Fund purchase amount and purchase deadline estimation method and device, equipment and medium

Abstract

The embodiment of the invention provides a fund purchase amount and purchase deadline estimation method and device, equipment and a medium, and relates to the field of artificial intelligence. The method comprises the steps that a time node when a user logs in a fund purchase interface is determined, historical fund data of the user and historical fund data of a fund are determined based on the time node, and the historical fund data comprise historical transaction data and historical asset data; the historical transaction data is used for indicating a corresponding historical transaction condition and an investment period; and based on the historical transaction data, the historical asset data and the historical fund data, performing pre-estimation processing on a fund purchase amount and a purchase deadline by using a pre-estimation model to obtain a pre-estimated purchase amount and a pre-estimated purchase deadline, the pre-estimation model being constructed based on a cyclic convolutional neural network. According to the method provided by the invention, the investment experience when the user purchases the financial product is improved, and the user satisfaction and the user stickiness are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a method, device, equipment and medium for estimating a fund purchase amount and purchase period. Background Art

[0002] With the rapid development of the Internet, online platforms provide users with comprehensive fund information, allowing users to purchase funds through the Internet, providing users with great convenience. Online fund purchase has gradually been favored by many users for its comprehensiveness, convenience, and speed.

[0003] Currently, the existing technology obtains a fund prediction model through training of historical fund net value data, and predicts whether to purchase funds and fund shares on the next day based on the fund net value data.

[0004] However, the variable input into the fund prediction model is only a single variable, lacking multi-feature information, and the accuracy of the model prediction results is difficult to guarantee; in addition, there is a lack of estimation of the fund purchase period, so it is difficult for users to determine the fund investment duration, and it is difficult for users to grasp the fund purchase period, resulting in losses on the selling day, affecting user experience and causing bank customer loss. Summary of the invention

[0005] The present application provides a method, device, equipment and medium for estimating the fund purchase amount and purchase period, which are used to solve the technical problem of inputting a single variable into a fund prediction model and determining the estimated fund purchase period in the prior art.

[0006] In a first aspect, the present application provides a method for estimating a fund purchase amount and a purchase period, including:

[0007] Determine the time point when the user logs into the fund purchase interface, and based on the time point, determine the historical fund data of the user and the historical fund data of the fund, the historical fund data includes historical transaction data and historical asset data, and the historical transaction data is used to indicate the corresponding historical transaction situation and investment period;

[0008] Based on the historical transaction data, historical asset data and historical fund data, an estimation model is used to estimate the fund purchase amount and purchase period to obtain an estimated purchase amount and an estimated purchase period. The estimation model is constructed based on a recurrent convolutional neural network.

[0009] In a possible implementation, the above-mentioned estimating model is used to estimate the fund purchase amount and purchase period based on the historical transaction data, historical asset data and historical fund data to obtain the estimated purchase amount and the estimated purchase period, including:

[0010] Based on the historical transaction data, determining whether the user is a first-time purchasing user;

[0011] If the user is not a first-time purchaser, based on the historical transaction data, historical asset data, historical fund data and historical experience data, an estimation model is used to estimate the fund purchase amount and purchase period to obtain an estimated purchase amount and an estimated purchase period;

[0012] Among them, the historical experience data includes historical purchase amount difference and historical purchase period difference, the historical purchase amount difference is used to indicate the absolute difference between the historical estimated purchase amount and the historical purchase amount, and the historical purchase period difference is used to indicate the absolute difference between the historical estimated purchase period and the historical purchase period.

[0013] In a possible implementation, the historical fund data includes fund yields corresponding to various types of funds, and the historical asset data includes: historical account balances and historical annual average daily assets. The method further includes:

[0014] In the case where the user is a first-time purchaser, the fund with the largest fund yield among multiple types of funds is used as the target fund, and the purchase period of the target fund is used as the estimated purchase period;

[0015] The estimated purchase amount is determined based on the historical account balance, the historical annual average daily assets and the fund yield of the target fund, wherein the historical account balance, the historical annual average daily assets and the fund yield are used to indicate data corresponding to multiple types of standard periods.

[0016] In a possible implementation, the estimation model includes a sequence prediction network and a decision network. Based on the historical transaction data, historical asset data, historical fund data, and historical experience data, the estimation model is used to estimate the fund purchase amount and purchase period to obtain the estimated purchase amount and the estimated purchase period, including:

[0017] Standardize the historical transaction data, historical asset data, and historical fund data according to multiple standard periods, and determine a subset of historical standard data corresponding to each standard period;

[0018] For any one of the multiple historical standard data subsets, multiple historical standard data and historical experience data in the historical standard data subset are input into the sequence prediction network to obtain data prediction sequences in multiple time periods;

[0019] The data prediction sequence is input into the decision network to obtain the estimated fund purchase amount and purchase period.

