Transaction volume prediction method, training method, device, equipment, medium and product

Abstract

The invention provides a transaction volume prediction method and device, a training method and device, equipment, a medium and a product, which can be applied to the fields of big data and financial science and technology. The method comprises the steps that a transaction network diagram of a target historical time period is acquired, the transaction network diagram comprises transaction nodes and edge relations, the transaction nodes represent places where transactions are executed, the edge relations represent transaction activeness of transactions between different transaction places, the transaction nodes comprise transaction and environment data related to the transaction places, and the transaction nodes represent the transaction activeness of the transactions between different transaction places; the transaction position data represents the position of the transaction place, and the transaction and environment data have time sequence attributes; and processing the transaction node and edge relationship of the transaction network diagram by using the transaction volume prediction model to obtain a predicted transaction volume.

Description

Technical Field

[0001] The present disclosure relates to the fields of big data and financial technology, and specifically to a transaction volume prediction method, training method, device, equipment, medium and product. Background Art

[0002] In order to better respond to business needs, trading platforms usually need to predict the total trading volume of the day before the start of a day, so as to adjust the service resource calling strategy of the trading platform based on the predicted trading volume to meet the trading needs of users. For example, the storage space of database resources and the cache space of distributed cache nodes can be adjusted based on the predicted trading volume to improve the processing efficiency of massive business transactions. However, since the accuracy of the transaction volume prediction method is not high, it is necessary to provide a more accurate transaction volume prediction method. Summary of the invention

[0003] In view of the above problems, the present disclosure provides a transaction volume prediction method, training method, device, equipment, medium and product.

[0004] According to a first aspect of the present disclosure, a transaction volume prediction method is provided, comprising: obtaining a transaction network diagram of a target historical period, wherein the transaction network diagram comprises transaction nodes and edge relationships, the transaction nodes represent places where transactions are performed, the edge relationships represent the transaction activity between different transaction places, the transaction nodes comprise transaction and environmental data related to the transaction places, and transaction location data representing the locations of the transaction places, the transaction and environmental data having time series attributes; and using a transaction volume prediction model to process the transaction nodes and edge relationships of the transaction network diagram to obtain predicted transaction volume.

[0005] According to an embodiment of the present disclosure, a transaction volume prediction model includes N spatiotemporal graph convolutional neural networks, where N is an integer greater than 0, and the transaction and environment data include transaction data and transaction environment data with time series attributes; the transaction volume prediction model is used to process transaction nodes and edge relationships of a transaction network graph to obtain predicted transaction volume, including: using a feature extraction network to process the transaction network graph to obtain graph features, wherein the graph features include transaction environment features that characterize transaction environment data, transaction location features that characterize transaction location data, and transaction activity features that characterize edge relationships; based on the time series attributes, the first spatiotemporal graph convolutional neural network of the transaction volume prediction model is used to process the transaction environment features, transaction location features, and transaction activity features to obtain the first spatiotemporal fusion feature; the nth spatiotemporal graph convolutional neural network of the transaction volume prediction model is used to process the n-1th spatiotemporal fusion feature to obtain the nth spatiotemporal fusion feature; when N=n, the Nth spatiotemporal fusion feature is processed based on a fully connected layer to output a predicted transaction volume, wherein N≥n>1.

[0006] According to an embodiment of the present disclosure, the nth spatiotemporal graph convolutional neural network includes a first gated sequence convolutional layer, a spatial graph convolutional layer and a second gated sequence convolutional layer, and the gated sequence convolutional layer is used to retain the temporal attribute information related to transaction and environmental data; wherein, the nth spatiotemporal graph convolutional neural network of the transaction volume prediction model is used to process the n-1th spatiotemporal fusion feature to obtain the nth spatiotemporal fusion feature, including: inputting the n-1th spatiotemporal fusion feature into the first gated sequence convolutional layer, and outputting the nth transaction temporal fusion feature; based on the attention mechanism, the spatial graph convolutional layer is used to process the nth transaction temporal fusion feature to obtain the nth transaction spatial fusion feature; the nth transaction spatial fusion feature is input into the second gated sequence convolutional layer, and the nth spatiotemporal fusion feature is output.

[0007] According to an embodiment of the present disclosure, the n-1th spatiotemporal fusion feature is input into the first gated sequential convolution layer, and the nth transaction timing fusion feature is output, including: inputting the n-1th spatiotemporal fusion feature into the first convolution sublayer, and outputting the first graph convolution feature; processing the n-1th spatiotemporal fusion feature and the first graph convolution feature based on the gated linear unit, and outputting the nth timing fusion feature.

[0008] According to an embodiment of the present disclosure, the nth transaction space fusion feature is input into the second gated sequence convolution layer, and the nth spatiotemporal fusion feature is output, including: inputting the nth transaction space fusion feature into the second convolution sublayer, and outputting the second graph convolution feature; processing the nth transaction space fusion feature and the second graph convolution feature based on the gated linear unit, and outputting the nth spatiotemporal fusion feature.

[0009] According to an embodiment of the present disclosure, a transaction network diagram is obtained based on the following method: acquiring transaction data, wherein the transaction data includes transaction and environmental data and transaction location data; preprocessing the transaction data to obtain target transaction data, wherein the preprocessing includes data cleaning and denoising; and determining a transaction network diagram based on the target transaction data.

[0010] According to an embodiment of the present disclosure, the transaction volume prediction method further includes: outputting the predicted transaction volume corresponding to the target transaction node; and visually displaying the predicted transaction volume corresponding to the target transaction node.

