recommended financial product information based on the user information and the information of the financial product to be purchased; Based on the recommended financial product information, a financial product recommendation list is generated, and the financial product recommendation list is used to display the recommended financial products.

2. The method according to claim 1, characterized in that Generating the financial product recommendation list based on the recommended financial product information includes: Obtaining a recommended financial product score in the recommended financial product information; The recommended financial products are sorted based on their scores to generate the financial product recommendation list.

3. The method according to claim 1 or 2, characterized in that: Determining the recommended financial product information based on the user information includes: When it is determined that the browsing interest information in the user information is not browsing financial products, cluster analysis is performed on the user information to obtain user classification information; Based on the user classification information, first recommended financial product information in the recommended financial product information is determined, wherein the first recommended financial product information is recommended financial product information obtained according to cluster analysis.

4. The method according to any one of claims 1 or 2, characterized in that Determining the recommended financial product information based on the user information includes: When it is determined that the browsing interest information in the user information is browsing financial products, and / or the financial transaction information in the user information is purchasing financial products, Based on the target neural network prediction model, the user information and the information of the financial product to be purchased are analyzed to obtain the second recommended financial product information in the financial product recommendation information, wherein the second recommended financial product information is the recommended financial product information obtained based on the target neural network prediction model.

5. The method according to claim 4, characterized in that Before analyzing the user information based on the target neural network prediction model, the method includes: Construct the initial model of the target neural network; Acquire the target user's preferred financial product information, where the preferred financial product information is information about financial products that the target user has subscribed to, collected, and purchased; Based on the target neural network initial model, the user information and the rate of return and risk rate in the preferred financial product information are input to obtain a financial product prediction score; Based on the preferred financial product scores in the preferred financial product information and the financial product prediction scores, the target neural network initial model is trained to obtain the target neural network prediction model.

6. The method according to claim 5, characterized in that Constructing the target neural network initial model, including: Performing structured processing on the user information to obtain a user information vector; Based on the user information vector, construct a convolutional layer; Based on the convolutional layer, the target neural network initial model is constructed.

7. The method according to claim 6, characterized in that Based on the user information vector, constructing the convolution layer includes: Based on a first user feature information vector in the user information vector, constructing a first convolutional layer, wherein the first user feature information vector includes an age feature vector and a city feature vector; constructing a second convolutional layer based on a second user feature information vector in the user information vector, wherein the second user feature information vector includes an investment capability information vector; Based on the third user feature information vector in the user information vector, a third convolutional layer is constructed; wherein the third user feature information vector includes a training degree vector, an experience value information vector and a financial management score information vector.

8. A financial product recommendation device, characterized in that: include: A first acquisition module, wherein the first acquisition module is used to acquire user information of a target user; A second acquisition module, the second acquisition module is used to acquire information of financial products to be purchased; A determination module, the determination module is used to determine recommended financial product information based on the user information and the information of the financial product to be purchased; The generation module and the generation determination module are used to generate a wealth management product recommendation list based on the recommended wealth management product information, and the wealth management product recommendation list is used to display the recommended wealth management products.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is executed, the device where the computer-readable storage medium is located is controlled to execute the financial product recommendation method 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 a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images