[0020] In a possible implementation, the multiple historical standard data and historical experience data in the historical standard data subset are input into the sequence prediction network to obtain data prediction sequences in multiple time periods, including:

[0021] Normalizing the multiple historical standard data and historical experience data in the historical standard data subset;

[0022] Inputting the normalized multiple historical standard data and historical experience data into a sequence prediction network to determine a hidden state corresponding to each time period, wherein the hidden state is used to indicate a long-term dependency relationship in the sequence data;

[0023] The hidden states corresponding to each time period are processed by the fully connected layer and the average pooling layer and then spliced ​​to obtain the data prediction sequence in multiple time periods.

[0024] In a possible implementation, the historical standard data subset includes a first training subset and a first validation subset, the historical experience data second training subset and a second validation subset, and the method further includes: estimating the fund purchase amount and purchase period based on the historical transaction data, historical asset data, historical fund data, and historical experience data using an estimation model to obtain the estimated purchase amount and the estimated purchase period.

[0025] Inputting the data in the first training subset and the second training subset into a pre-built estimation model for training to obtain an alternative estimation model;

[0026] Verifying the candidate estimation model according to the data in the first validation subset and the second validation subset to obtain a verification result, wherein the verification result is used to indicate whether the candidate estimation model converges;

[0027] If the alternative estimation model does not converge, optimizing the model parameters of the alternative estimation model, and performing training based on the optimized alternative estimation model;

[0028] If the candidate estimation model converges, the candidate estimation model is used as the estimation model.

[0029] In a possible implementation, the method further includes:

[0030] Obtaining the actual purchase amount and actual purchase period of the user;

[0031] Performing absolute difference processing on the actual purchase amount and the estimated purchase amount, the actual purchase period and the estimated purchase period, respectively, to obtain a purchase amount difference and a purchase period difference;

[0032] Based on the purchase amount difference and the purchase period difference, the historical experience data is updated, and based on the updated historical experience data, the estimation model is optimized.

[0033] In a second aspect, the present application provides a device for estimating a fund purchase amount and a purchase period, comprising:

[0034] A determination module, used to determine the time node when the user logs into the fund purchase interface, and based on the time node, determine the historical fund data of the user and the historical fund data of the fund, wherein the historical fund data includes historical transaction data and historical asset data, and the historical transaction data is used to indicate the corresponding historical transaction situation and investment period;

[0035] The estimation module is used to estimate the fund purchase amount and purchase period based on the historical transaction data, historical asset data and historical fund data using an estimation model to obtain an estimated purchase amount and an estimated purchase period.

[0036] In a possible implementation manner, the device further includes: a judgment module;

[0037] The judgment module is used to judge whether the user is a first-time purchase user based on the historical transaction data;

[0038] In the case where the user is not a first-time purchaser, the estimation module is further used to estimate the fund purchase amount and purchase period using an estimation model based on the historical transaction data, historical asset data, historical fund data and historical experience data to obtain an estimated purchase amount and an estimated purchase period;

[0039] Among them, the historical experience data includes historical purchase amount difference and historical purchase period difference, the historical purchase amount difference is used to indicate the absolute difference between the historical estimated purchase amount and the historical purchase amount, and the historical purchase period difference is used to indicate the absolute difference between the historical estimated purchase period and the historical purchase period.

[0040] In a possible implementation, the historical fund data includes fund yields corresponding to various types of funds, and the historical asset data includes: historical account balances, historical annual average daily assets,

[0041] In the case where the user is a first-time buyer, the estimation module is further used to take the fund with the largest fund yield among multiple types of funds as the target fund, and take the purchase period of the target fund as the estimated purchase period; determine the estimated purchase amount based on the historical account balance, historical annual average daily assets and the fund yield of the target fund, wherein the historical account balance, historical annual average daily assets and fund yield are used to indicate data corresponding to multiple types of standard periods.

[0042] In one possible implementation, the estimation model includes a sequence prediction network and a decision network. The estimation module is further used to standardize the historical transaction data, historical asset data, and historical fund data according to multiple standard periods to determine a historical standard data subset corresponding to each standard period; for any one of the multiple historical standard data subsets, multiple historical standard data and historical experience data in the historical standard data subset are input into the sequence prediction network to obtain data prediction sequences within multiple time periods; the data prediction sequence is input into the decision network to obtain an estimated fund purchase amount and purchase period.

[0043] In a possible implementation, the estimation module is also used to normalize the multiple historical standard data and historical experience data in the historical standard data subset; input the normalized multiple historical standard data and historical experience data into the sequence prediction network to determine the hidden state corresponding to each time period, and the hidden state is used to indicate the long-term dependency relationship in the sequence data; the hidden state corresponding to each time period is processed by the fully connected layer and the average pooling layer and then spliced ​​to obtain the data prediction sequence in multiple time periods.