[0011] A second aspect of the present disclosure provides a training method for a trading volume prediction model for predicting trading volume, comprising: obtaining trading network graph samples and label data for a historical period, wherein the label data represents the trading volume corresponding to the trading node in the trading network graph sample; processing the trading network graph sample using an initial trading volume prediction model to obtain the sample predicted trading volume; processing the sample predicted trading volume and label data according to a loss function to obtain a target loss value; and training the initial trading volume prediction model according to the target loss value to obtain a trading volume prediction model.

[0012] The third aspect of the present disclosure provides a transaction volume prediction device, including: an acquisition module, used to obtain a transaction network diagram of a target historical period, wherein the transaction network diagram includes transaction nodes and edge relationships, the transaction nodes represent the places where transactions are executed, the edge relationships represent the transaction activity between different transaction venues, the transaction nodes include transaction and environmental data related to the transaction venues, and transaction location data representing the locations of the transaction venues, and the transaction and environmental data have time series attributes; a predicted transaction volume acquisition module, used to use a transaction volume prediction model to process the transaction nodes and edge relationships of the transaction network diagram to obtain predicted transaction volume.

[0013] The fourth aspect of the present disclosure provides a training device, including: a second acquisition module, used to acquire transaction network graph samples and label data for a historical period, wherein the label data represents the transaction volume corresponding to the transaction node in the transaction network graph sample; a sample predicted transaction volume acquisition module, used to process the transaction network graph sample using an initial transaction volume prediction model to obtain the sample predicted transaction volume; a target loss value acquisition module, used to process the sample predicted transaction volume and label data according to a loss function to obtain a target loss value; a training module, used to train the initial transaction volume prediction model according to the target loss value to obtain a transaction volume prediction model.

[0014] A fifth aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the above method.

[0015] The sixth aspect of the present disclosure further provides a computer-readable storage medium having a computer program or instructions stored thereon, which implements the steps of the above method when the above computer program or instructions are executed by a processor.

[0016] The seventh aspect of the present disclosure further provides a computer program product, including a computer program or instructions, which implement the steps of the above method when the above computer program or instructions are executed by a processor.

[0017] According to an embodiment of the present disclosure, a transaction network diagram of a target historical period can be obtained. The transaction network diagram may include transaction nodes and edge relationships. The transaction nodes may represent the place where the transaction is executed, and the edge relationship may represent the transaction activity between different transaction places. The transaction nodes may include transaction and environmental data related to the transaction place, and transaction location data representing the location of the transaction place. The transaction and environmental data have time series attributes. The obtained transaction network diagram efficiently integrates data with spatiotemporal dynamic characteristics. Then, the obtained transaction network diagram is processed using a transaction volume prediction model to extract high-dimensional features in the transaction network diagram and obtain a more accurate predicted transaction volume. The predicted transaction volume can also be used to identify and prevent illegal trading behaviors and enhance data security. Accurately predicting transaction volume can also help avoid losses due to waste or shortage of resources, further reducing the operating costs of financial institutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0019] Figure 1 A diagram schematically shows an application scenario of a transaction volume prediction method, a training method, and a device according to an embodiment of the present disclosure;

[0020] Figure 2 A flowchart of a transaction volume prediction method according to an embodiment of the present disclosure is schematically shown;

[0021] Figure 3 The structure diagram of a transaction prediction model including multiple spatiotemporal graph convolutional neural networks according to an embodiment of the present disclosure is schematically shown;

[0022] Figure 4 Schematically shows a structural diagram of an n-th spatiotemporal graph convolutional neural network according to an embodiment of the present disclosure;

[0023] Figure 5A The structure diagram of the first gated sequential convolution layer according to an embodiment of the present disclosure is schematically shown;

[0024] Figure 5B Schematically shows a structural diagram of a second gated sequential convolutional layer according to an embodiment of the present disclosure;

[0025] Figure 6 A flowchart of a method for training a transaction volume prediction model for predicting transaction volume according to an embodiment of the present disclosure is schematically shown;

[0026] Figure 7 Schematically shows a schematic diagram according to another embodiment of the present disclosure;

[0027] Figure 8The structure block diagram of the transaction volume prediction device according to the embodiment of the present disclosure is schematically shown;

[0028] Fig. 9 A structural block diagram of a training device according to an embodiment of the present disclosure is schematically shown; and

[0029] Fig.10 A block diagram of an electronic device suitable for implementing a product recommendation method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0030] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0031] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0032] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0033] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0034] In order to better respond to business needs, trading platforms usually need to predict the total transaction volume of the day before the start of a certain day, so as to adjust the service resource calling strategy of the trading platform based on the predicted transaction volume to meet the transaction needs of users. For example, the storage space of database resources and the cache space of distributed cache nodes can be adjusted based on the predicted transaction volume to improve the processing efficiency of massive business transactions. However, since the transaction volume prediction method has low accuracy in predicting transaction volume, a more accurate transaction volume prediction method needs to be provided.

[0035] In view of this, the present disclosure provides a transaction volume prediction method, training method, device, equipment, medium and product. The method includes: obtaining a transaction network diagram of a target historical period, wherein the transaction network diagram includes transaction nodes and edge relationships, the transaction nodes represent the places where transactions are executed, the edge relationships represent the transaction activity between different transaction places, the transaction nodes include transaction and environmental data related to the transaction places, and transaction location data representing the location of the transaction places, and the transaction and environmental data have time series attributes; using a transaction volume prediction model to process the transaction nodes and edge relationships of the transaction network diagram to obtain predicted transaction volume.

