Financial management preference prediction method and device, equipment, storage medium and product

Abstract

The invention discloses a financing preference prediction method and device, equipment, a storage medium and a product, relates to the technical field of computers, and discloses a financing preference prediction method comprising the following steps: obtaining a historical purchase type and a corresponding historical purchase amount; and based on a preset financing preference prediction model, carrying out financing preference prediction on the historical purchase type and the historical purchase amount to obtain a financing preference prediction result, the financing preference prediction model being obtained by training a preset to-be-trained model by a purchase historical sample and a corresponding real purchase label. According to the invention, the purchase historical sample and the corresponding real purchase label are adopted to train the preset to-be-trained model, and the trained financing preference prediction model carries out financing preference prediction on the historical purchase type and the historical purchase amount to obtain the financing preference prediction result, thereby achieving the more accurate prediction of the customer preference, and improving the user experience. And therefore, the reliability of the financing preference prediction is remarkably improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to methods, devices, equipment, storage media and products for predicting financial management preferences. Background Art

[0002] With the development of Internet technology and the enhancement of people's financial management awareness, the financial management market is becoming increasingly prosperous. Financial management products can be divided into long-term, medium-term, medium-term and short-term products according to their terms. Specifically, long-term products have the highest returns but the worst liquidity; short-term products have the lowest returns but the best liquidity. Although the number of buyers of long-term financial management products is small, they help retain customers; while although the number of buyers of short-term financial management products is large, such customers are prone to loss. Therefore, the evaluation of customer preferences for financial management products has become increasingly important.

[0003] The existing financial management preference prediction technology uses the longest financial management period that a customer has ever purchased as a label, constructs a longest financial management preference model, and then predicts the customer's longest financial management period preference through the model. However, the existing technology can only evaluate whether the customer will buy when predicting financial management preferences, and does not distinguish the value of products with different periods, resulting in an inability to accurately predict the customer's longest financial management period preference, thereby reducing the accuracy of financial management preference prediction.

[0004] Therefore, how to improve the accuracy of financial management preference prediction is an urgent problem to be solved.

[0005] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0006] The main purpose of this application is to provide a financial management preference prediction method, device, equipment, storage medium and product, aiming to solve the technical problem of low accuracy of financial management preference prediction.

[0007] To achieve the above purpose, the present application proposes a method for predicting financial management preferences, the method comprising:

[0008] Get the historical purchase types and corresponding historical purchase amounts;

[0009] Based on a preset financial management preference prediction model, financial management preference prediction is performed on the historical purchase types and the historical purchase amounts to obtain a financial management preference prediction result, wherein the financial management preference prediction model is obtained by training a preset model to be trained using purchase history samples and corresponding real purchase labels.

[0010] In one embodiment, before the step of obtaining the historical purchase types and the corresponding historical purchase amounts, the step further includes:

[0011] Obtain purchase history samples and corresponding real purchase labels;

[0012] Performing value weight allocation on the purchase history samples to obtain a weighted sample of term products;

[0013] Determine the purchase amount sample corresponding to the purchase history sample;

[0014] Input the term product weight sample and the corresponding purchase amount sample into a preset model to be trained to obtain a financial management preference prediction label;

[0015] Based on the financial management preference prediction label and the real purchase label, query a preset loss weight table to obtain a loss weight sample;

[0016] By using a preset loss function and the loss weight sample, the predicted loss between the financial management preference prediction label and the real purchase label is calculated to obtain a sample loss value;

[0017] If the sample loss value does not meet the preset loss standard, return to the step of inputting the term product weight sample and the corresponding purchase amount sample into the preset model to be trained to obtain a financial management preference prediction label, and stop training until the sample loss value meets the loss standard to obtain a financial management preference prediction model that meets the loss standard.

[0018] In one embodiment, the step of inputting the term product weight samples and the corresponding purchase amount samples into a preset model to be trained to obtain a financial management preference prediction label includes:

[0019] The term product weight samples and the corresponding purchase amount samples are input into a preset prediction equation to obtain a financial management preference prediction value, wherein the prediction equation includes:

[0020]

[0021] Among them, pred is the predicted value of financial management preference, C i is the purchase amount of the ith product, T i is the weight of the i-th financial product;

[0022] According to a preset mapping relationship, the financial management preference prediction value is mapped to a financial management preference prediction label.

[0023] In one embodiment, the step of assigning value weights to the purchase history samples to obtain term product weight samples includes:

[0024] Determine a sample of purchase categories in a sample of purchase histories;

[0025] Classifying the purchase type samples by term days to obtain a plurality of term product samples and corresponding term days;

[0026] According to the term days and the preset value weights of different types, value weights are allocated to the different term product samples to obtain term product weight samples.

