Transaction quantity prediction method and device, computer equipment and storage medium
By decomposing the sequence generated by transaction data, extracting attribute characteristics and change trend characteristics, and combining with prediction models, the problem of low transaction volume prediction accuracy in the prior art is solved, and higher prediction accuracy is achieved.
Patent Information
- Application Number
- CN202411507057.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-28
- Publication Date
- 2025-02-18
AI Technical Summary
The prior art only relies on the numerical values of transaction data in transaction volume prediction, ignores changes in time characteristics, resulting in a decrease in prediction accuracy.
By obtaining the transaction data of the target business, generating a transaction sequence, decomposing it, extracting transaction attribute characteristics and change trend characteristics, and predicting transaction volume with the prediction model.
The accuracy of transaction volume prediction is improved, and the effectiveness of the prediction model is enhanced by considering time characteristics and transaction characteristics.
Smart Images
Figure CN119477504A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a transaction volume prediction method, apparatus, computer equipment and storage medium, which can be applied to the financial field or other technical fields. Background Art
[0002] With the development of the financial sector, the number of various financial businesses has also increased. In order to ensure the stability of transaction processing under financial business, transaction volume prediction methods for financial business have emerged. The transaction volume of future periods can be predicted based on transaction data of historical periods.
[0003] However, since changes in time characteristics will directly affect transaction volume, using the current transaction volume prediction method to predict transaction volume only based on the numerical value of transaction data will reduce the accuracy of transaction volume prediction. Summary of the invention
[0004] Based on this, it is necessary to provide a transaction volume prediction method, device, computer equipment and storage medium that can improve the accuracy of transaction volume prediction in response to the above technical problems.
[0005] In a first aspect, the present application provides a transaction volume prediction method. The method comprises:
[0006] Obtain target transaction data of target business;
[0007] Based on the generation time of the target transaction data, a target transaction sequence is generated according to the target transaction data;
[0008] Decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0009] Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0010] In one embodiment, the target transaction sequence is decomposed to obtain target transaction attribute characteristics and target change trend characteristics, including:
[0011] The target transaction sequence is decomposed by the empirical mode decomposition method to obtain multiple candidate transaction attribute features and target change trend features; wherein the candidate transaction attribute features include a time feature component and a transaction feature component; according to the target transaction sequence and the candidate transaction attribute features, the target transaction attribute features are selected from the candidate transaction attribute features.
[0012] In one embodiment, selecting a target transaction attribute feature from the candidate transaction attribute features according to the target transaction sequence and the candidate transaction attribute features includes:
[0013] According to the association between each candidate transaction attribute feature and the target transaction sequence, the correlation coefficient value corresponding to each candidate transaction attribute feature is determined; according to the correlation coefficient value corresponding to each candidate transaction attribute feature, the target transaction attribute feature is selected from the candidate transaction attribute features.
[0014] In one embodiment, based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics, including:
[0015] According to the generation time of the target transaction data, a target prediction model is selected from the candidate prediction models corresponding to different time categories; through the target prediction model, the predicted transaction volume of the target business is determined according to the target transaction attribute characteristics and the target change trend characteristics.
[0016] In one embodiment, the predicted transaction volume of the target business is determined by the target prediction model according to the target transaction attribute characteristics and the target change trend characteristics, including:
[0017] Through the target prediction model, the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics are determined; according to the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics, the predicted transaction volume of the target business is determined.
[0018] In one embodiment, the candidate prediction model corresponding to any time category is trained in the following manner:
[0019] Obtain sample transaction data under the time category; input sample transaction attribute features and sample change trend features corresponding to the sample transaction data into the initial prediction model to obtain prediction results of the sample transaction data; wherein the initial prediction model includes a trend change sub-model, a cycle change sub-model and a holiday impact sub-model; determine the loss value according to the actual results and the predicted results of the sample transaction data; use the loss value to adjust the network parameters of the initial prediction model to obtain a candidate prediction model under the time category.