[0036] It should be noted that the transaction volume prediction method, training method, prediction device and training device provided in the present disclosure can be used in the field of financial technology, such as banks and other financial institutions, and can also be used in any field other than the field of financial technology, such as the field of big data. Therefore, the application field of the transaction volume prediction method, training method, prediction device and training device provided in the present disclosure is not limited.

[0037] In the technical solution of the present disclosure, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved 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 relevant laws, regulations and standards, 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.

[0038] In the scenario of using personal information for automated decision-making, the methods, devices, and systems provided by the embodiments of the present disclosure provide users with corresponding operation portals for users to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered. The expression "automated decision-making" here refers to the activity of automatically analyzing and evaluating a person's behavioral habits, interests and hobbies, or economic, health, credit status, etc. through computer programs, and making decisions. The expression "expert decision-making" here refers to the activity of making decisions by people who specialize in a certain field, have specialized experience, knowledge and skills, and have reached a certain level of professionalism.

[0039] Figure 1 The application scenario diagram of the transaction volume prediction method, training method and device according to the embodiments of the present disclosure is schematically shown.

[0040] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a scenario of recommending bank financial products. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0041] The user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only for example).

[0042] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0043] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0044] It should be noted that the transaction volume prediction method and training method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the transaction volume prediction device and training device provided in the embodiment of the present disclosure can generally be set in the server 105. The transaction volume prediction method and training method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the transaction volume prediction device and training device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0045] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0046] Figure 2 The flowchart of the transaction volume prediction method according to the embodiment of the present disclosure is schematically shown.

[0047] like Figure 2 As shown, the transaction volume prediction method of this embodiment includes operations S210 to S240, and the transaction volume prediction method can be executed by an electronic device.

[0048] In operation S210, a transaction network diagram of a target historical period is obtained.

[0049] In operation S220, the transaction volume prediction model is used to process the transaction nodes and edge relationships of the transaction network graph to obtain the predicted transaction volume.

[0050] According to an embodiment of the present disclosure, a transaction network diagram may include transaction nodes and edge relationships. The transaction nodes may represent the place where the transaction is performed, and the edge relationship may represent the transaction activity between different transaction places. The transaction nodes may include transaction and environmental data related to the transaction place, and transaction location data representing the location of the transaction place. The transaction and environmental data have time series attributes.

[0051] According to an embodiment of the present disclosure, transaction and environmental data may include transaction data and environmental data. Transaction data may include transaction volume, transaction amount, transaction type, etc. Transaction type may include consumption transactions in which payment is made directly by swiping or inserting a card, cash withdrawal transactions in which cash is withdrawn through ATM machines or counters, or transfer transactions in which funds are transferred between different bank accounts through bank counters, online banking, mobile banking, etc., or installment payment transactions in which the consumption amount is paid in installments when making consumption, etc., but are not limited to this. The embodiments of the present disclosure do not limit the types of transaction types.

[0052] According to the embodiments of the present disclosure, environmental data may include weather, traffic flow, holiday information, holiday type, etc. during the transaction period, but is not limited thereto. The embodiments of the present disclosure do not limit the transaction and environmental data.

[0053] According to the embodiments of the present disclosure, the transaction volume prediction model can select a graph neural network model suitable for processing data with spatiotemporal attributes, such as a spatiotemporal graph neural network, but is not limited thereto. The embodiments of the present disclosure do not limit this. The transaction volume prediction model can be used to process the transaction nodes and edge relationships of the transaction network graph to obtain the predicted transaction volume.

[0054] According to an embodiment of the present disclosure, a transaction network diagram of a target historical period can be obtained. The transaction network diagram may include transaction nodes and edge relationships. The transaction nodes may represent the place where the transaction is executed, and the edge relationship may represent the transaction activity between different transaction places. The transaction nodes may include transaction and environmental data related to the transaction place, and transaction location data representing the location of the transaction place. The transaction and environmental data have time series attributes. The obtained transaction network diagram efficiently integrates data with spatiotemporal dynamic characteristics. Then, the obtained transaction network diagram is processed using a transaction volume prediction model to extract high-dimensional features in the transaction network diagram and obtain a more accurate predicted transaction volume. The predicted transaction volume can also be used to identify and prevent illegal trading behaviors and enhance data security. Accurately predicting transaction volume can also help avoid losses due to waste or shortage of resources, further reducing the operating costs of financial institutions.

[0055] According to an embodiment of the present disclosure, a transaction volume prediction model includes N spatiotemporal graph convolutional neural networks, where N is an integer greater than 0, and the transaction and environment data include transaction data and transaction environment data with time series attributes; the transaction volume prediction model is used to process transaction nodes and edge relationships of a transaction network graph to obtain predicted transaction volume, including: using a feature extraction network to process the transaction network graph to obtain graph features, wherein the graph features include transaction environment features that characterize transaction environment data, transaction location features that characterize transaction location data, and transaction activity features that characterize edge relationships; based on the time series attributes, the first spatiotemporal graph convolutional neural network of the transaction volume prediction model is used to process the transaction environment features, transaction location features, and transaction activity features to obtain the first spatiotemporal fusion feature; the nth spatiotemporal graph convolutional neural network of the transaction volume prediction model is used to process the n-1th spatiotemporal fusion feature to obtain the nth spatiotemporal fusion feature; when N=n, the Nth spatiotemporal fusion feature is processed based on a fully connected layer to output a predicted transaction volume, wherein N≥n>1.