[0027] In one embodiment, the step of inputting the term product weight samples and the corresponding purchase amount samples into a preset model to be trained to obtain a financial management preference prediction label includes:

[0028] Segmenting the time in the purchase history sample according to a preset time interval to obtain a purchase history time sample;

[0029] Performing seasonal analysis on the purchase history time samples to obtain seasonal purchase characteristics;

[0030] Performing periodic analysis on the purchase history time samples to obtain periodic purchase features;

[0031] The seasonal purchase characteristics, the periodic purchase characteristics, the term product weight samples and the corresponding purchase amount samples are input into a preset model to be trained to obtain a financial management preference prediction label.

[0032] In one embodiment, the step of performing financial management preference prediction on the historical purchase types and the historical purchase amounts based on a preset financial management preference prediction model, and obtaining the financial management preference prediction result further includes:

[0033] Receive the current batch purchase type and the corresponding current batch purchase amount;

[0034] According to the current batch purchase type and the corresponding current batch purchase amount, updating the purchase history sample and the corresponding real purchase label;

[0035] The financial management preference prediction model is incrementally updated based on the purchase history samples and the corresponding real purchase labels.

[0036] In addition, to achieve the above purpose, the present application also proposes a financial management preference prediction device, which includes:

[0037] The acquisition module is used to obtain the historical purchase types and corresponding historical purchase amounts;

[0038] The prediction module is used to perform financial management preference prediction on the historical purchase types and the historical purchase amounts based on a preset financial management preference prediction model to obtain a financial management preference prediction result, wherein the financial management preference prediction model is obtained by training a preset model to be trained with purchase history samples and corresponding real purchase labels.

[0039] In addition, to achieve the above-mentioned purpose, the present application also proposes a financial management preference prediction device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the financial management preference prediction method described above.

[0040] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the financial management preference prediction method described above are implemented.

[0041] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the financial management preference prediction method as described above.

[0042] One or more technical solutions proposed in this application have at least the following technical effects:

[0043] Compared with the related art, in which the longest financial management period that the customer has ever bought in history is used as a label to construct a longest financial management preference model, and then the model is used to predict the customer's longest financial management period preference, the present application obtains the historical purchase type and the corresponding historical purchase amount; based on the preset financial management preference prediction model, the historical purchase type and the historical purchase amount are predicted for financial management preference, and the financial management preference prediction result is obtained, wherein the financial management preference prediction model is obtained by training the preset model to be trained by the purchase history sample and the corresponding real purchase label. It can be understood that the present application uses the purchase history sample and the corresponding real purchase label to train the preset model to be trained. When the historical purchase type and the corresponding historical purchase amount are obtained, the financial management preference prediction model is used to predict the financial management preference for the historical purchase type and the historical purchase amount, and the financial management preference prediction result is obtained. By taking the historical purchase type as an influencing factor, the customer's preference for different product types can be accurately evaluated. By adding the historical purchase amount as an influencing factor, the customer's preference can be more accurately predicted. Therefore, the financial management preference prediction based on the historical purchase type and the historical purchase amount can improve the accuracy of the financial management preference prediction, thereby significantly improving the reliability of the financial management preference prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0046] Figure 1 A flowchart of the first embodiment of the financial management preference prediction method of the present application is provided;

[0047] Figure 2 A label schematic diagram provided for the first embodiment of the financial management preference prediction method of this application;

[0048] Figure 3 A brief flow chart of the first embodiment of the financial management preference prediction method of the present application is provided;

[0049] Figure 4 This is a schematic diagram of the module structure of the financial management preference prediction device according to an embodiment of the present application;

[0050] Figure 5 Schematic diagram of the device structure of the hardware operating environment involved in the financial management preference prediction method in the embodiment of the present application.

[0051] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0052] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0053] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0054] The main solution of the embodiment of the present application is: obtaining historical purchase types and corresponding historical purchase amounts; based on a preset financial preference prediction model, performing financial preference prediction on the historical purchase types and the historical purchase amounts to obtain financial preference prediction results, wherein the financial preference prediction model is obtained by training a preset model to be trained using purchase history samples and corresponding real purchase labels.

[0055] In this embodiment, the present application uses the financial preference prediction device as the execution subject. For the convenience of description, it is specifically described below as "device".

[0056] The existing technology uses the longest financial management period that a customer has ever purchased as a label to construct a longest financial management preference model, and then predicts the customer's longest financial management period preference through the model. However, the existing technology can only evaluate whether the customer will buy when predicting financial management preferences, and does not distinguish the value of products with different terms, resulting in an inability to accurately predict the customer's longest financial management period preference, thereby reducing the accuracy of financial management preference prediction.