[0020] In a second aspect, the present application also provides a transaction volume prediction device. The device comprises:
[0021] A data acquisition module is used to acquire target transaction data of a target business;
[0022] A sequence generation module, used for generating a target transaction sequence according to the target transaction data based on the generation time of the target transaction data;
[0023] The sequence decomposition module is used to decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0024] The prediction module is used to determine the predicted transaction volume of the target business based on the target prediction model, according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0025] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:
[0026] Obtain target transaction data of target business;
[0027] Based on the generation time of the target transaction data, a target transaction sequence is generated according to the target transaction data;
[0028] Decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0029] Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0030] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:
[0031] Obtain target transaction data of target business;
[0032] Based on the generation time of the target transaction data, a target transaction sequence is generated according to the target transaction data;
[0033] Decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0034] Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0035] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:
[0036] Obtain target transaction data of target business;
[0037] Based on the generation time of the target transaction data, a target transaction sequence is generated according to the target transaction data;
[0038] Decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0039] Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0040] The transaction volume prediction method, device, computer equipment and storage medium obtain the target transaction data of the target business, generate the target transaction sequence, and decompose the target transaction sequence to obtain the target transaction attribute characteristics and the target change trend characteristics; then, based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics. Compared with the related art, which only predicts the transaction volume based on the numerical value of the transaction data, the above method decomposes the target transaction sequence into the target change trend characteristics and the target transaction attribute characteristics containing the time characteristics and the transaction characteristics, so that the transaction volume can be predicted in combination with the time characteristics of the target transaction data, thereby ensuring the accuracy of the transaction volume prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 A diagram showing an application environment of a transaction volume prediction method in an embodiment;
[0042] Figure 2 A schematic diagram of a process flow of a transaction volume prediction method in one embodiment;
[0043] Figure 3 A schematic diagram of a process for determining a target transaction attribute feature in one embodiment;
[0044] Figure 4 A schematic diagram of a process for determining a predicted transaction volume of a target business in one embodiment;
[0045] Figure 5 A schematic diagram of a flow chart of a transaction volume prediction method in another embodiment;
[0046] Figure 6 is a structural block diagram of a transaction volume prediction device in one embodiment;
[0047] Figure 7 is a structural block diagram of a transaction volume prediction device in another embodiment;
[0048] Figure 8 is a structural block diagram of a transaction volume prediction device in yet another embodiment;
[0049] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0050] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0051] The transaction volume prediction method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. For example, the transaction data of each business, etc. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. For example, the server 104 obtains the target transaction data of the target business, generates a target transaction sequence based on the target transaction data based on the generation time of the target transaction data, decomposes the target transaction sequence, obtains the target transaction attribute characteristics and the target change trend characteristics, and determines the predicted transaction volume of the target business based on the target prediction model according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics; further, the server 104 can send the predicted transaction volume of the target business to the terminal 102 for display. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and Internet of Things devices. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.
[0052] With the development of the financial sector, the number of various financial businesses has also increased. In order to ensure the stability of transaction processing under financial business, transaction volume prediction methods for financial business have emerged. The transaction volume of future periods can be predicted based on transaction data of historical periods.
[0053] However, since changes in time characteristics will directly affect transaction volume, using current transaction volume prediction methods to make predictions based only on the numerical values of transaction data will reduce the accuracy of transaction volume predictions.
[0054] Based on this, in one embodiment, Figure 2 As shown, a transaction volume prediction method is provided, which is applied to Figure 1 The server in the example is used as an example to illustrate the specific steps:
[0055] S201, obtaining target transaction data of a target business.
[0056] The target business refers to the financial business provided by the institution; the target transaction data refers to the business-related data under the target business, for example, it can be the transaction data of the target business.
[0057] Optionally, the operation and maintenance personnel can select the target business to be predicted through the page on the terminal and click Send. At this time, the terminal can generate a prediction request containing the target business and send it to the server; then, after receiving the prediction request, the server can obtain the target transaction data within the historical period based on the target business carried in the prediction request.
[0058] S202, generating a target transaction sequence according to the target transaction data based on the generation time of the target transaction data.
[0059] The target transaction sequence refers to a time series sequence generated based on the target transaction data.
[0060] Optionally, after acquiring the target transaction data, the target transaction data may be sorted from front to back according to the generation time of the target transaction data; then, the target transaction data may be processed according to the sorting result to generate a target transaction sequence.
[0061] S203, decomposing the target transaction sequence to obtain target transaction attribute characteristics and target change trend characteristics.
[0062] The target transaction attribute feature refers to the feature used to represent the transaction attribute of the target transaction sequence; the target change trend feature refers to the feature used to represent the change trend of the target transaction sequence.