[0044] In a possible implementation, each historical standard data subset includes a first training subset and a first validation subset, the historical experience data second training subset and a second validation subset, and the device further includes: a training module;

[0045] The training module is used to input the data in the first training subset and the second training subset into a pre-built estimation model for training to obtain an alternative estimation model; verify the alternative estimation model according to the data in the first verification subset and the second verification subset to obtain a verification result, and the verification result is used to indicate whether the alternative estimation model converges; if the alternative estimation model does not converge, optimize the model parameters of the alternative estimation model and train based on the optimized alternative estimation model; if the alternative estimation model converges, use the alternative estimation model as the estimation model.

[0046] In a possible implementation, the device also includes: an updating module, used to obtain the actual purchase amount and actual purchase period of the user; performing absolute difference processing on the actual purchase amount and estimated purchase amount, and the actual purchase period and estimated purchase period, respectively, to obtain a purchase amount difference and a purchase period difference; based on the purchase amount difference and the purchase period difference, updating the historical experience data, and optimizing the estimation model based on the updated historical experience data.

[0047] In a third aspect, an embodiment of the present application provides a device for estimating a fund purchase amount and a purchase period, including: a memory, a processor;

[0048] The memory stores computer-executable instructions;

[0049] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0050] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementations of the first aspect.

[0051] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect.

[0052] The method, device, equipment and medium for estimating the fund purchase amount and purchase period provided in the present application estimate the fund purchase amount and purchase period for users who are not purchasing the fund for the first time through an estimation model trained based on historical standard data and historical experience data; for users who are purchasing the fund for the first time, the fund purchase amount and purchase period are estimated through historical standard data, which improves the accuracy of the estimated purchase amount and the estimated purchase period, enhances the user's investment experience when purchasing financial products, and improves user satisfaction and user stickiness. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] 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.

[0054] Figure 1 A schematic diagram of a scenario of a method for estimating a fund purchase amount and purchase period provided in an embodiment of the present application;

[0055] Figure 2 A schematic diagram of a method for estimating a fund purchase amount and purchase period provided in an embodiment of the present application Figure 1 ;

[0056] Figure 3 A schematic diagram of a method for estimating a fund purchase amount and purchase period provided in an embodiment of the present application Figure 2 ;

[0057] Figure 4 A schematic diagram of the structure of a device for estimating fund purchase amount and purchase period provided in an embodiment of the present application;

[0058] Figure 5 A schematic diagram of the structure of a device for estimating fund purchase amount and purchase period provided in an embodiment of the present application.

[0059] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0060] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0061] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0062] In addition, this application involves conducting big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.), and using artificial intelligence technology to make automated decisions, and making technical solutions that have a significant impact on personal rights and interests based on the results of automated decisions. The application provides users with corresponding operation entrances for them to choose to agree or reject the results of automated decisions; if the user chooses to reject, the expert decision-making process will be entered.

[0063] It should be noted that the method, device, equipment and medium for estimating the fund purchase amount and purchase period provided in this application can be used in the field of artificial intelligence, and can also be used in any field other than artificial intelligence. The application field of the method, device, equipment and medium for estimating the fund purchase amount and purchase period in this application is not limited.

[0064] With the rapid development of the Internet, online platforms provide users with comprehensive fund information, allowing users to purchase funds through the Internet, providing users with great convenience. Online fund purchase has gradually been favored by many users for its comprehensiveness, convenience, and speed.

[0065] The Convolutional Recurrent Neural Network (CRNN) enables the network to process variable-length input sequences by expanding the network on the time series. The output of each time step depends not only on the input of the current time step, but also on the previous hidden state. This structure enables the recurrent convolutional network to capture the temporal dependency in sequence data.

[0066] Figure 1 A schematic diagram of a scenario of a method for estimating a fund purchase amount and purchase period provided in an embodiment of the present application, such as Figure 1 As shown, the user logs into the fund purchase interface of the fund purchase APP through the user terminal 101, and the server 102 obtains the time node of the user logging into the fund purchase interface pushed by the fund purchase APP. The server 102 can exchange information with the user terminal 101 in a wired or wireless manner.

[0067] The server 102 obtains historical transaction data, historical asset data and historical fund data based on the time node pushed by the user terminal 101, and estimates the fund purchase amount and purchase period based on the estimation model, and then pushes the estimated purchase amount and estimated purchase period to the fund purchase interface to provide a reference for users to purchase funds.

[0068] Combined with the above scenario, it can be seen that in the prior art, the fund prediction model is obtained by training the historical fund net value data, and the fund net value data is used to predict whether to purchase the fund and the fund shares on the next day. Since the variable input into the fund prediction model is only a single variable and lacks multi-feature information, the accuracy of the model prediction result is difficult to guarantee; in addition, the prior art lacks an estimate of the fund purchase period, making it difficult for users to determine the fund investment duration. For users who do not understand financial management knowledge, they only use the fund net value of the day as the basis for fund purchase without considering the change of the fund net value, which is easy to suffer losses on the selling day, affecting the user experience and leading to the loss of bank customers; the prior art lacks a comprehensive estimate of the fund purchase amount based on the user's own asset data, and the user's life may be affected by the consumption of funds.