[0056] According to an embodiment of the present disclosure, the transaction volume prediction model may include N spatiotemporal graph convolutional neural networks, where N is an integer greater than 0, and the transaction and environment data may include transaction data and transaction environment data with time series attributes.

[0057] Figure 3 The structure diagram of a transaction prediction model including multiple spatiotemporal graph convolutional neural networks according to an embodiment of the present disclosure is schematically shown.

[0058] like Figure 3 As shown, the transaction network graph 310 can be input into the feature extraction network 320 to obtain graph features, wherein the graph features may include transaction environment features characterizing transaction environment data, transaction location features characterizing transaction location data, and transaction activity features characterizing edge relationships; based on time series attributes, the first spatiotemporal graph convolutional neural network 330 of the transaction volume prediction model can be used to process the transaction environment features, transaction location features, and transaction activity features to obtain the first spatiotemporal fusion features; the nth spatiotemporal graph convolutional neural network 340 of the transaction volume prediction model can be used to process the n-1th spatiotemporal fusion features to obtain the nth spatiotemporal fusion features; in the case of N=n, the Nth spatiotemporal fusion features can be processed based on the fully connected layer 350 to output the predicted transaction volume 360, wherein N≥n>1.

[0059] According to an embodiment of the present disclosure, the nth spatiotemporal graph convolutional neural network includes a first gated sequence convolutional layer, a spatial graph convolutional layer and a second gated sequence convolutional layer; wherein, the nth spatiotemporal graph convolutional neural network of the transaction volume prediction model is used to process the n-1th spatiotemporal fusion feature to obtain the nth spatiotemporal fusion feature, including: inputting the n-1th spatiotemporal fusion feature into the first gated sequence convolutional layer, and outputting the nth transaction timing fusion feature; based on the attention mechanism, the spatial graph convolutional layer is used to process the nth transaction timing fusion feature to obtain the nth transaction spatial fusion feature; the nth transaction spatial fusion feature is input into the second gated sequence convolutional layer, and the nth spatiotemporal fusion feature is output.

[0060] According to an embodiment of the present disclosure, the nth spatiotemporal graph convolutional neural network may include a first gated sequence convolutional layer, a spatial graph convolutional layer, and a second gated sequence convolutional layer, and the gated sequence convolutional layer may be used to retain temporal attribute information related to transaction and environmental data.

[0061] Figure 4 The structure diagram of the nth spatiotemporal graph convolutional neural network according to an embodiment of the present disclosure is schematically shown.

[0062] like Figure 4 As shown, the nth spatiotemporal graph convolutional neural network 410 may include a first gated sequence convolutional layer 411, a spatial graph convolutional layer 412 and a second gated sequence convolutional layer 413. The n-1th spatiotemporal fusion feature is input into the first gated sequence convolutional layer 411, and the nth transaction time series fusion feature is output. Based on the attention mechanism, the spatial graph convolutional layer 412 is used to process the nth transaction time series fusion feature to obtain the nth transaction space fusion feature. The nth transaction space fusion feature is input into the second gated sequence convolutional layer 413, and the nth spatiotemporal fusion feature is output.

[0063] According to an embodiment of the present disclosure, the n-1th spatiotemporal fusion feature is input into the first gated sequential convolution layer, and the nth transaction timing fusion feature is output, including: inputting the n-1th spatiotemporal fusion feature into the first convolution sublayer, and outputting the first graph convolution feature; processing the n-1th spatiotemporal fusion feature and the first graph convolution feature based on the first gated linear unit, and outputting the nth timing fusion feature.

[0064] According to an embodiment of the present disclosure, the first gated sequential convolutional layer can introduce a gating mechanism so that the model can selectively pass information, thereby improving the expressiveness and performance of the model.

[0065] Figure 5A The structure diagram of the first gated sequential convolutional layer according to an embodiment of the present disclosure is schematically shown.

[0066] like Figure 5AAs shown, the first gated sequential convolution layer 510 may include a first convolution sublayer 512 and a first gated linear unit 511. The n-1th spatiotemporal fusion feature is input to the first convolution sublayer 512, and the first graph convolution feature can be output; based on the first gated linear unit 511, the n-1th spatiotemporal fusion feature and the first graph convolution feature are processed, and the nth temporal fusion feature can be output.

[0067] According to an embodiment of the present disclosure, the nth transaction space fusion feature is input into the second gated sequence convolution layer, and the nth spatiotemporal fusion feature is output, including: inputting the nth transaction space fusion feature into the second convolution sublayer, and outputting the second graph convolution feature; processing the nth transaction space fusion feature and the second graph convolution feature based on the second gated linear unit, and outputting the nth spatiotemporal fusion feature.

[0068] According to an embodiment of the present disclosure, the second gated linear unit can introduce a gating mechanism to enable the model to selectively pass information, thereby improving the model's expressiveness and performance. The nth transaction space fusion feature is input into the second convolution sublayer, and the second graph convolution feature can be output; based on the second gated linear unit, the nth transaction space fusion feature and the second graph convolution feature are processed, and the nth spatiotemporal fusion feature can be output.

[0069] Figure 5B The structure diagram of the second gated sequential convolution layer according to an embodiment of the present disclosure is schematically shown.