[0063] Optionally, after determining the target transaction sequence, the target transaction sequence can be decomposed to obtain one or more target transaction attribute features and a target change trend feature. It is understandable that in order to be able to predict the transaction volume of the target business in the next few periods in combination with the time feature, each target transaction attribute feature needs to include data under the time dimension and data under the transaction dimension.
[0064] S204, based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0065] Among them, the target prediction model refers to a trained neural network model that can predict the transaction volume of the target business based on the target transaction data.
[0066] Optionally, since the transaction volume of the target business in different time categories changes periodically, the candidate prediction models corresponding to different time categories can be generated according to the periodic changes in the transaction volume of the target business in each time category; then, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the candidate prediction models corresponding to different time categories, the target transaction attribute characteristics and the target change trend characteristics. The candidate prediction model refers to any trained prediction model.
[0067] Exemplarily, a target prediction model can be selected from candidate prediction models corresponding to different time categories according to the generation time of the target transaction data, and the predicted transaction volume of the target business can be determined through the target prediction model according to the target transaction attribute characteristics and the target change trend characteristics.
[0068] Optionally, the time category corresponding to the target data can be determined based on the generation time of the target transaction data; then, based on the time category corresponding to the target data, the target prediction model is selected from the candidate prediction models corresponding to different time categories; finally, the target transaction attribute characteristics and the target change trend characteristics are input into the target prediction model, and the target prediction model determines the predicted transaction volume of the target business based on the target transaction attribute characteristics, the target change trend characteristics and the model parameters.
[0069] In the above transaction volume prediction method, the target transaction data of the target business is obtained to generate a target transaction sequence, and the target transaction sequence is decomposed to obtain the target transaction attribute characteristics and the target change trend characteristics; then, based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics. Compared with the related art, in which the transaction volume is predicted only based on the numerical value of the transaction data, the above method can predict the transaction volume by combining the time characteristics of the target transaction data by decomposing the target transaction sequence into the target change trend characteristics and the target transaction attribute characteristics containing the time characteristics and the transaction characteristics, thereby ensuring the accuracy of the transaction volume prediction.
[0070] In order to ensure the accuracy of the target transaction attribute characteristics, based on the above embodiment, in this embodiment, an optional method for determining the target transaction attribute characteristics is provided, such as Figure 3 As shown, the specific steps include:
[0071] S301, using the empirical mode decomposition method to decompose the target transaction sequence to obtain multiple candidate transaction attribute features and target change trend features.
[0072] The candidate transaction attribute feature refers to any transaction attribute feature obtained by decomposing the target transaction sequence, and may include a time feature component and a transaction feature component.
[0073] Optionally, the target transaction sequence may be decomposed by using an empirical mode decomposition (EMD) method to obtain a target change trend feature and a plurality of candidate transaction attribute features including a time feature component and a transaction feature component.
[0074] The essence of the empirical mode decomposition method is to stabilize nonlinear time series data. The basic principle can be referred to in formula (1), which is to decompose a high-frequency time component series signal into several signal components and a trend term, and then arrange them in order from high frequency to low frequency at any time. Among them, Y t Represents the high-frequency time component sequence signal, that is, the target transaction sequence; IMF i (t) represents the i-th signal component, i.e., the i-th candidate transaction attribute feature; R n (t) represents the trend term, that is, the target change trend characteristic.
[0075] (1)
[0076] First, according to the attribute characteristics of candidate transactions, the target transaction sequence can be converted into a signal graph, and the local maximum points of the target transaction sequence can be connected using the cubic spline interpolation method to form an upper envelope u 0 (t); Use the cubic spline interpolation method to connect the local minimum points of the target transaction sequence to form the lower envelope l 0 (t); then, refer to formula (2) to find u 0 (t) and l 0 The mean value m of (t) 0 (t).
[0077] (2)
[0078] Furthermore, the target transaction sequence minus the mean m is used. 0 (t) Determine the component p 1 (t), and check to obtain the component p 1 (t) Whether the component condition is met, if the component p 1 (t) satisfies the component condition, then the first component is p 1 (t); if the component p 1 (t) does not meet the component condition, repeat the above operation until the component condition is met or the stopping criterion is reached, where the stopping criterion SD is shown in formula (3).
