A model training method, order recognition method and related device

Abstract

The embodiments of the present application disclose a model training method, an order recognition method and related devices, which can be applied to various scenarios such as cloud technology, artificial intelligence, smart transportation, and assisted driving. In the model training method, a plurality of simulated ordinary orders in a simulated resource transfer scenario are generated through a resource transfer scenario model; a plurality of simulated abnormal orders in a simulated resource transfer scenario are generated through an abnormal resource transfer model; based on simulated ordinary orders and simulated abnormal orders, an order recognition model is trained to recognize order types in a simulated resource scenario. In this way, a real resource transfer scenario is simulated, and then based on the two different order types of simulated abnormal orders and simulated ordinary orders in the simulated resource transfer scenario, an order recognition model is trained so that it can accurately and reliably identify ordinary orders and abnormal orders in a real resource transfer scenario, thereby improving the detection accuracy of abnormal orders and ensuring the security of resource transfer.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a model training method, an order recognition method and related devices. Background Art

[0002] In recent years, with the rapid development of computer technology and Internet technology, online resource transfer business has become more and more common. Online resource transfer business can be, for example, online transaction business, where users can use applications with online transaction functions to perform online order payment, online transfer and other operations.

[0003] For the backend system that supports online resource transfer business, in order to ensure the security of online resource transfer, it is necessary to detect the relevant resource transfer orders to determine whether they are abnormal orders (such as illegal transaction orders, etc.), and take relevant intervention measures for abnormal orders in a timely manner. In related technologies, abnormal orders are usually detected and identified based on preset rules (obtained from statistical summaries of the characteristics of historical abnormal orders); however, this method is difficult to accurately and reliably identify complex resource transfer orders, and is prone to false detection or missed detection of abnormal orders. Summary of the invention

[0004] The embodiments of the present application provide a model training method, an order recognition method and related devices, which can train an order recognition model that can accurately and reliably recognize various types of orders, improve the accuracy of detecting abnormal orders, and thereby ensure the security of resource transfer.

[0005] The first aspect of the present application provides a model training method, the method comprising:

[0006] Generate multiple simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model; the resource transfer scenario model is trained based on virtual common orders, and the virtual common orders are constructed according to a first system dynamics model for indicating characteristics of common resource transfer orders;

[0007] Generate multiple simulated abnormal orders in the simulated resource transfer scenario through an abnormal resource transfer model; the abnormal resource transfer model is trained based on virtual abnormal orders, and the virtual abnormal orders are constructed according to a second system dynamics model for indicating characteristics of abnormal resource transfer orders;

[0008] Based on the simulated ordinary orders and the simulated abnormal orders, the order recognition model is trained to identify the order type in the simulated resource scenario; the order recognition model that meets the training end conditions is used to identify the type of real resource transfer orders in the actual resource transfer scenario.

[0009] A second aspect of the present application provides an order identification method, the method comprising:

[0010] Obtain target resource transfer order;

[0011] Determine the order type corresponding to the target resource transfer order through an order recognition model; the order recognition model is obtained by training the model training method described in the first aspect;

[0012] According to the order type corresponding to the target resource transfer order, a target operation is performed on the target resource transfer order.

[0013] A third aspect of the present application provides a model training device, the device comprising:

[0014] A first order generation module is used to generate a plurality of simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model; the resource transfer scenario model is trained based on virtual common orders, and the virtual common orders are constructed according to a first system dynamics model for indicating characteristics of common resource transfer orders;

[0015] A second order generation module is used to generate a plurality of simulated abnormal orders in the simulated resource transfer scenario through an abnormal resource transfer model; the abnormal resource transfer model is trained based on virtual abnormal orders, and the virtual abnormal orders are constructed according to a second system dynamics model for indicating characteristics of abnormal resource transfer orders;

[0016] A model training module is used to train an order recognition model to recognize order types in the simulated resource scenario based on the simulated ordinary orders and the simulated abnormal orders; the order recognition model that meets the training end conditions is used to recognize the type of real resource transfer orders in actual resource transfer scenarios.