[0070] like Figure 5B As shown, the second gated sequential convolution layer 510 may include a second convolution sublayer 522 and a second gated linear unit 521. The second gated sequential convolution layer may introduce a gating mechanism to enable the model to selectively pass information, thereby improving the expression ability and performance of the model.

[0071] The nth transaction space fusion feature is input into the second convolution sublayer 522, and the second graph convolution feature can be output; based on the second gated linear unit 521, the nth transaction space fusion feature and the second graph convolution feature are processed to output the nth spatiotemporal fusion feature.

[0072] According to an embodiment of the present disclosure, a transaction network diagram is obtained based on the following method: acquiring transaction data, wherein the transaction data includes transaction and environmental data and transaction location data; preprocessing the transaction data to obtain target transaction data, wherein the preprocessing includes data cleaning and denoising; and determining a transaction network diagram based on the target transaction data.

[0073] According to an embodiment of the present disclosure, transaction data may include transaction and environmental data and transaction location data, and the acquired transaction data may be preprocessed, and the preprocessing includes data cleaning and denoising, etc., but is not limited thereto, and the embodiment of the present disclosure does not limit the specific steps of preprocessing. By cleaning and processing the transaction data, it can be ensured that the data quality meets the requirements of model training. A transaction network diagram can be determined based on the target transaction data. Specifically, transactions in different trading venues can be connected. In the transaction network diagram, various places are regarded as nodes in the transaction network diagram, and information such as transaction volume, weather, holidays, etc. are regarded as attributes of the nodes, and there will also be influences between the nodes.

[0074] According to the embodiments of the present disclosure, the features of transaction data such as holidays, weather, and locations that may affect the transaction volume are integrated to obtain a transaction network diagram, which solves the problem of insufficient prediction accuracy due to the limitation of the types and number of features used, and improves the prediction accuracy.

[0075] According to an embodiment of the present disclosure, the transaction volume prediction method further includes: outputting the predicted transaction volume corresponding to the target transaction node; and visually displaying the predicted transaction volume corresponding to the target transaction node.

[0076] According to the embodiments of the present disclosure, the predicted transaction volume corresponding to the output target transaction node can be visualized to help users understand and analyze the prediction results. The visualized predicted transaction volume can enhance market transparency and help regulators and the market to detect abnormal trading behavior in a timely manner.

[0077] Figure 6 A flowchart of a method for training a transaction volume prediction model for predicting transaction volume according to an embodiment of the present disclosure is schematically shown.

[0078] like Figure 6 As shown, the training method of the transaction volume prediction model for predicting transaction volume in this embodiment includes operations S610 to S640, and the transaction volume prediction method can be executed by an electronic device.

[0079] In operation S610, a transaction network graph sample and label data of a historical period are obtained.

[0080] In operation S620, the transaction network graph sample is processed using the initial transaction volume prediction model to obtain the sample predicted transaction volume.

[0081] In operation S630, the sample predicted transaction volume and label data are processed according to the loss function to obtain a target loss value.

[0082] In operation S640, the initial transaction volume prediction model is adjusted according to the target loss value to obtain a transaction volume prediction model.

[0083] According to the embodiments of the present disclosure, a transaction network diagram and label data of a historical period can be obtained, and the label data can characterize the transaction volume corresponding to the transaction node in the transaction network diagram sample. The sample predicted transaction volume can be obtained by processing the transaction network diagram sample using the initial transaction volume prediction model, and the sample predicted transaction volume and label data can be processed using the loss function to obtain the target loss value. Finally, the initial transaction volume prediction model can be adjusted according to the target loss value to obtain the transaction volume prediction model.

[0084] According to the embodiments of the present disclosure, the performance of the trained transaction volume prediction model can be evaluated and further optimized based on the evaluation results.

[0085] According to the embodiments of the present disclosure, different types of input data will have different effects on the overall prediction performance of the model. Specifically, the transaction volume prediction model is predicted based on the temporal and spatial relationship between the data, so when the characteristics of the training data are very different from the actual data to be predicted, it will affect the model performance. For example, using data from developed areas to train the network to predict the transaction volume in remote areas, or vice versa, it will affect the accuracy of the model. Alternatively, the transaction volume of transfer transactions and cash withdrawal transactions during the epidemic will be significantly different from the same type of transaction types in normal times. Therefore, the training data needs to have relatively complete temporal and spatial characteristics, and also have a high similarity with the object to be predicted.

[0086] Figure 7 A schematic diagram according to another embodiment of the present disclosure is schematically shown.

[0087] like Figure 7 As shown, the data acquisition module 710 can be used to acquire data to obtain information such as the current date, transaction volume in various places, and corresponding weather, as shown in Table 1.

[0088] Table 1 Input data acquired by the data acquisition module

[0089] Data Item describe time Timestamp of transaction volume Trading location The geographic location or institution code where the transaction took place Weather conditions Weather conditions on this date Holiday Information Is the date a holiday? Holiday Type If it is a holiday, is it a statutory holiday or another type of holiday? Weekdays / Weekends Whether the transaction occurred on a weekday or weekend

[0090] According to an embodiment of the present disclosure, the acquired data is processed by the data acquisition module 720, and the data acquisition module 720 may include data cleaning and data storage, etc. Data cleaning can clean and process the collected transaction data to ensure that the data quality meets the requirements of model training, and data storage can store the cleaned transaction data for subsequent construction of the transaction network diagram and model training.