[0057] The present application provides a solution, which uses purchase history samples and corresponding real purchase labels to train a preset model to be trained. When the historical purchase types and the corresponding historical purchase amounts are obtained, the trained financial preference prediction model is used to predict the financial preference of the historical purchase types and the historical purchase amounts to obtain a financial preference prediction result. By taking the historical purchase types as an influencing factor, an accurate assessment of the customer's preferences for different product types is achieved. By adding the historical purchase amount as an influencing factor, a more accurate prediction of the customer's preferences is achieved. Therefore, the accuracy of the financial preference prediction based on the historical purchase types and the historical purchase amounts can be improved, thereby significantly improving the reliability of the financial preference prediction.

[0058] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a financial preference prediction device, etc. The following takes the financial preference prediction device as an example to illustrate this embodiment and the following embodiments.

[0059] Based on this, the present application embodiment provides a financial management preference prediction method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the financial management preference prediction method of the present application.

[0060] In this embodiment, the financial management preference prediction method includes steps S10 to S20:

[0061] Step S10, obtaining historical purchase types and corresponding historical purchase amounts;

[0062] It should be noted that the historical purchase types refer to the different financial products that the customer has purchased in the past period of time. The historical purchase amounts refer to the specific purchase amounts of different financial products purchased in the past.

[0063] For example, a customer purchased three different financial products in the past year, namely a 10,000 yuan stock product, a 5,000 yuan fund product and a 2,000 yuan bond product. By obtaining these historical purchase types and the corresponding purchase amounts, the device can analyze and predict the customer's financial preferences, so as to provide customers with more personalized and accurate financial advice and services.

[0064] Step S20, based on a preset financial preference prediction model, financial preference prediction is performed on the historical purchase types and the historical purchase amounts to obtain a financial preference prediction result, wherein the financial preference prediction model is obtained by training a preset model to be trained by using purchase history samples and corresponding real purchase labels.

[0065] It should be noted that the financial management preference prediction model is a trained prediction model, which is used to predict the financial management preferences of customers based on their financial management product purchase history. The historical purchase categories are the categories of different financial management products that customers have purchased in the past period of time. The historical purchase amount is the specific amount of money spent by the customer when purchasing financial management products. The purchase history samples are one or more sets of data sets containing the customer's purchase history, which are used to train the financial management preference prediction model so that it can more accurately predict the customer's financial management preferences. The real purchase label refers to the actual purchase situation corresponding to the purchase history sample, that is, the customer's real purchase behavior, which is used to evaluate and adjust the prediction accuracy of the model during the training process.

[0066] Exemplarily, the device inputs the acquired historical purchase types and corresponding historical purchase amounts into a preset financial preference prediction model. The model outputs financial preference prediction results by analyzing and processing the input data, combined with the patterns and rules learned in the previous training process. This result may include the customer's preference for different types of financial products, predictions of possible future purchase behaviors, etc. The device can use the customer's historical purchase data, combined with advanced machine learning technology, to scientifically and accurately predict the customer's financial preferences, thereby improving customer experience and satisfaction, while also bringing more business opportunities and value-added space to financial institutions.

[0067] Exemplarily, the device inputs the acquired historical purchase types and corresponding historical purchase amounts into a preset financial preference prediction model. The model outputs the customer's preference score for each financial product by analyzing and processing the input data, and selects the financial product type corresponding to the maximum preference score as the financial preference prediction result.

[0068] It is understandable that the device uses the trained financial preference prediction model to analyze the customer's past purchasing behavior to infer the customer's financial management tendencies.

[0069] The present embodiment provides a financial management preference prediction method, which uses purchase history samples and corresponding real purchase labels to train a preset model to be trained. When historical purchase types and corresponding historical purchase amounts are obtained, the trained financial management preference prediction model is used to predict the historical purchase types and the historical purchase amounts to obtain financial management preference prediction results. By using the historical purchase types as an influencing factor, an accurate assessment of the customer's preferences for different product types is achieved. By adding the historical purchase amount as an influencing factor, a more accurate prediction of the customer's preferences is achieved. Therefore, based on the financial management preference prediction of the historical purchase types and the historical purchase amounts, the accuracy of the financial management preference prediction can be improved, thereby significantly improving the reliability of the financial management preference prediction.