[0017] Optionally, the resource transfer scenario model is trained by the following units:

[0018] A first model building unit, configured to build the first system dynamics model; the first system dynamics model includes a first resource transfer process model and a first resource transfer variable model;

[0019] A first order building unit, configured to build a plurality of the virtual common orders according to the first resource transfer process model and the first resource transfer variable model;

[0020] The first model training unit is used to train the resource transfer scenario model based on the virtual ordinary order, so that the resource transfer scenario model learns the characteristics of the ordinary resource transfer order.

[0021] Optionally, the first order construction unit is used to determine the virtual resource transfer node in the virtual ordinary order according to the resource transfer node included in the first resource transfer process model; determine the virtual resource transfer node in the virtual ordinary order according to the resource transfer variable information included in the first resource transfer variable model; and construct the virtual ordinary order according to the virtual resource transfer node and the virtual resource transfer variable.

[0022] Optionally, the abnormal resource transfer model is trained by the following units:

[0023] A second model building unit, configured to build the second system dynamics model based on historical abnormal transaction orders; the second system dynamics model includes a second resource transfer process model and a second resource transfer variable model;

[0024] A second order construction unit, configured to construct a plurality of the virtual abnormal orders according to the second resource transfer process model and the second resource transfer variable model;

[0025] The second model training unit is used to train the abnormal resource transfer model based on the virtual abnormal order so that the abnormal resource transfer model learns the characteristics of the abnormal resource transfer order.

[0026] Optionally, the second order construction unit is used to determine the virtual resource transfer node in the virtual abnormal order according to the resource transfer node included in the second resource transfer process model; determine the virtual resource transfer variable in the virtual abnormal order according to the resource transfer variable information included in the second resource transfer variable model; and construct the virtual abnormal order according to the virtual resource transfer node and the virtual resource transfer variable.

[0027] Optionally, the model training module includes:

[0028] A type identification unit, used to identify the prediction type corresponding to the training simulation order through the order identification model; the training simulation order is the simulated ordinary order or the simulated abnormal order;

[0029] The model parameter adjustment unit is used to adjust the model parameters of the order recognition model according to the predicted type corresponding to the training simulation order and the actual type of the training simulation order.

[0030] A fourth aspect of the present application provides an order identification device, the device comprising:

[0031] An order acquisition module is used to obtain target resource transfer orders;

[0032] An order recognition model, used to determine the order type corresponding to the target resource transfer order through the order recognition model; the order recognition model is trained by the model training method described in the first aspect;

[0033] The order processing module is used to perform a target operation on the target resource transfer order according to the order type corresponding to the target resource transfer order.

[0034] Optionally, the order processing module is used to perform a target alarm operation on a related object of the target resource transfer order when the order type corresponding to the target resource transfer order indicates that the target resource transfer order is an abnormal resource transfer order.

[0035] A fifth aspect of the present application provides a computer device, the device comprising a processor and a memory:

[0036] The memory is used to store computer programs;

[0037] The processor is used to execute the steps of the model training method described in the first aspect or the steps of the order identification method described in the second aspect according to the computer program.

[0038] In a sixth aspect, the present application provides a computer-readable storage medium, which is used to store a computer program, and the computer program is used to execute the steps of the model training method described in the first aspect above, or the computer program is used to execute the steps of the order identification method described in the second aspect above.

[0039] In a seventh aspect, the present application provides a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the steps of the model training method described in the first aspect, or the computer device executes the steps of the order identification method described in the second aspect.