[0079] (3)
[0080] According to the target change trend characteristics, the remaining components after the above processing can be converted into a new sequence to be decomposed, and the above three steps are repeated again until the component or the remaining component is less than a specific value or the remaining component is converted into a monotonic function.
[0081] S302: Select a target transaction attribute feature from the candidate transaction attribute features according to the target transaction sequence and the candidate transaction attribute features.
[0082] Optionally, after the candidate transaction attribute features are determined, the target transaction attribute features may be selected from the candidate transaction attribute features according to the association between each candidate transaction attribute feature and the target transaction sequence.
[0083] Exemplarily, the correlation coefficient value corresponding to each candidate transaction attribute feature can be determined according to the association between each candidate transaction attribute feature and the target transaction sequence, and the target transaction attribute feature can be selected from the candidate transaction attribute features according to the correlation coefficient value corresponding to each candidate transaction attribute feature.
[0084] Optionally, a correlation coefficient method may be used to determine the correlation coefficient value corresponding to each candidate transaction attribute feature according to the association between each candidate transaction attribute feature and the target transaction sequence; then, the candidate transaction attribute features are sorted in descending order according to the correlation coefficient value corresponding to each candidate transaction attribute feature, and a preset number of candidate transaction attribute features ranked first are used as target transaction attribute features.
[0085] In this embodiment, candidate transaction attribute features are introduced, and the target transaction attribute features are determined from the candidate transaction attribute features, so that the accuracy of determining the target transaction attribute features can be guaranteed.
[0086] In order to ensure the accuracy of the predicted transaction volume, based on the above implementation, in this embodiment, an optional method for determining the predicted transaction volume of the target business is provided, such as Figure 4 As shown, the specific steps include:
[0087] S401, determining the component prediction value corresponding to the target transaction attribute feature and the trend prediction value corresponding to the target change trend feature through the target prediction model.
[0088] Among them, the component prediction value refers to the numerical value determined according to the target transaction attribute characteristics and used to predict the transaction volume; the trend prediction value refers to the numerical value determined according to the target change trend characteristics and used to represent the transaction volume trend.
[0089] Optionally, after the target transaction attribute characteristics and the target change trend characteristics are input into the target prediction model, the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics can be determined by the target prediction model.
[0090] S402, determining the predicted transaction volume of the target business according to the component prediction value corresponding to the target transaction attribute characteristic and the trend prediction value corresponding to the target change trend characteristic.
[0091] Optionally, the component prediction values corresponding to the target transaction attribute characteristics and the trend prediction values corresponding to the target change trend characteristics may be aggregated to determine the predicted transaction volume of the target business.
[0092] In this embodiment, component prediction values and trend prediction values are introduced, and the predicted transaction volume of the target business is determined by the component prediction values and trend prediction values, so as to ensure the accuracy of the predicted transaction volume.
[0093] In order to ensure the accuracy of the candidate prediction model, on the basis of the above implementation, in this embodiment, a training method for the candidate prediction model is provided, specifically, obtaining sample transaction data under the time category, inputting the sample transaction attribute characteristics and sample change trend characteristics corresponding to the sample transaction data into the initial prediction model, obtaining the prediction result of the sample transaction data, determining the loss value according to the actual result and the predicted result of the sample transaction data, and then using the loss value to adjust the network parameters of the initial prediction model to obtain the candidate prediction model under the time category.
[0094] Optionally, the initial prediction model may be constructed as a model including a trend change sub-model, a cycle change sub-model and a holiday impact sub-model with reference to formula (4).
[0095] y(t)=g(t)+s(t)+h(t)+ε t (4)
[0096] Among them, y(t) represents the initial prediction model; g(t) represents the trend term used to fit random changes, that is, the trend change sub-model; s(t) represents the periodic change of time series, that is, the periodic change sub-model; h(t) represents the irregular holiday impact, that is, the holiday impact sub-model; ε t Represents all error factors not considered by the model.
[0097] Furthermore, the models for the trend change sub-model, cycle change sub-model and holiday impact sub-model can be constructed by referring to formulas (5)-(7). Where C represents the maximum asymptotic value of the model; k represents the growth rate of the model; t represents time; m represents the offset of the trend change sub-model; T is the cycle length; N is the number of cycles; a n 、b n is the estimated parameter; n is half of the number of cycles used in the periodic variation submodel; Z(t) is the indicator function, K i Indicates the parameters under each time category.