[0091] According to an embodiment of the present disclosure, a transaction network graph can be constructed based on the stored data, wherein nodes represent different transaction locations or places where transactions are conducted, edges represent connections between transaction locations, and edge relationships can characterize the transaction activity between different transaction locations. Transaction nodes can include transaction and environmental data related to the transaction locations, as well as transaction location data characterizing the locations of the transaction locations, such as geographic location, transaction type, etc.

[0092] According to an embodiment of the present disclosure, the model can be trained using the model training module 730, which can specifically include parameter adjustment and model evaluation. The parameters of the transaction volume prediction model can be adjusted according to historical transaction data to improve the prediction accuracy, the performance of the model can be evaluated using historical transaction volume, and further optimized according to the evaluation results.

[0093] According to the embodiments of the present disclosure, the prediction module 740 can be used to predict future transaction volumes, and business decisions can be made based on the prediction results. The prediction results can be visualized to help users understand and analyze the prediction results, and the actual transaction amount can be adjusted based on the predicted transaction volume to meet business needs.

[0094] Figure 8 The structural block diagram of the transaction volume prediction device according to an embodiment of the present disclosure is schematically shown.

[0095] like Figure 8 As shown, the transaction volume prediction device 800 of this embodiment includes a first acquisition module 810 and a predicted transaction volume acquisition module 820.

[0096] The first acquisition module 810 is used to obtain a transaction network diagram of a target historical period, wherein the transaction network diagram includes transaction nodes and edge relationships, the transaction nodes represent the places where transactions are performed, the edge relationships represent the transaction activity between different transaction places, the transaction nodes include transaction and environmental data related to the transaction places, and transaction location data representing the locations of the transaction places, and the transaction and environmental data have time series attributes.

[0097] The conversion module 820 is used to process the transaction nodes and edge relationships of the transaction network graph using the transaction volume prediction model to obtain the predicted transaction volume.

[0098] According to an embodiment of the present disclosure, a transaction network diagram of a target historical period can be obtained. The transaction network diagram may include transaction nodes and edge relationships. The transaction nodes may represent the place where the transaction is executed, and the edge relationship may represent the transaction activity between different transaction places. The transaction nodes may include transaction and environmental data related to the transaction place, and transaction location data representing the location of the transaction place. The transaction and environmental data have time series attributes. The obtained transaction network diagram efficiently integrates data with spatiotemporal dynamic characteristics. Then, the acquired transaction network diagram is processed using a transaction volume prediction model to extract high-dimensional features in the transaction network diagram, thereby obtaining a more accurate predicted transaction volume. The predicted transaction volume can also be used to identify and prevent illegal trading behaviors, thereby enhancing data security.

[0099] According to an embodiment of the present disclosure, the transaction volume prediction model includes N spatiotemporal graph convolutional neural networks, where N is an integer greater than 0, and the transaction and environment data include transaction data and transaction environment data with time series attributes.

[0100] According to an embodiment of the present disclosure, the predicted transaction volume obtaining module includes: a graph feature obtaining submodule, a first spatiotemporal fusion feature obtaining submodule, an nth spatiotemporal fusion feature obtaining submodule and a predicted transaction volume obtaining submodule.

[0101] The graph feature acquisition submodule is used to process the transaction network graph using a feature extraction network to obtain graph features, wherein the graph features include transaction environment features that characterize transaction environment data, transaction location features that characterize transaction location data, and transaction activity features that characterize edge relationships.

[0102] The first spatiotemporal fusion feature obtaining submodule is used to process the transaction environment features, transaction location features and transaction activity features based on time series attributes using the first spatiotemporal graph convolutional neural network of the transaction volume prediction model to obtain the first spatiotemporal fusion features.

[0103] The nth spatiotemporal fusion feature obtaining submodule is used to process the n-1th spatiotemporal fusion feature using the nth spatiotemporal graph convolutional neural network of the transaction volume prediction model to obtain the nth spatiotemporal fusion feature.

[0104] The predicted transaction volume obtaining submodule is used to process the Nth spatiotemporal fusion feature based on the fully connected layer when N=n, and output the predicted transaction volume, where N≥n>1.

[0105] According to an embodiment of the present disclosure, the nth spatiotemporal graph convolutional neural network includes a first gated sequential convolutional layer, a spatial graph convolutional layer, and a second gated sequential convolutional layer.

[0106] According to an embodiment of the present disclosure, the nth spatiotemporal fusion feature obtaining submodule includes: an nth transaction time series fusion feature output unit, an nth transaction space fusion feature obtaining unit and an nth spatiotemporal fusion feature output unit.

[0107] The nth transaction timing fusion feature output unit is used to input the n-1th spatiotemporal fusion feature into the first gated sequential convolution layer and output the nth transaction timing fusion feature.

[0108] The nth transaction spatial fusion feature obtaining unit is used to process the nth transaction temporal fusion feature based on the attention mechanism and utilize the spatial graph convolution layer to obtain the nth transaction spatial fusion feature.

[0109] The nth spatiotemporal fusion feature output unit is used to input the nth transaction space fusion feature into the second gated sequential convolution layer and output the nth spatiotemporal fusion feature.

[0110] According to an embodiment of the present disclosure, the nth transaction timing fusion feature output unit includes: a first graph convolution feature output subunit and an nth timing fusion feature output subunit.

[0111] The first graph convolution feature output subunit is used to input the n-1th spatiotemporal fusion feature into the first convolution sublayer and output the first graph convolution feature.