[0040] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0041] This application first generates multiple simulated ordinary orders in a simulated resource transfer scenario through a resource transfer scenario model; wherein the resource transfer scenario model is trained based on a virtual ordinary order, and the virtual ordinary order is constructed according to a first system dynamics model for indicating the characteristics of an ordinary resource transfer order. Then, through an abnormal resource transfer model, multiple simulated abnormal orders in a simulated resource transfer scenario are generated; wherein the abnormal resource transfer model is trained based on a virtual abnormal order, and the virtual abnormal order is constructed according to a second system dynamics model for indicating the characteristics of an abnormal resource transfer order; subsequently, based on simulated ordinary orders and simulated abnormal orders, an order recognition model is trained to identify order types in a simulated resource scenario; when the training end condition is met, the training is terminated to obtain an order recognition model, which can be used to identify the type of real resource transfer orders in an actual resource transfer scenario. In this way, through the resource transfer scenario model and the abnormal resource transfer model, multiple simulated ordinary orders and multiple simulated abnormal orders in a simulated resource transfer scenario are generated respectively to simulate a real resource transfer scenario. Based on the two different order types, simulated abnormal orders and simulated ordinary orders in the simulated resource transfer scenario, the order recognition model is trained so that it can accurately and reliably identify ordinary orders and abnormal orders in real resource transfer scenarios, thereby improving the detection accuracy of abnormal orders and ensuring the security of resource transfer. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A scenario architecture diagram of a model training method provided in an embodiment of the present application;

[0043] Figure 2a A flowchart of a model training method provided in an embodiment of the present application;

[0044] Figure 2b A flowchart of an order identification method provided in an embodiment of the present application;

[0045] Figure 3 A signaling diagram of an online transaction process provided in an embodiment of the present application;

[0046] Figure 4 A flowchart of a specific model training method provided in an embodiment of the present application;

[0047] Figure 5 A technical architecture diagram for identifying abnormal orders provided in an embodiment of the present application;

[0048] Figure 6 A schematic diagram of calculating the payment security level provided in an embodiment of the present application;

[0049] Figure 7a A schematic diagram of the structure of a model training device provided in an embodiment of the present application;

[0050] Figure 7b A schematic diagram of the structure of an order identification device provided in an embodiment of the present application;

[0051] Figure 8 A schematic diagram of the structure of a terminal device provided in an embodiment of the present application;

[0052] Fig. 9 A schematic diagram of the structure of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0054] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0055] Online resource transfer business refers to a business in which one party transfers relevant resources to another party using virtual media such as the Internet. For example, the online resource transfer business can be an online transaction business, and users can use applications with online transaction functions to perform online order payments, online transfers, and other operations. Currently, abnormal orders may occur, affecting the security of resource transfer. Therefore, for the background system that supports the online resource transfer business, it is necessary to detect the resource transfer orders related to the online resource transfer to determine whether the resource transfer order is an abnormal order (such as an illegal transaction order, etc.), and take relevant intervention measures for abnormal orders in a timely manner to ensure the security of resource transfer.

[0056] In the related art, statistics and summaries are usually performed based on the characteristics of historical abnormal orders to obtain preset rules for detecting and identifying abnormal orders. However, the preset rules are generally simple and cannot accurately and reliably identify the more complex resource transfer orders in resource transfer scenarios. In other words, it is easy to identify abnormal orders as ordinary orders, resulting in false detection or missed detection of abnormal orders, thereby affecting the security of resource transfer.

[0057] In view of the above problems, a model training method, an order recognition method and related devices are provided in the present application. In the model training method, a plurality of simulated ordinary orders in a simulated resource transfer scenario are generated through a resource transfer scenario model; the resource transfer scenario model is trained based on a virtual ordinary order, and the virtual ordinary order is constructed according to a first system dynamics model for indicating the characteristics of an ordinary resource transfer order; a plurality of simulated abnormal orders in a simulated resource transfer scenario are generated through an abnormal resource transfer model; the abnormal resource transfer model is trained based on a virtual abnormal order, and the virtual abnormal order is constructed according to a second system dynamics model for indicating the characteristics of an abnormal resource transfer order; based on simulated ordinary orders and simulated abnormal orders, an order recognition model is trained to recognize order types in a simulated resource scenario; an order recognition model that meets the training end condition is used to recognize the type of a real resource transfer order in an actual resource transfer scenario.