[0069] The method for estimating the fund purchase amount and purchase period provided in the present application solves the technical problems in the prior art of a single variable input into a fund prediction model and lacking multi-feature information through the user's historical transaction data and historical asset data and the fund's historical fund data; estimates the fund purchase amount and purchase period based on a recurrent convolutional neural network estimation model to capture the time dependency in the sequence data, thereby solving the technical problems in the prior art of lack of estimation of the fund purchase period and the loss caused by changes in the fund net value that affect customer experience; estimates the fund purchase amount based on the user's own asset data to solve the technical problems in the prior art of insufficient cash flow and impact on life caused by improper allocation of investment amounts when users purchase funds.

[0070] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0071] Figure 2 A schematic diagram of a method for estimating a fund purchase amount and purchase period provided in an embodiment of the present application Figure 1 The execution subject of this embodiment may be, for example, Figure 1 The server shown, such as Figure 2 As shown, the method includes:

[0072] S201, determining a time point when a user logs into a fund purchase interface, and based on the time point, determining historical fund data of the user and historical fund data of the fund, wherein the historical fund data includes historical transaction data and historical asset data;

[0073] Among them, historical transaction data is used to indicate the corresponding historical transaction status and investment period, and the time node is used to indicate the date when the user logs into the fund purchase interface; historical fund data includes fund yield, fund fees, etc.; historical transaction data includes historical fund transaction data and historical other transaction data, among which historical fund transaction data includes historical fund purchase amount and historical fund purchase period, and each historical fund purchase amount corresponds to a historical fund purchase period; historical asset data includes account balance, annual average daily assets, periodic balance, etc.

[0074] For example, after user A logs into the mobile banking and enters the fund purchase interface, the server uses the date when user A enters the fund purchase interface as the time node, and within the time node, obtains historical fund data such as fund yield and fund fees from the fund platform; and obtains the user's historical fund transaction data and historical asset data from the bank database.

[0075] S202. Based on the historical transaction data, historical asset data and historical fund data, an estimation model is used to estimate the fund purchase amount and purchase period to obtain an estimated purchase amount and an estimated purchase period.

[0076] Among them, the estimation model is built based on a recurrent convolutional neural network.

[0077] A prediction model based on a recurrent convolutional neural network is used to process historical transaction data, historical asset data, and historical fund data. By capturing long-term dependencies in time series, the trends of transaction data, asset data, and fund data are predicted, and the fund purchase amount and purchase period are estimated to obtain the estimated purchase amount and estimated purchase period.

[0078] The method for estimating the fund purchase amount and purchase period provided in this embodiment determines the time node when the user logs into the fund purchase interface, and obtains the user's historical transaction data and historical asset data and the fund's historical fund data according to the time node, and uses the estimation model for estimation, by capturing the long-term dependency in the time series, predicting the trend of transaction data, asset data and fund data, and estimating the fund purchase amount and purchase period to obtain the estimated purchase amount and estimated purchase period, providing the user with a more accurate estimated purchase amount and estimated purchase period; the fund purchase amount and purchase period are estimated by combining the user's own historical transaction data and historical asset data, so as to avoid the user being affected by insufficient cash flow due to improper allocation of investment amount. The user can use the estimated purchase amount and estimated purchase period as a reference to purchase funds, which improves the user's investment experience when purchasing financial products, and improves user satisfaction and user stickiness.

[0079] Figure 3 A schematic diagram of a method for estimating a fund purchase amount and purchase period provided in an embodiment of the present application Figure 2 In this embodiment, the historical fund data includes fund yields corresponding to various types of funds; the estimation model includes a sequence prediction network and a decision network; in this embodiment, Figure 2 Based on the embodiment, a method for estimating the fund purchase amount and purchase period is described in detail, and the method includes:

[0080] S301, determining a time point when a user logs into a fund purchase interface, and based on the time point, determining historical fund data of the user and historical fund data of the fund, wherein the historical fund data includes historical transaction data and historical asset data;

[0081] Step S301 is similar to step S201 and will not be described again here.

[0082] S302, based on the historical transaction data, determine whether the user is a first-time purchaser; if not, execute steps S303 to S307; if yes, execute steps S308 to S309;

[0083] Among them, historical transaction data includes historical fund purchase amounts.