[0098] (5)
[0099] (6)
[0100] (7)
[0101] Optionally, after constructing the initial prediction model, for each time category, sample transaction data under the time category can be obtained; then, the sample transaction attribute characteristics and sample change trend characteristics corresponding to the sample transaction data are input into the initial prediction model to obtain the prediction results of the sample transaction data.
[0102] Furthermore, a loss function can be used to determine the loss value based on the actual results and predicted results of the sample transaction data, and the loss value can be used to adjust the network parameters of the initial prediction model until the accuracy of the initial prediction model meets the model training conditions. The initial prediction model can then be used as a candidate prediction model under this time category.
[0103] In this embodiment, a training method for a candidate prediction model is provided, which can ensure the accuracy of the candidate prediction model.
[0104] Figure 5 FIG. 2 is a flow chart of a transaction volume prediction method in another embodiment. Based on the above embodiment, this embodiment provides an optional example of a transaction volume prediction method. Figure 5 The specific implementation process is as follows:
[0105] S501, obtaining target transaction data of a target business.
[0106] S502, generating a target transaction sequence according to the target transaction data based on the generation time of the target transaction data.
[0107] S503, using the empirical mode decomposition method to decompose the target transaction sequence to obtain multiple candidate transaction attribute features and target change trend features.
[0108] The candidate transaction attribute features include a time feature component and a transaction feature component.
[0109] S504: Determine the correlation coefficient value corresponding to each candidate transaction attribute feature according to the association between each candidate transaction attribute feature and the target transaction sequence.
[0110] S505 , selecting a target transaction attribute feature from the candidate transaction attribute features according to the correlation coefficient value corresponding to each candidate transaction attribute feature.
[0111] S506: Select a target prediction model from candidate prediction models corresponding to different time categories according to the generation time of the target transaction data.
[0112] S507, determining the component prediction value corresponding to the target transaction attribute feature and the trend prediction value corresponding to the target change trend feature through the target prediction model.
[0113] S508, determining the predicted transaction volume of the target business according to the component prediction value corresponding to the target transaction attribute characteristic and the trend prediction value corresponding to the target change trend characteristic.
[0114] Optionally, obtain sample transaction data under the time category; input sample transaction attribute features and sample change trend features corresponding to the sample transaction data into the initial prediction model to obtain prediction results of the sample transaction data; wherein the initial prediction model includes a trend change sub-model, a cycle change sub-model and a holiday impact sub-model; determine the loss value based on the actual results and the predicted results of the sample transaction data; use the loss value to adjust the network parameters of the initial prediction model to obtain a candidate prediction model under the time category.
[0115] The specific process of the above S501-S508 can refer to the description of the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.
[0116] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these 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 part of the steps or stages in other steps.
[0117] Based on the same inventive concept, the embodiment of the present application also provides a transaction volume prediction device for implementing the transaction volume prediction method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more transaction volume prediction device embodiments provided below can refer to the limitations of the transaction volume prediction method above, and will not be repeated here.
[0118] In one embodiment, Figure 6 As shown, a transaction volume prediction device 1 is provided, comprising: a data acquisition module 10, a sequence generation module 20, a sequence decomposition module 30 and a prediction module 40, wherein:
[0119] A data acquisition module 10 is used to acquire target transaction data of a target business;
[0120] A sequence generation module 20, for generating a target transaction sequence according to the target transaction data based on the generation time of the target transaction data;
[0121] A sequence decomposition module 30 is used to decompose the target transaction sequence to obtain target transaction attribute characteristics and target change trend characteristics;
[0122] The prediction module 40 is used to determine the predicted transaction volume of the target business based on the target prediction model, according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0123] In one embodiment, Figure 7 As shown, the sequence decomposition module 30 includes:
[0124] The sequence decomposition unit 31 is used to decompose the target transaction sequence by using the empirical mode decomposition method to obtain a plurality of candidate transaction attribute features and target change trend features; wherein the candidate transaction attribute features include a time feature component and a transaction feature component;
[0125] The determination unit 32 is configured to select a target transaction attribute feature from the candidate transaction attribute features according to the target transaction sequence and the candidate transaction attribute features.