[0070] Based on the first embodiment of the present application, on this basis, before step S10, the financial management preference prediction method further includes:

[0071] Obtain purchase history samples and corresponding real purchase labels;

[0072] Performing value weight allocation on the purchase history samples to obtain a weighted sample of term products;

[0073] Determine the purchase amount sample corresponding to the purchase history sample;

[0074] Input the term product weight sample and the corresponding purchase amount sample into a preset model to be trained to obtain a financial management preference prediction label;

[0075] Based on the financial management preference prediction label and the real purchase label, query a preset loss weight table to obtain a loss weight sample;

[0076] By using a preset loss function and the loss weight sample, the predicted loss between the financial management preference prediction label and the real purchase label is calculated to obtain a sample loss value;

[0077] If the sample loss value does not meet the preset loss standard, return to the step of inputting the term product weight sample and the corresponding purchase amount sample into the preset model to be trained to obtain a financial management preference prediction label, and stop training until the sample loss value meets the loss standard to obtain a financial management preference prediction model that meets the loss standard.

[0078] It should be noted that the purchase history sample is a set of sample data selected from the customer's historical purchase records for training the model. The real purchase label is the actual purchase label corresponding to the purchase history sample, which is used to evaluate the accuracy of the model prediction. The value weight allocation is a method of weighting sample data, and different weights are allocated according to the importance and value of the sample. The term product weight sample is the sample data for different term products after value weight allocation, and each sample has a corresponding weight value. The purchase amount sample is the purchase amount data corresponding to the purchase history sample, which is used to reflect the specific amount of the customer's purchase behavior. The financial management preference prediction label is the customer financial management preference label predicted by the model based on the input data, which is used to indicate the customer's preference for different types of financial products. The loss weight table is a preset table used to query and allocate loss weights based on the difference between the predicted label and the real label to help optimize the model. The loss function is a function that evaluates the difference between the model prediction result and the real result. Commonly used loss functions include mean square error, cross entropy loss, etc. The loss weight sample is the sample data queried according to the loss weight table and used to adjust the weight of the loss function during the model training process. The sample loss value is calculated by the loss function, reflecting the predicted loss value of the model on a specific sample. The loss standard is a threshold set in advance, used to determine whether the predicted loss of the model is within an acceptable range. The model to be trained is a model that has not been fully trained and needs to be continuously iterated and optimized through input data. The financial preference prediction model that meets the loss standard is a model whose predicted loss value reaches or is lower than the preset loss standard after training, and is considered to have good prediction ability.

[0079] It is understandable that the above process is a typical machine learning model training process. Through continuous iteration and optimization, the model can more accurately predict the customer's financial preferences. The whole process can be divided into the following steps:

[0080] Obtaining purchase history samples and corresponding real purchase labels is the basic data preparation stage for training models. Purchase history samples are records of customers’ past purchase behaviors, while real purchase labels are the actual results of these purchase behaviors, which are used for label comparison in subsequent model training.

[0081] In the step of assigning value weights to the purchase history samples and obtaining weight samples of term products, the purchase history samples are assigned weights, possibly to highlight the impact of the purchase behavior of certain specific term products on the model training, thereby obtaining more accurate prediction results.

[0082] The purchase amount sample reflects the specific amount of the customer's purchase behavior. The step of determining the purchase amount sample corresponding to the purchase history sample is to use the purchase amount as an important factor in model training.

[0083] Input the term product weight samples and the corresponding purchase amount samples into the preset model to be trained. The financial management preference prediction label is obtained by inputting the processed data into the model for training. The model predicts the customer's financial management preference label based on the input data.

[0084] Based on the predicted labels of financial preference and the real purchase labels, the preset loss weight table is queried to obtain the loss weight samples by comparing the predicted labels and the real labels and querying the loss weight table. The purpose is to optimize the prediction performance of the model in subsequent steps.

[0085] Through the preset loss function and loss weight sample, the prediction loss between the financial preference prediction label and the actual purchase label is calculated. The sample loss value is obtained by using the loss function to calculate the loss between the predicted label and the actual label, and multiplying it by the corresponding loss weight to obtain the sample loss value.

[0086] If the sample loss value does not meet the preset loss standard, return to the retraining step until the sample loss value meets the loss standard and then stop training. Obtaining a financial preference prediction model that meets the loss standard is an iterative optimization process. If the model's predicted loss does not meet the preset standard, the model parameters need to be repeatedly adjusted until the predetermined loss standard is met.

[0087] In a feasible implementation, the above financial management preference prediction method may include:

[0088] The term product weight samples and the corresponding purchase amount samples are input into a preset prediction equation to obtain a financial management preference prediction value, wherein the prediction equation includes:

[0089]

[0090] Among them, pred is the predicted value of financial management preference, C i is the purchase amount of the ith product, T i is the weight of the i-th financial product;

[0091] According to a preset mapping relationship, the financial management preference prediction value is mapped to a financial management preference prediction label.