[0084] Based on the historical fund purchase amount in the historical transaction data, the user attributes of the user are judged. If the historical fund purchase amount of the user is not zero, it means that the user has purchased funds before and is not a first-time purchaser. Therefore, the historical experience data of the user includes the historical purchase amount difference and the historical purchase period difference, wherein the historical purchase amount difference is used to indicate the absolute difference between the historical estimated purchase amount and the historical purchase amount, and the historical purchase period difference is used to indicate the absolute difference between the historical estimated purchase period and the historical purchase period. Then, according to steps S303 to S307, the fund purchase amount and purchase period of the user are estimated to obtain the estimated fund purchase amount and purchase period;

[0085] If the user's historical fund purchase amount is zero, it indicates that the user has never purchased a fund and is a first-time purchaser. Therefore, the user's historical experience data is empty. The user's fund purchase amount and purchase period are estimated according to steps S308~S309 to obtain the estimated fund purchase amount and purchase period.

[0086] S303, performing standardization processing on the historical transaction data, historical asset data, and historical fund data according to multiple standard periods, and determining a subset of historical standard data corresponding to each standard period;

[0087] Among them, the standard period is used to indicate day, week and month, and the standardization processing includes mean processing and feature serialization processing.

[0088] Taking the day as the standard period, the historical transaction data, historical asset data and historical fund data are directly used as the historical data subset corresponding to the standard period;

[0089] Taking the week as the standard cycle, the following formula is used to average the historical transaction data, historical asset data and historical fund data:

[0090]

[0091] in, Used to indicate the kth category of data in week j; It is the historical data of each day in the unit period.

[0092] The historical transaction data, historical asset data and historical fund data after mean processing are used as the historical data subset corresponding to the standard period of one week;

[0093] Taking the month as the standard cycle, the following formula is used to average the historical transaction data, historical asset data and historical fund data:

[0094]

[0095] in, Used to indicate the k-th category of data in month m; It is the historical data of each day in the unit period.

[0096] The historical transaction data, historical asset data and historical fund data after mean processing are used as the historical data subset corresponding to the monthly standard period.

[0097] The three historical data subsets constructed according to the standard cycle are processed by feature serialization, and a two-dimensional feature vector is formed in the order of cycle time. , get the historical standard data subset corresponding to each standard period, where, is the standard cycle, is the jth type of data in the i-th type of standard period.

[0098] S304, for any one of the plurality of historical standard data subsets, normalizing the plurality of historical standard data and historical experience data in the historical standard data subset;

[0099] The following formula is used to normalize the various historical standard data and historical experience data in the historical standard data subset:

[0100]

[0101] Among them, It is the normalization function softmax.

[0102] S305, inputting the normalized multiple historical standard data and historical experience data into a sequence prediction network to determine the hidden state corresponding to each time period;

[0103] Among them, the hidden state is used to indicate the long-term dependencies in the sequence data.

[0104] Will According to the periodic time sequence, it is input into the sequence prediction network to obtain the hidden state corresponding to each period. The formula is as follows:

[0105]

[0106] in is a sequence prediction network, It is the hidden state of the last layer of the sequence prediction network at each time period.

[0107] S306, concatenating the hidden states corresponding to each time period after being processed by a fully connected layer and an average pooling layer to obtain a data prediction sequence within multiple time periods;

[0108] Among them, the fully connected layer is used to integrate the previously extracted features; the average pooling layer is used to reduce the amount of calculation and retain the overall features.

[0109] The following formula is used to concatenate the hidden states corresponding to each period after being processed by the fully connected layer and the average pooling layer:

[0110]

[0111] in, is used to indicate the data prediction sequence, is a fully connected layer, is the average pooling layer, It is a splicing connection operation.

[0112] S307, inputting the data prediction sequence into the decision network to obtain an estimated fund purchase amount and purchase period;

[0113] The data prediction sequence is used as input to the decision network to obtain the estimated fund purchase amount and purchase period. The formula is as follows:

[0114]

[0115] in, is the estimated fund purchase amount; The estimated fund purchase period; For decision-making network.

[0116] In one possible implementation, before using the estimation model to estimate the fund purchase amount and purchase period, the estimation model needs to be trained. In this embodiment, each historical standard data subset includes a first training subset and a first verification subset, and a second training subset and a second verification subset of historical experience data.

[0117] The specific process is as follows: input the data in the first training subset and the second training subset into a pre-built prediction model for training to obtain an alternative prediction model; verify the alternative prediction model according to the data in the first verification subset and the second verification subset to obtain a verification result, and the verification result is used to indicate whether the alternative prediction model converges; if the alternative prediction model does not converge, optimize the model parameters of the alternative prediction model, and train based on the optimized alternative prediction model; if the alternative prediction model converges, use the alternative estimation model as the prediction model.

[0118] The training set and validation set are determined according to a preset ratio for each historical standard data subset and historical experience data of the customer. For example, 80% of the data in each historical standard data subset can be used as the first training subset and 20% of the data can be used as the first validation subset; 80% of the data in the historical experience data can be used as the second training subset and 20% of the data can be used as the second validation subset. The data in the first training subset and the second training subset are input into the estimation model for training to obtain an alternative estimation model, and the alternative estimation model is validated with the first validation subset and the second validation subset to obtain a validation result. If the validation result shows that the estimation model does not converge, the model parameters of the alternative estimation model are optimized, the model parameters of the alternative estimation model are updated, and the training is continued with the data in the training set, and the validation is continued with the data in the validation set until the validation result shows that the estimation model converges, and the alternative estimation model is used as the estimation model.