[0126] In one embodiment, the determining unit 32 is specifically configured to:
[0127] According to the association between each candidate transaction attribute feature and the target transaction sequence, the correlation coefficient value corresponding to each candidate transaction attribute feature is determined; according to the correlation coefficient value corresponding to each candidate transaction attribute feature, the target transaction attribute feature is selected from the candidate transaction attribute features.
[0128] In one embodiment, Figure 8 As shown, the prediction module 40 includes:
[0129] A model determination unit 41 is used to select a target prediction model from candidate prediction models corresponding to different time categories according to the generation time of the target transaction data;
[0130] The prediction unit 42 is used to determine the predicted transaction volume of the target business according to the target transaction attribute characteristics and the target change trend characteristics through the target prediction model.
[0131] In one embodiment, the prediction unit 42 is specifically configured to:
[0132] Through the target prediction model, the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics are determined; according to the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics, the predicted transaction volume of the target business is determined.
[0133] In one embodiment, the transaction volume prediction device 1 further includes a model training module, wherein the model training module is specifically used for:
[0134] Obtain sample transaction data under the time category; input sample transaction attribute features and sample change trend features corresponding to the sample transaction data into the initial prediction model to obtain prediction results of the sample transaction data; wherein the initial prediction model includes a trend change sub-model, a cycle change sub-model and a holiday impact sub-model; determine the loss value according to the actual results and the predicted results of the sample transaction data; use the loss value to adjust the network parameters of the initial prediction model to obtain a candidate prediction model under the time category.
[0135] Each module in the transaction volume prediction device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.
[0136] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig. 9 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store transaction data of each business. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a transaction volume prediction method is implemented.
[0137] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0138] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:
[0139] Obtain target transaction data of target business;
[0140] Based on the generation time of the target transaction data, a target transaction sequence is generated according to the target transaction data;
[0141] Decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0142] Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0143] In one embodiment, when the processor executes the logic of decomposing the target transaction sequence in the computer program to obtain the target transaction attribute characteristics and the target change trend characteristics, the following steps are specifically implemented:
[0144] The target transaction sequence is decomposed by the empirical mode decomposition method to obtain multiple candidate transaction attribute features and target change trend features; wherein the candidate transaction attribute features include a time feature component and a transaction feature component; according to the target transaction sequence and the candidate transaction attribute features, the target transaction attribute features are selected from the candidate transaction attribute features.
[0145] In one embodiment, when the processor executes the logic of selecting the target transaction attribute feature from the candidate transaction attribute features according to the target transaction sequence and the candidate transaction attribute features in the computer program, the following steps are specifically implemented:
[0146] According to the association between each candidate transaction attribute feature and the target transaction sequence, the correlation coefficient value corresponding to each candidate transaction attribute feature is determined; according to the correlation coefficient value corresponding to each candidate transaction attribute feature, the target transaction attribute feature is selected from the candidate transaction attribute features.
[0147] In one embodiment, when the processor executes the logic of determining the predicted transaction volume of the target business based on the target prediction model according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics in the computer program, the following steps are specifically implemented:
[0148] According to the generation time of the target transaction data, a target prediction model is selected from the candidate prediction models corresponding to different time categories; through the target prediction model, the predicted transaction volume of the target business is determined according to the target transaction attribute characteristics and the target change trend characteristics.
[0149] In one embodiment, when the processor executes the logic of determining the predicted transaction volume of the target business according to the target transaction attribute characteristics and the target change trend characteristics through the target prediction model in the computer program, the following steps are specifically implemented:
[0150] Through the target prediction model, the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics are determined; according to the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics, the predicted transaction volume of the target business is determined.
[0151] In one embodiment, the candidate prediction model corresponding to any time category is trained in the following manner:
[0152] Obtain sample transaction data under the time category; input sample transaction attribute features and sample change trend features corresponding to the sample transaction data into the initial prediction model to obtain prediction results of the sample transaction data; wherein the initial prediction model includes a trend change sub-model, a cycle change sub-model and a holiday impact sub-model; determine the loss value according to the actual results and the predicted results of the sample transaction data; use the loss value to adjust the network parameters of the initial prediction model to obtain a candidate prediction model under the time category.