[0112] The nth temporal fusion feature output subunit is used to process the n-1th spatiotemporal fusion feature and the first graph convolution feature based on the gated linear unit, and output the nth temporal fusion feature.

[0113] According to an embodiment of the present disclosure, the nth spatiotemporal fusion feature output unit includes: a second graph convolution feature output subunit and an nth spatiotemporal fusion feature output subunit.

[0114] The second graph convolution feature output subunit is used to input the nth transaction space fusion feature into the second convolution sublayer and output the second graph convolution feature.

[0115] The nth spatiotemporal fusion feature output subunit is used to process the nth transaction space fusion feature and the second graph convolution feature based on the gated linear unit, and output the nth spatiotemporal fusion feature.

[0116] According to an embodiment of the present disclosure, the first acquisition module includes: a transaction data acquisition submodule, a preprocessing submodule and a transaction network diagram determination submodule.

[0117] The transaction data acquisition submodule is used to acquire transaction data, wherein the transaction data includes transaction and environment data and transaction location data.

[0118] The preprocessing submodule is used to preprocess the transaction data to obtain target transaction data, wherein the preprocessing includes data cleaning and denoising.

[0119] The transaction network diagram determination submodule is used to determine the transaction network diagram based on the target transaction data.

[0120] According to an embodiment of the present disclosure, the transaction volume prediction device 800 of this embodiment includes: a predicted transaction volume output module and a visualization module.

[0121] The predicted transaction volume output module is used to output the predicted transaction volume corresponding to the target transaction node.

[0122] The visualization module is used to visualize the predicted transaction volume corresponding to the target transaction node.

[0123] Fig. 9 The structural block diagram of the training device according to an embodiment of the present disclosure is schematically shown.

[0124] like Fig. 9 As shown, the training device 900 of this embodiment includes a second acquisition module 910, a sample predicted transaction volume acquisition module 920, a target loss value acquisition module 930 and a training module 940.

[0125] The second acquisition module 910 is used to acquire a transaction network graph sample and label data of a historical period, wherein the label data represents the transaction volume corresponding to the transaction node in the transaction network graph sample.

[0126] The sample predicted transaction volume obtaining module 920 is used to process the transaction network graph sample using the initial transaction volume prediction model to obtain the sample predicted transaction volume.

[0127] The target loss value obtaining module 930 is used to process the sample predicted transaction volume and the label data according to the loss function to obtain a target loss value.

[0128] The training module 940 is used to train the initial transaction volume prediction model according to the target loss value to obtain a transaction volume prediction model.

[0129] According to an embodiment of the present disclosure, any multiple modules among the first acquisition module 910, the predicted transaction volume acquisition module 920, the second acquisition module 910, the sample predicted transaction volume acquisition module 920, the target loss value acquisition module 930 and the training module 940 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the first acquisition module 910, the predicted transaction volume acquisition module 920, the second acquisition module 910, the sample predicted transaction volume acquisition module 920, the target loss value acquisition module 930 and the training module 940 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in an appropriate combination of any of them. Alternatively, at least one of the first acquisition module 910, the predicted transaction volume acquisition module 920, the second acquisition module 910, the sample predicted transaction volume acquisition module 920, the target loss value acquisition module 930 and the training module 940 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is executed.

[0130] Fig.10 A block diagram of an electronic device suitable for implementing a product recommendation method according to an embodiment of the present disclosure is schematically shown.

[0131] like Fig.10 As shown, the electronic device 1000 according to an embodiment of the present disclosure includes a processor 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage part 1008 to a random access memory (RAM) 1003. The processor 1001 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 1001 may also include an onboard memory for caching purposes. The processor 1001 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0132] In RAM 1003, various programs and data required for the operation of electronic device 1000 are stored. Processor 1001, ROM 1002 and RAM 1003 are connected to each other via bus 1004. Processor 1001 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 1002 and / or RAM 1003. It should be noted that the program can also be stored in one or more memories other than ROM 1002 and RAM 1003. Processor 1001 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0133] According to an embodiment of the present disclosure, the electronic device 1000 may further include an input / output (I / O) interface 1005, which is also connected to the bus 1004. The electronic device 1000 may further include one or more of the following components connected to the input / output (I / O) interface 1005: an input portion 1006 including a keyboard, a mouse, etc.; an output portion 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 1008 including a hard disk, etc.; and a communication portion 1009 including a network interface card such as a LAN card, a modem, etc. The communication portion 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the input / output (I / O) interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage portion 1008 as needed.

[0134] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0135] According to an embodiment of the present disclosure, a computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present disclosure, a computer-readable storage medium may include the ROM 1002 and / or RAM 1003 described above and / or one or more memories other than ROM 1002 and RAM 1003.

[0136] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the product recommendation method provided by the embodiment of the present disclosure.

[0137] The above functions defined in the system / device of the embodiment of the present disclosure are performed when the computer program is executed by the processor 1001. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0138] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 10010, and / or installed from the removable medium 1011. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0139] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 10010, and / or installed from the removable medium 1011. When the computer program is executed by the processor 1001, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0140] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0141] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0142] It will be appreciated by those skilled in the art that the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present disclosure. In particular, without departing from the spirit and teachings of the present disclosure, the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0143] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A transaction volume prediction method, characterized in that: include: Obtaining a transaction network diagram for a target historical period, wherein the transaction network diagram includes transaction nodes and edge relationships, the transaction nodes represent places where transactions are performed, the edge relationships represent transaction activity between different transaction places, the transaction nodes include transaction and environment data related to the transaction places, and transaction location data representing the location of the transaction places, and the transaction and environment data have a time series attribute; The transaction volume prediction model is used to process the transaction nodes and edge relationships of the transaction network graph to obtain the predicted transaction volume.