[0119] The estimation model is trained through the user's corresponding historical transaction data, historical asset data and historical fund data, which improves the accuracy of the estimation model in estimating the estimated purchase amount and estimated purchase period for each user. The estimated purchase amount and estimated purchase period obtained have high reference value, which improves customer experience and satisfaction.

[0120] S308, taking the fund with the largest fund yield among the multiple types of funds as the target fund, and taking the purchase period of the target fund as the estimated purchase period;

[0121] Calculate the fund yields of various types of funds, compare multiple fund yields, determine the maximum fund yield, and use the fund corresponding to the maximum fund yield as the target fund, and use the purchase period of the target fund as the user's estimated purchase period.

[0122] Use the following formula to determine the estimated purchase period:

[0123]

[0124] in, To estimate the purchase period; is the fund yield index; is the fund yield in the i-th period; is the number of cycles with the day as the standard period; is the number of cycles with a standard period of one week; is the number of periods in months.

[0125] S309: Determine the estimated purchase amount based on the historical account balance, historical annual average daily assets and the fund yield of the target fund.

[0126] Among them, historical account balances, historical annual average daily assets and fund yields are used to indicate data corresponding to multiple standard periods.

[0127] Use the following formula to calculate the user's estimated purchase amount:

[0128]

[0129] in, is the estimated purchase amount, is the account balance index; is the annual average daily asset index; is the account balance of the ith period of the periodic type determined based on the purchase period; is the annual average daily assets of the ith period of the periodic type determined based on the purchase period.

[0130] In one possible implementation, after a user purchases a fund based on an estimated purchase amount and an estimated purchase period as a reference, the user's actual purchase amount and actual purchase period can also be obtained; the actual purchase amount and the estimated purchase amount, the actual purchase period and the estimated purchase period are subjected to absolute difference processing to obtain a purchase amount difference and a purchase period difference; based on the purchase amount difference and the purchase period difference, the historical experience data is updated, and the estimation model is optimized based on the updated historical experience data.

[0131] The absolute difference between the actual purchase amount and the estimated purchase amount is processed using the following formula:

[0132]

[0133] in, The difference in purchase amount; is the actual purchase amount; This is the estimated purchase amount.

[0134] The following formula is used to calculate the absolute difference between the actual purchase period and the estimated purchase period:

[0135]

[0136] in, Find for purchase period; The actual purchase period; Estimated purchase period.

[0137] The user uses the estimated purchase amount and the estimated purchase period as a reference. After purchasing the fund, the user's actual purchase amount and actual purchase period are obtained. The purchase amount is obtained by performing absolute difference processing on the actual purchase amount and the estimated purchase amount. The actual purchase period and the estimated purchase period are processed by performing absolute difference processing to obtain the purchase period difference. The user's historical experience data is updated through the purchase amount difference and the purchase period difference. The updated historical experience data is used as the input variable of the estimation model, and the estimation model is optimized to improve the accuracy and timeliness of the model.

[0138] The method for estimating the fund purchase amount and purchase period provided in the embodiment of the present application obtains the user's historical transaction data and historical asset data and the fund's historical fund data according to the time node when the user logs into the fund purchase interface. For users who purchase funds for the first time, the user's historical account balance, historical annual average daily assets and fund yield are comprehensively considered to estimate the purchase amount and purchase period of the fund purchased by the user; for users who are not purchasing funds for the first time, the estimated purchase amount and the estimated purchase period are obtained by better processing the relationship between sequence data through the self-attention mechanism, and then using the estimation model for estimation, thereby improving the accuracy of the estimated purchase amount and period of the fund; the user can use the estimated purchase amount and the estimated purchase period as a reference to purchase funds, thereby improving the user's investment experience when purchasing financial products, and improving user satisfaction and user stickiness.

[0139] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0140] It should be further noted that, although the various steps in the flowchart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0141] Figure 4 A schematic diagram of a device for estimating a fund purchase amount and purchase period provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, the fund purchase amount and purchase period estimation device 40 provided in this embodiment includes:

[0142] Determination module 401, used to determine the time node when the user logs into the fund purchase interface, and based on the time node, determine the user's historical fund data and the fund's historical fund data, the historical fund data includes historical transaction data and historical asset data, and the historical transaction data is used to indicate the corresponding historical transaction situation and investment period;

[0143] The estimation module 402 is used to estimate the fund purchase amount and purchase period based on historical transaction data, historical asset data and historical fund data using an estimation model to obtain an estimated purchase amount and an estimated purchase period.