[0153] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
[0154] Obtain target transaction data of target business;
[0155] Based on the generation time of the target transaction data, a target transaction sequence is generated according to the target transaction data;
[0156] Decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0157] Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0158] In one embodiment, when the code logic of decomposing the target transaction sequence to obtain the target transaction attribute characteristics and the target change trend characteristics in the computer program is executed by the processor, the following steps are specifically implemented:
[0159] The target transaction sequence is decomposed by the empirical mode decomposition method to obtain multiple candidate transaction attribute features and target change trend features; wherein the candidate transaction attribute features include a time feature component and a transaction feature component; according to the target transaction sequence and the candidate transaction attribute features, the target transaction attribute features are selected from the candidate transaction attribute features.
[0160] In one embodiment, when the code logic of selecting the target transaction attribute feature from the candidate transaction attribute features according to the target transaction sequence and the candidate transaction attribute features in the computer program is executed by the processor, the following steps are specifically implemented:
[0161] According to the association between each candidate transaction attribute feature and the target transaction sequence, the correlation coefficient value corresponding to each candidate transaction attribute feature is determined; according to the correlation coefficient value corresponding to each candidate transaction attribute feature, the target transaction attribute feature is selected from the candidate transaction attribute features.
[0162] In one embodiment, when the code logic of determining the predicted transaction volume of the target business based on the target prediction model according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics in the computer program is executed by the processor, the following steps are specifically implemented:
[0163] According to the generation time of the target transaction data, a target prediction model is selected from the candidate prediction models corresponding to different time categories; through the target prediction model, the predicted transaction volume of the target business is determined according to the target transaction attribute characteristics and the target change trend characteristics.
[0164] In one embodiment, when the code logic of determining the predicted transaction volume of the target business through the target prediction model according to the target transaction attribute characteristics and the target change trend characteristics in the computer program is executed by the processor, the following steps are specifically implemented:
[0165] Through the target prediction model, the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics are determined; according to the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics, the predicted transaction volume of the target business is determined.
[0166] In one embodiment, the candidate prediction model corresponding to any time category is trained in the following manner:
[0167] Obtain sample transaction data under the time category; input sample transaction attribute features and sample change trend features corresponding to the sample transaction data into the initial prediction model to obtain prediction results of the sample transaction data; wherein the initial prediction model includes a trend change sub-model, a cycle change sub-model and a holiday impact sub-model; determine the loss value according to the actual results and the predicted results of the sample transaction data; use the loss value to adjust the network parameters of the initial prediction model to obtain a candidate prediction model under the time category.
[0168] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:
[0169] Obtain target transaction data of target business;
[0170] Based on the generation time of the target transaction data, a target transaction sequence is generated according to the target transaction data;
[0171] Decompose the target transaction sequence to obtain the target transaction attribute characteristics and target change trend characteristics;
[0172] Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
[0173] In one embodiment, when the computer program is executed by the processor to decompose the target transaction sequence and obtain the target transaction attribute characteristics and the target change trend characteristics, the following steps are specifically implemented:
[0174] The target transaction sequence is decomposed by the empirical mode decomposition method to obtain multiple candidate transaction attribute features and target change trend features; wherein the candidate transaction attribute features include a time feature component and a transaction feature component; according to the target transaction sequence and the candidate transaction attribute features, the target transaction attribute features are selected from the candidate transaction attribute features.
[0175] In one embodiment, when the computer program is executed by a processor to select a target transaction attribute feature from the candidate transaction attribute features according to the target transaction sequence and the candidate transaction attribute features, the following steps are specifically implemented:
[0176] According to the association between each candidate transaction attribute feature and the target transaction sequence, the correlation coefficient value corresponding to each candidate transaction attribute feature is determined; according to the correlation coefficient value corresponding to each candidate transaction attribute feature, the target transaction attribute feature is selected from the candidate transaction attribute features.
[0177] In one embodiment, when the computer program is executed by the processor to determine the predicted transaction volume of the target business based on the target prediction model according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics, the following steps are specifically implemented:
[0178] According to the generation time of the target transaction data, a target prediction model is selected from the candidate prediction models corresponding to different time categories; through the target prediction model, the predicted transaction volume of the target business is determined according to the target transaction attribute characteristics and the target change trend characteristics.