[0092] It should be noted that the term product weight sample is the sample data for different term products after value weight allocation, and each sample has a corresponding weight value, which is used to reflect the importance of different term products in the customer's purchasing behavior. The purchase amount sample is the purchase amount data corresponding to the purchase history sample, which is used to reflect the specific amount of the customer's purchasing behavior. The preset prediction equation is a mathematical model, which is used to calculate the financial management preference prediction value based on the input purchase amount sample and weight sample. The financial management preference prediction value is calculated by the prediction equation, which reflects the customer's preference for a certain type of financial management product. The mapping relationship is a rule or function that converts the financial management preference prediction value into a financial management preference prediction label, which is used to convert the prediction value into a discrete label form. The financial management preference prediction label is a label form obtained by converting the financial management preference prediction value according to the mapping relationship, which is used to indicate the customer's preference for different types of financial management products, which is convenient for subsequent model evaluation and application.

[0093] In a feasible implementation manner, the above financial management preference prediction method may further include:

[0094] Determine a sample of purchase categories in a sample of purchase histories;

[0095] Classifying the purchase type samples by term days to obtain a plurality of term product samples and corresponding term days;

[0096] According to the term days and the preset value weights of different types, value weights are allocated to the different term product samples to obtain term product weight samples.

[0097] It should be noted that the purchase category samples are sample data on purchase categories selected from the customer's historical purchase records, which are used to reflect the specific types of customer purchase behavior. The term day classification is a method of classifying purchase category samples according to the term days in which the purchase behavior occurs, which is used to distinguish purchase behaviors within different terms. The term product samples are purchase behavior samples within different terms after term day classification, and each sample carries corresponding term day information. The value weight is a method of weighting sample data, and different weights are assigned according to the importance and value of the sample. The different types of value weights are different weight values ​​assigned according to different types of sample data (such as products with different terms), which are used to reflect the importance of different types of samples in model training. The term product weight samples are sample data for products with different terms after value weight assignment, and each sample carries a corresponding weight value, which is used to reflect the importance of products with different terms in customer purchase behavior.

[0098] In a feasible implementation, the above financial management preference prediction method may include:

[0099] Segmenting the time in the purchase history sample according to a preset time interval to obtain a purchase history time sample;

[0100] Performing seasonal analysis on the purchase history time samples to obtain seasonal purchase characteristics;

[0101] Performing periodic analysis on the purchase history time samples to obtain periodic purchase features;

[0102] The seasonal purchase characteristics, the periodic purchase characteristics, the term product weight samples and the corresponding purchase amount samples are input into a preset model to be trained to obtain a financial management preference prediction label.

[0103] It should be noted that the preset time interval is a pre-set time span used to segment the time in the purchase history sample. The purchase history time sample is sample data containing the time information of the customer's purchase behavior after time segmentation processing. The seasonal analysis is a time series analysis method used to identify and extract seasonal patterns and features in purchase behavior. The seasonal purchase feature is obtained through seasonal analysis, and is characteristic data reflecting the seasonal change law in the customer's purchase behavior. The periodic analysis is a time series analysis method used to identify and extract periodic patterns and features in purchase behavior. The periodic purchase feature is obtained through periodic analysis, and is characteristic data reflecting the periodic change law in the customer's purchase behavior. The model to be trained is a machine learning model that has not yet been trained, which is used to predict the customer's financial management preferences based on input data. The financial management preference prediction label is calculated by the model to be trained, and is in the form of a label that reflects the customer's preference for a certain type of financial product, and is used for subsequent model evaluation and application.

[0104] It is understandable that first a fixed time period or interval needs to be defined, such as quarterly, semi-annual or annual divisions. Then, the transaction time in the purchase history sample is segmented according to this preset time interval to form a purchase history time sample. The purpose of this is to capture possible seasonal or cyclical purchase behavior patterns. Next, the purchase history time sample is subjected to seasonal analysis. This analysis usually uses statistical methods such as seasonal decomposition and moving average to identify and quantify purchase behavior patterns within a specific time period. For example, it may be found that the sales of financial products in a certain quarter are significantly higher than those in other quarters. This is the so-called seasonal purchase feature. Similarly, the purchase history time sample is also subjected to periodic analysis to identify any potential periodic purchase patterns. This may involve analysis of different periods such as weeks, months or years. Periodic features may include weekly weekend sales peaks, or annual sales surges on specific holidays. Finally, the seasonal purchase features, periodic purchase features, and term product weight samples are input into the preset model to be trained together with the corresponding purchase amount samples. The goal of this model is to learn and understand the complex relationships between the input data in order to make accurate predictions on new data in the future. After the model training is completed, the financial preference labels of new purchase history data can be predicted. These labels may include preference levels or priorities for various financial products, thereby helping financial institutions better understand customer needs and provide personalized services.