[0144] In a possible implementation manner, the device further includes: a determination module 403;

[0145] The determination module 403 is used to determine whether the user is a first-time purchaser based on historical transaction data;

[0146] In the case where the user is not a first-time purchaser, the estimation module 402 is further used to estimate the fund purchase amount and purchase period using an estimation model based on historical transaction data, historical asset data, historical fund data, and historical experience data to obtain an estimated purchase amount and an estimated purchase period;

[0147] Among them, the historical experience data includes the historical purchase amount difference and the historical purchase period difference. The historical purchase amount difference is used to indicate the absolute difference between the historical estimated purchase amount and the historical purchase amount, and the historical purchase period difference is used to indicate the absolute difference between the historical estimated purchase period and the historical purchase period.

[0148] In a possible implementation, the historical fund data includes fund yields corresponding to various types of funds, and the historical asset data includes: historical account balances, historical annual average daily assets,

[0149] In the case where the user is a first-time buyer, the estimation module 402 is also used to select the fund with the largest fund yield among multiple types of funds as the target fund, and use the purchase period of the target fund as the estimated purchase period; determine the estimated purchase amount based on the historical account balance, the historical annual average daily assets and the fund yield of the target fund, wherein the historical account balance, the historical annual average daily assets and the fund yield are used to indicate the data corresponding to multiple types of standard periods.

[0150] In one possible implementation, the estimation model includes a sequence prediction network and a decision network. The estimation module 402 is also used to standardize historical transaction data, historical asset data, and historical fund data according to multiple standard periods, and determine the historical standard data subset corresponding to each standard period; for any one of the multiple historical standard data subsets, multiple historical standard data and historical experience data in the historical standard data subset are input into the sequence prediction network to obtain data prediction sequences within multiple time periods; the data prediction sequence is input into the decision network to obtain an estimated fund purchase amount and purchase period.

[0151] In one possible implementation, the estimation module 402 is also used to normalize the various historical standard data and historical experience data in the historical standard data subset; input the normalized various historical standard data and historical experience data into the sequence prediction network to determine the hidden state corresponding to each time period, and the hidden state is used to indicate the long-term dependency relationship in the sequence data; the hidden state corresponding to each time period is processed by the fully connected layer and the average pooling layer and then spliced ​​to obtain the data prediction sequence in multiple time periods.

[0152] In a possible implementation, each historical standard data subset includes a first training subset and a first validation subset, and a historical experience data second training subset and a second validation subset, and the apparatus further includes: a training module 404;

[0153] The training module 404 is used to input the data in the first training subset and the second training subset into a pre-built estimation model for training to obtain an alternative estimation model; verify the alternative estimation model according to the data in the first verification subset and the second verification subset to obtain a verification result, and the verification result is used to indicate whether the alternative estimation model converges; if the alternative estimation model does not converge, optimize the model parameters of the alternative estimation model and train based on the optimized alternative estimation model; if the alternative estimation model converges, use the alternative estimation model as the estimation model.

[0154] In a possible implementation, the device further includes: an updating module 405;

[0155] The update module 405 is used to obtain the user's actual purchase amount and actual purchase period; perform absolute difference processing on the actual purchase amount and estimated purchase amount, and the actual purchase period and estimated purchase period, respectively, to obtain the purchase amount difference and the purchase period difference; based on the purchase amount difference and the purchase period difference, update the historical experience data, and optimize the estimation model based on the updated historical experience data.

[0156] The device for estimating the fund purchase amount and purchase period provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and this embodiment will not be described in detail here.

[0157] It should be understood that the above-mentioned device embodiments are only illustrative, and the device of the present application can also be implemented in other ways. For example, the division of units / modules in the above-mentioned embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0158] In addition, unless otherwise specified, each functional unit / module in each embodiment of the present application may be integrated into one unit / module, each unit / module may exist physically separately, or two or more units / modules may be integrated together. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of a software program module.

[0159] Figure 5 This is a schematic diagram of a device for estimating fund purchase amount and purchase period provided in an embodiment of the present application. Figure 5 As shown, the electronic device 50 provided in this embodiment includes: at least one processor 501 and a memory 502. Optionally, the device 50 also includes a communication component 503. The processor 501, the memory 502 and the communication component 503 are connected via a bus 504.

[0160] In a specific implementation process, at least one processor 501 executes the computer-executable instructions stored in the memory 502, so that at least one processor 501 executes the above method.

[0161] The specific implementation process of the processor 501 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.

[0162] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0163] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0164] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0165] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0166] In the above embodiments, the description of each embodiment has its own emphasis. For the part not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0167] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0168] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for estimating fund purchase amount and purchase period, characterized in that: include: Determine the time point when the user logs into the fund purchase interface, and based on the time point, determine the historical fund data of the user and the historical fund data of the fund, the historical fund data includes historical transaction data and historical asset data, and the historical transaction data is used to indicate the corresponding historical transaction situation and investment period; Based on the historical transaction data, historical asset data and historical fund data, an estimation model is used to estimate the fund purchase amount and purchase period to obtain an estimated purchase amount and an estimated purchase period. The estimation model is constructed based on a recurrent convolutional neural network.