[0179] In one embodiment, when the computer program is executed by the processor to determine the predicted transaction volume of the target business according to the target transaction attribute characteristics and the target change trend characteristics through the target prediction model, the following steps are specifically implemented:
[0180] Through the target prediction model, the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics are determined; according to the component prediction value corresponding to the target transaction attribute characteristics and the trend prediction value corresponding to the target change trend characteristics, the predicted transaction volume of the target business is determined.
[0181] In one embodiment, the candidate prediction model corresponding to any time category is trained in the following manner:
[0182] Obtain sample transaction data under the time category; input sample transaction attribute features and sample change trend features corresponding to the sample transaction data into the initial prediction model to obtain prediction results of the sample transaction data; wherein the initial prediction model includes a trend change sub-model, a cycle change sub-model and a holiday impact sub-model; determine the loss value according to the actual results and the predicted results of the sample transaction data; use the loss value to adjust the network parameters of the initial prediction model to obtain a candidate prediction model under the time category.
[0183] It should be noted that the data involved in this application (including but not limited to transaction data of various businesses, etc.) are all data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0184] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.
[0185] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are 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.
[0186] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A transaction volume prediction method, characterized in that: The method comprises: Obtain target transaction data of target business; Based on the generation time of the target transaction data, generate a target transaction sequence according to the target transaction data; Decomposing the target transaction sequence to obtain target transaction attribute characteristics and target change trend characteristics; Based on the target prediction model, the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
2. The method according to claim 1, characterized in that Decomposing the target transaction sequence to obtain target transaction attribute characteristics and target change trend characteristics includes: Decomposing the target transaction sequence by using the empirical mode decomposition method to obtain a plurality of candidate transaction attribute features and target change trend features; wherein the candidate transaction attribute features include a time feature component and a transaction feature component; According to the target transaction sequence and the candidate transaction attribute characteristics, the target transaction attribute characteristics are selected from the candidate transaction attribute characteristics.
3. The method according to claim 2, characterized in that The step of selecting the target transaction attribute feature from the candidate transaction attribute features according to the target transaction sequence and the candidate transaction attribute features includes: Determine the correlation coefficient value corresponding to each candidate transaction attribute feature according to the correlation between each candidate transaction attribute feature and the target transaction sequence; According to the correlation coefficient value corresponding to each candidate transaction attribute feature, a target transaction attribute feature is selected from each candidate transaction attribute feature.
4. The method according to any one of claims 1 to 3, characterized in that: The target prediction model is based on which the predicted transaction volume of the target business is determined according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics, including: According to the generation time of the target transaction data, a target prediction model is selected from candidate prediction models corresponding to different time categories; The target prediction model is used to determine the predicted transaction volume of the target business according to the target transaction attribute characteristics and the target change trend characteristics.
5. The method according to claim 4, characterized in that Determining the predicted transaction volume of the target business according to the target transaction attribute characteristics and the target change trend characteristics through the target prediction model includes: Determine, by means of the target prediction model, a component prediction value corresponding to the target transaction attribute feature and a trend prediction value corresponding to the target change trend feature; The predicted transaction volume of the target business is determined according to the component prediction value corresponding to the target transaction attribute feature and the trend prediction value corresponding to the target change trend feature.
6. The method according to claim 4, characterized in that The candidate prediction model corresponding to any time category is trained in the following way: Get sample transaction data under this time category; Inputting the sample transaction attribute features and sample change trend features corresponding to the sample transaction data into the initial prediction model to obtain the prediction results of the sample transaction data; wherein the initial prediction model includes a trend change sub-model, a cycle change sub-model and a holiday impact sub-model; Determine the loss value based on the actual results and predicted results of the sample transaction data; The loss value is used to adjust the network parameters of the initial prediction model to obtain a candidate prediction model under the time category.
7. A transaction volume prediction device, characterized in that: The device comprises: A data acquisition module is used to acquire target transaction data of a target business; A sequence generation module, configured to generate a target transaction sequence according to the target transaction data based on the generation time of the target transaction data; A sequence decomposition module is used to decompose the target transaction sequence to obtain target transaction attribute characteristics and target change trend characteristics; The prediction module is used to determine the predicted transaction volume of the target business based on the target prediction model, according to the generation time of the target transaction data, the target transaction attribute characteristics and the target change trend characteristics.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.