2. The method according to claim 1, characterized in that The transaction volume prediction model includes N spatiotemporal graph convolutional neural networks, where N is an integer greater than 0, and the transaction and environment data includes transaction data and transaction environment data with time series attributes; The step of using the transaction volume prediction model to process the transaction nodes and edge relationships of the transaction network graph to obtain the predicted transaction volume includes: Processing the transaction network graph using a feature extraction network to obtain graph features, wherein the graph features include transaction environment features characterizing the transaction environment data, transaction location features characterizing the transaction location data, and transaction activity features characterizing the edge relationships; Based on the time series attributes, the transaction environment feature, the transaction location feature, and the transaction activity feature are processed using the first spatiotemporal graph convolutional neural network of the transaction volume prediction model to obtain a first spatiotemporal fusion feature; Using the nth spatiotemporal graph convolutional neural network of the transaction volume prediction model to process the n-1th spatiotemporal fusion feature to obtain the nth spatiotemporal fusion feature; In the case of N=n, the Nth spatiotemporal fusion feature is processed based on the fully connected layer to output the predicted transaction volume, where N≥n>1.

3. The method according to claim 2, characterized in that The nth spatiotemporal graph convolutional neural network includes a first gated sequential convolutional layer, a spatial graph convolutional layer, and a second gated sequential convolutional layer; The process of using the n-1th spatiotemporal fusion feature using the nth spatiotemporal graph convolutional neural network of the transaction volume prediction model to obtain the nth spatiotemporal fusion feature includes: Input the n-1th spatiotemporal fusion feature into the first gated sequential convolutional layer, and output the nth transaction temporal fusion feature; Based on the attention mechanism, the spatial graph convolution layer is used to process the nth transaction temporal fusion feature to obtain the nth transaction spatial fusion feature; The nth transaction space fusion feature is input into the second gated sequential convolutional layer, and the nth spatiotemporal fusion feature is output.

4. The method according to claim 3, characterized in that The step of inputting the n-1th spatiotemporal fusion feature into the first gated sequential convolutional layer and outputting the nth transaction time series fusion feature comprises: Input the n-1th spatiotemporal fusion feature into the first convolution sublayer, and output the first graph convolution feature; The n-1th spatiotemporal fusion feature and the first graph convolution feature are processed based on a first gated linear unit to output an nth temporal fusion feature.

5. The method according to claim 3, characterized in that: The step of inputting the nth transaction space fusion feature into the second gated sequence convolution layer and outputting the nth spatiotemporal fusion feature comprises: Inputting the nth transaction space fusion feature into the second convolution sublayer, and outputting the second graph convolution feature; The nth transaction space fusion feature and the second graph convolution feature are processed based on a second gated linear unit to output an nth spatiotemporal fusion feature.

6. The method according to claim 1, characterized in that The transaction network diagram is obtained based on the following method: Acquiring transaction data, wherein the transaction data includes transaction and environment data and transaction location data; Preprocessing the transaction data to obtain target transaction data, wherein the preprocessing includes data cleaning and denoising; Based on the target transaction data, a transaction network diagram is determined.

7. The method according to claim 1, characterized in that The method further comprises: Outputting the predicted transaction volume corresponding to the target transaction node; and The predicted transaction volume corresponding to the target transaction node is visualized.

8. A method for training a trading volume prediction model for predicting trading volume, characterized in that: The method comprises: Acquire a transaction network graph sample and label data for a historical period, wherein the label data represents a transaction volume corresponding to a transaction node in the transaction network graph sample; Processing the transaction network graph sample using the initial transaction volume prediction model to obtain the sample predicted transaction volume; Processing the sample predicted transaction volume and the label data according to a loss function to obtain a target loss value; The initial transaction volume prediction model is trained according to the target loss value to obtain a transaction volume prediction model.

9. A transaction volume prediction device, characterized in that: include: A first acquisition module is used to acquire a transaction network diagram of a target historical period, wherein the transaction network diagram includes transaction nodes and edge relationships, the transaction nodes represent places where transactions are performed, the edge relationships represent transaction activity between different transaction places, the transaction nodes include transaction and environment data related to the transaction places, and transaction location data representing the location of the transaction places, and the transaction and environment data have a time series attribute; The predicted transaction volume obtaining module is used to process the transaction nodes and edge relationships of the transaction network graph using the transaction volume prediction model to obtain the predicted transaction volume.

10. A training device, characterized in that: include: A second acquisition module is used to acquire a transaction network graph sample and label data of a historical period, wherein the label data represents a transaction volume corresponding to a transaction node in the transaction network graph sample; A sample predicted transaction volume obtaining module is used to process the transaction network graph sample using the initial transaction volume prediction model to obtain the sample predicted transaction volume; A target loss value obtaining module, used for processing the sample predicted transaction volume and the label data according to a loss function to obtain a target loss value; A training module is used to train the initial transaction volume prediction model according to the target loss value to obtain a transaction volume prediction model.

11. An electronic device, comprising: one or more processors; a memory for storing one or more computer programs, It is characterized in that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 8.

12. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

13. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images