2. The method according to claim 1, characterized in that Based on the historical transaction data, historical asset data and historical fund data, the fund purchase amount and purchase period are estimated using an estimation model to obtain an estimated purchase amount and an estimated purchase period, including: Based on the historical transaction data, determining whether the user is a first-time purchasing user; If the user is not a first-time purchaser, based on the historical transaction data, historical asset data, historical fund data and historical experience data, an estimation model is used to estimate the fund purchase amount and purchase period to obtain an estimated purchase amount and an estimated purchase period; Among them, the historical experience data includes historical purchase amount difference and historical purchase period difference, the historical purchase amount difference is used to indicate the absolute difference between the historical estimated purchase amount and the historical purchase amount, and the historical purchase period difference is used to indicate the absolute difference between the historical estimated purchase period and the historical purchase period.

3. The method according to claim 2, characterized in that The historical fund data includes fund yields corresponding to various types of funds, the historical asset data includes: historical account balances, historical annual average daily assets, and the method further includes: In the case where the user is a first-time purchaser, the fund with the largest fund yield among multiple types of funds is used as the target fund, and the purchase period of the target fund is used as the estimated purchase period; The estimated purchase amount is determined based on the historical account balance, the historical annual average daily assets and the fund yield of the target fund, wherein the historical account balance, the historical annual average daily assets and the fund yield are used to indicate data corresponding to multiple types of standard periods.

4. The method according to claim 2, characterized in that: The estimation model includes a sequence prediction network and a decision network. Based on the historical transaction data, historical asset data, historical fund data and historical experience data, the estimation model is used to estimate the fund purchase amount and purchase period to obtain the estimated purchase amount and the estimated purchase period, including: Standardize the historical transaction data, historical asset data, and historical fund data according to multiple standard periods, and determine a subset of historical standard data corresponding to each standard period; For any one of the multiple historical standard data subsets, multiple historical standard data and historical experience data in the historical standard data subset are input into the sequence prediction network to obtain data prediction sequences in multiple time periods; The data prediction sequence is input into the decision network to obtain the estimated fund purchase amount and purchase period.

5. The method according to claim 4, characterized in that The multiple historical standard data and historical experience data in the historical standard data subset are input into the sequence prediction network to obtain data prediction sequences in multiple time periods, including: Normalizing the multiple historical standard data and historical experience data in the historical standard data subset; Inputting the normalized multiple historical standard data and historical experience data into a sequence prediction network to determine a hidden state corresponding to each time period, wherein the hidden state is used to indicate a long-term dependency relationship in the sequence data; The hidden states corresponding to each time period are processed by the fully connected layer and the average pooling layer and then spliced ​​to obtain the data prediction sequence in multiple time periods.

6. The method according to claim 2, characterized in that Each historical standard data subset includes a first training subset and a first validation subset, the historical experience data second training subset and a second validation subset, and the method further includes: estimating the fund purchase amount and purchase period based on the historical transaction data, historical asset data, historical fund data and historical experience data by using an estimation model to obtain the estimated purchase amount and the estimated purchase period. Inputting the data in the first training subset and the second training subset into a pre-built estimation model for training to obtain an alternative estimation model; Verifying the candidate estimation model according to the data in the first validation subset and the second validation subset to obtain a verification result, wherein the verification result is used to indicate whether the candidate estimation model converges; If the alternative estimation model does not converge, optimizing the model parameters of the alternative estimation model, and performing training based on the optimized alternative estimation model; If the candidate estimation model converges, the candidate estimation model is used as the estimation model.

7. The method according to claim 2, characterized in that The method further comprises: Obtaining the actual purchase amount and actual purchase period of the user; Performing absolute difference processing on the actual purchase amount and the estimated purchase amount, the actual purchase period and the estimated purchase period, respectively, to obtain a purchase amount difference and a purchase period difference; Based on the purchase amount difference and the purchase period difference, the historical experience data is updated, and based on the updated historical experience data, the estimation model is optimized.

8. A device for estimating fund purchase amount and purchase period, characterized in that: include: A determination module, used to determine the time node when the user logs into the fund purchase interface, and based on the time node, determine the historical fund data of the user and the historical fund data of the fund, wherein the historical fund data includes historical transaction data and historical asset data, and the historical transaction data is used to indicate the corresponding historical transaction situation and investment period; The estimation module is used to estimate the fund purchase amount and purchase period based on the historical transaction data, historical asset data and historical fund data using an estimation model to obtain an estimated purchase amount and an estimated purchase period.

9. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

11. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 7 when being executed by a processor.

Images