[0105] In a feasible implementation, after step S20, the financial management preference prediction method may include:

[0106] Receive the current batch purchase type and the corresponding current batch purchase amount;

[0107] According to the current batch purchase type and the corresponding current batch purchase amount, updating the purchase history sample and the corresponding real purchase label;

[0108] The financial management preference prediction model is incrementally updated based on the purchase history samples and the corresponding real purchase labels.

[0109] It should be noted that the current batch purchase types are different types of financial products purchased by customers at the current time point or time period. The current batch purchase amount is the purchase amount corresponding to the current batch purchase type, that is, the specific amount spent by the customer on each financial product. The purchase history sample is a sample data that records the customer's past purchase behavior of financial products, including information such as the customer's purchase amount at different times and for different types of financial products. The real purchase label is a mark of the purchase history sample, which reflects the characteristics of the customer's actual purchase behavior and is used for real value comparison during model training. The incremental update is a model update method, which means that based on the existing model, the model is locally adjusted and optimized using the newly collected data (i.e., the current batch purchase types and the corresponding purchase amounts), rather than retraining the entire model. The financial preference prediction model is a trained machine learning model that is used to predict the customer's future financial preferences based on the input purchase history samples and real purchase labels.

[0110] For example, in order to help understand the implementation process of the financial management preference prediction method obtained by combining this embodiment with the above embodiments, please refer to Figure 3 , Figure 3 A brief flowchart of a financial management preference prediction method is provided, specifically:

[0111] In banks, financial products can be divided into long-term (D), medium- and long-term (C), medium-term (B) and short-term (A) types according to their terms. In terms of returns, the long-term type has the highest yield and the short-term type has the lowest; in terms of liquidity, the long-term type has the worst and the short-term type has the best.

[0112] When training the financial management preference prediction model, we first define the weights of different types of financial management products based on the investment period. As shown in Table 1, the weight value is positively correlated with the term value of the product. The larger the term value, the larger the weight value.

[0113] Table 1 The weights of the maturity of different financial products

[0114]

[0115] Then, based on the weight and amount of the customer's purchased products, the predicted value of financial management preference is calculated using the formula:

[0116]

[0117] Among them, pred is the predicted value of financial management preference, C i is the purchase amount of the ith product, T i is the weight of the i-th financial product; the pred value reflects the customer's preference and the value of the term. The larger the pred value, the more the customer prefers products with higher term values.

[0118] In order to correspond the pred value to the label, such as Figure 2 As shown in the figure, the pred value is mapped to different types of financial products according to the threshold. The mapping relationship between the pred value and the label is as follows:

[0119] If the pred value is less than or equal to 1, the corresponding labeling result is a Class A product.

[0120] If the pred value is greater than 1 but less than or equal to 2, the corresponding labeling result is a Class B product.

[0121] If the pred value is greater than 2 but less than or equal to 4, the corresponding labeling result is a Class C product.

[0122] If the pred value is greater than 4, the corresponding labeling result is a Class D product.

[0123] The financial preference prediction model generates four-classification (A\B\C\D) model samples according to the above labeling method, and then conducts multi-classification model training. When the predicted value and the true value of each sample are compared to calculate the loss, the loss weight table is used to weight the loss weight of the model prediction and the true value, as shown in Table 2:

[0124] Table 2 Loss weights of model predictions and true values

[0125]

[0126] After the model is trained, model prediction is performed to obtain the customer's preference scores for the four types of financial products. Finally, the category corresponding to the largest of the four scores of each customer (A\B\C\D) is the customer's financial preference prediction result.

[0127] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the financial preference prediction method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0128] This application also provides a financial management preference prediction device, please refer to Figure 4 , the financial management preference prediction device comprises:

[0129] An acquisition module 10 is used to acquire historical purchase types and corresponding historical purchase amounts;

[0130] The prediction module 20 is used to perform financial preference prediction on the historical purchase types and the historical purchase amounts based on a preset financial preference prediction model to obtain a financial preference prediction result, wherein the financial preference prediction model is obtained by training a preset model to be trained with purchase history samples and corresponding real purchase labels.

[0131] And / or, the financial management preference prediction device includes:

[0132] The first acquisition module is used to obtain purchase history samples and corresponding real purchase tags;

[0133] A first weight module is used to assign value weights to the purchase history samples to obtain a weighted sample of term products;

[0134] A first determination module is used to determine a purchase amount sample corresponding to a purchase history sample;

[0135] A first labeling module is used to input the term product weight sample and the corresponding purchase amount sample into a preset to-be-trained model to obtain a financial management preference prediction label;

[0136] A first query module is used to query a preset loss weight table based on the financial management preference prediction label and the real purchase label to obtain a loss weight sample;

[0137] A first calculation module, used to calculate the predicted loss of the financial preference prediction label and the real purchase label through a preset loss function and the loss weight sample, and obtain a sample loss value;

[0138] The first loop module is used to return to the step of inputting the term product weight sample and the corresponding purchase amount sample into the preset model to be trained to obtain a financial management preference prediction label if the sample loss value does not meet the preset loss standard, and stop training after the sample loss value meets the loss standard to obtain a financial management preference prediction model that meets the loss standard.

[0139] And / or, the first marking module includes:

[0140] The first input module is used to input the term product weight sample and the corresponding purchase amount sample into a preset prediction equation to obtain a financial management preference prediction value, wherein the prediction equation includes:

[0141]

[0142] Among them, pred is the predicted value of financial management preference, C i is the purchase amount of the ith product, T i is the weight of the i-th financial product;

[0143] The first mapping module is used to map the financial management preference prediction value into a financial management preference prediction label according to a preset mapping relationship.

[0144] And / or, the first weight module includes:

[0145] A second determination module is used to determine a purchase category sample in the purchase history sample;

[0146] A first classification module is used to classify the purchase type samples by term days to obtain a plurality of term product samples and corresponding term days;

[0147] The first allocation module is used to allocate value weights to the different term product samples according to the term days and preset value weights of different types to obtain term product weight samples.

[0148] And / or, the first marking module includes:

[0149] A first segmentation module, used to segment the time in the purchase history sample according to a preset time interval to obtain a purchase history time sample;

[0150] A first analysis module is used to perform seasonal analysis on the purchase history time samples to obtain seasonal purchase characteristics;

[0151] A second analysis module is used to perform periodic analysis on the purchase history time samples to obtain periodic purchase features;

[0152] The second input module is used to input the seasonal purchase characteristics, the periodic purchase characteristics, the term product weight samples and the corresponding purchase amount samples into a preset model to be trained to obtain a financial management preference prediction label.

[0153] And / or, the financial management preference prediction device includes:

[0154] A first receiving module, used for receiving the current batch purchase type and the corresponding current batch purchase amount;

[0155] A first updating module, configured to update the purchase history sample and the corresponding real purchase label according to the current batch purchase type and the corresponding current batch purchase amount;

[0156] The second updating module is used to incrementally update the financial management preference prediction model based on the purchase history samples and the corresponding real purchase tags.

[0157] The financial management preference prediction device provided by the present application adopts the financial management preference prediction method in the above embodiment, which can solve the technical problem of low accuracy of financial management preference prediction. Compared with the prior art, the beneficial effects of the financial management preference prediction device provided by the present application are the same as the beneficial effects of the financial management preference prediction method provided by the above embodiment, and other technical features in the financial management preference prediction device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0158] The present application provides a financial management preference prediction device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the financial management preference prediction method in the above-mentioned embodiment one.

[0159] Reference below Figure 5 , which shows a schematic diagram of the structure of a financial preference prediction device suitable for implementing the embodiment of the present application. The financial preference prediction device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, tablet computers, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PMPs (Portable Media Players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5 The financial preference prediction device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0160] like Figure 5 As shown, the financial preference prediction device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the financial preference prediction device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the financial preference prediction device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a financial preference prediction device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0161] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0162] The financial management preference prediction device provided by the present application adopts the financial management preference prediction method in the above embodiment, which can solve the technical problem of low accuracy of financial management preference prediction. Compared with the prior art, the beneficial effects of the financial management preference prediction device provided by the present application are the same as the beneficial effects of the financial management preference prediction method provided by the above embodiment, and other technical features in the financial management preference prediction device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0163] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0164] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0165] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the financial management preference prediction method in the above-mentioned embodiment.

[0166] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0167] The computer-readable storage medium may be included in the financial management preference prediction device; or may exist independently without being assembled into the financial management preference prediction device.

[0168] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the financial preference prediction device, the financial preference prediction device: obtains historical purchase types and corresponding historical purchase amounts; performs financial preference prediction on the historical purchase types and the historical purchase amounts based on a preset financial preference prediction model to obtain a financial preference prediction result, wherein the financial preference prediction model is obtained by training a preset model to be trained by pairing purchase history samples with corresponding real purchase labels.

[0169] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0170] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0171] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0172] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned financial management preference prediction method, and can solve the technical problem of low accuracy of financial management preference prediction. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the financial management preference prediction method provided in the above-mentioned embodiment, and will not be repeated here.

[0173] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned financial management preference prediction method when executed by a processor.

[0174] The computer program product provided by the present application can solve the technical problem of low accuracy in financial management preference prediction. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as those of the financial management preference prediction method provided by the above embodiment, which will not be elaborated here.

[0175] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for predicting financial management preferences, characterized in that: The method includes: Get the historical purchase types and corresponding historical purchase amounts; Based on a preset financial management preference prediction model, financial management preference prediction is performed on the historical purchase types and the historical purchase amounts to obtain a financial management preference prediction result, wherein the financial management preference prediction model is obtained by training a preset model to be trained using purchase history samples and corresponding real purchase labels.

2. The method according to claim 1, characterized in that The step of obtaining the historical purchase types and the corresponding historical purchase amounts also includes: Obtain purchase history samples and corresponding real purchase labels; Performing value weight allocation on the purchase history samples to obtain a weighted sample of term products; Determine the purchase amount sample corresponding to the purchase history sample; Input the term product weight sample and the corresponding purchase amount sample into a preset model to be trained to obtain a financial management preference prediction label; Based on the financial management preference prediction label and the real purchase label, query a preset loss weight table to obtain a loss weight sample; By using a preset loss function and the loss weight sample, the predicted loss between the financial management preference prediction label and the real purchase label is calculated to obtain a sample loss value; If the sample loss value does not meet the preset loss standard, return to the step of inputting the term product weight sample and the corresponding purchase amount sample into the preset model to be trained to obtain a financial management preference prediction label, and stop training until the sample loss value meets the loss standard to obtain a financial management preference prediction model that meets the loss standard.

3. The method according to claim 2, characterized in that The step of inputting the term product weight sample and the corresponding purchase amount sample into a preset model to be trained to obtain a financial management preference prediction label includes: The term product weight samples and the corresponding purchase amount samples are input into a preset prediction equation to obtain a financial management preference prediction value, wherein the prediction equation includes: Among them, pred is the predicted value of financial management preference, C i is the purchase amount of the ith product, T i is the weight of the i-th financial product; According to a preset mapping relationship, the financial management preference prediction value is mapped to a financial management preference prediction label.

4. The method according to claim 2, characterized in that The step of assigning value weights to the purchase history samples to obtain term product weight samples comprises: Determine a sample of purchase categories in a sample of purchase histories; Classifying the purchase type samples by term days to obtain a plurality of term product samples and corresponding term days; According to the term days and the preset value weights of different types, value weights are allocated to the different term product samples to obtain term product weight samples.

5. The method according to claim 2, characterized in that The step of inputting the term product weight sample and the corresponding purchase amount sample into a preset model to be trained to obtain a financial management preference prediction label includes: Segmenting the time in the purchase history sample according to a preset time interval to obtain a purchase history time sample; Performing seasonal analysis on the purchase history time samples to obtain seasonal purchase characteristics; Performing periodic analysis on the purchase history time samples to obtain periodic purchase features; The seasonal purchase characteristics, the periodic purchase characteristics, the term product weight samples and the corresponding purchase amount samples are input into a preset model to be trained to obtain a financial management preference prediction label.

6. The method according to claim 1, characterized in that The step of performing financial management preference prediction on the historical purchase types and the historical purchase amounts based on the preset financial management preference prediction model to obtain the financial management preference prediction result further includes: Receive the current batch purchase type and the corresponding current batch purchase amount; According to the current batch purchase type and the corresponding current batch purchase amount, updating the purchase history sample and the corresponding real purchase label; The financial management preference prediction model is incrementally updated based on the purchase history samples and the corresponding real purchase labels.

7. A financial management preference prediction device, characterized in that: The device comprises: The acquisition module is used to obtain the historical purchase types and corresponding historical purchase amounts; The prediction module is used to perform financial preference prediction on the historical purchase types and the historical purchase amounts based on a preset financial preference prediction model to obtain a financial preference prediction result, wherein the financial preference prediction model is obtained by training a preset model to be trained with purchase history samples and corresponding real purchase labels.

8. A financial management preference prediction device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the financial management preference prediction method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the financial management preference prediction method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the financial management preference prediction method according to any one of claims 1 to 6 are implemented.

Images