[0058] In this way, through the resource transfer scenario model and the abnormal resource transfer model, multiple simulated ordinary orders and multiple simulated abnormal orders in the simulated resource transfer scenario are generated respectively to simulate the real resource transfer scenario. Then, based on the two different order types of simulated abnormal orders and simulated ordinary orders in the simulated resource transfer scenario, the order recognition model is trained to enable it to accurately and reliably identify ordinary orders and abnormal orders in the real resource transfer scenario, thereby improving the detection accuracy of abnormal orders and ensuring the security of resource transfer.

[0059] Next, the execution subject of the model training method provided in the embodiment of the present application is introduced in detail.

[0060] The execution subject of the model training method provided in the embodiment of the present application can be a computer device with data processing capabilities, and the computer device can specifically be a terminal device or a server. As an example, the terminal device can specifically include but is not limited to mobile phones, desktop computers, tablet computers, notebook computers, PDAs, intelligent voice interaction devices, smart home appliances, vehicle terminals, aircraft, etc. The server can be an independent physical server, or it can be a server cluster or distributed system composed of multiple physical servers. In addition, the server can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. In addition, the model training method provided in the embodiment of the present application can also be executed collaboratively by the terminal device and the server. Among them, the terminal and the server can be directly or indirectly connected by wired or wireless communication, and this application is not limited here. Therefore, in the embodiment of the present application, the implementation subject of the technical solution of the present application is not limited.

[0061] In addition, it should be noted that the executor of the order identification method provided in the embodiment of the present application may also be a computer device with data processing capabilities. The computer device may specifically be a terminal device or a server. Please refer to the above description for details, which will not be repeated here.

[0062] Next, the scenario of the model training method provided in the embodiment of the present application is introduced in detail.

[0063] See also Figure 1 , which is a scenario architecture diagram of a model training method provided in an embodiment of the present application, including a computer device 100. The computer device 100 can be a terminal device or a server in the above-mentioned various forms. Specifically, the following embodiment is exemplified by taking the computer device 100 as server A.

[0064] Server A calls the resource transfer scenario model to generate M simulated normal orders (including simulated normal order 1, simulated normal order 2, ... and simulated normal order M) in the simulated resource transfer scenario; Server A then calls the abnormal resource transfer model to generate N simulated abnormal orders (including simulated abnormal order 1, simulated abnormal order 2, ... and simulated abnormal order N) in the simulated resource transfer scenario. Subsequently, Server A trains the order recognition model based on the M simulated normal orders and the N simulated abnormal orders until the training end condition is met, and the training order recognition model is obtained, which is used to identify the type of real resource transfer orders in the actual resource transfer scenario.

[0065] It can be seen that by simulating ordinary orders and simulating abnormal orders to simulate real resource transfer scenarios, the order recognition model is trained to learn the different characteristics of these two different order types, namely, simulated ordinary orders and simulated abnormal orders, so that it can accurately and reliably identify ordinary orders and abnormal orders in real resource transfer scenarios, improve the detection accuracy of abnormal orders, and thus ensure the security of resource transfer.

[0066] Next, taking the server as the execution entity, the model training method provided in the embodiment of the present application is specifically introduced.

[0067] See also Figure 2a , which is a flow chart of a model training method provided in an embodiment of the present application, combined with Figure 2a As shown, the model training method may specifically include:

[0068] S201: Generate multiple simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model.

[0069] In the embodiments of the present application, the simulated resource transfer scenario refers to a virtual scenario used to simulate a real resource transfer scenario. The simulated ordinary order is used to simulate a real ordinary (non-abnormal) order in a real resource transfer scenario. For example, the simulated ordinary order can be a simulated online transaction ordinary order, which can include a transaction party, a transaction recipient, a payment institution, a transaction time, and a transaction amount, etc., which is not limited in the present application.

[0070] Through the resource transfer scenario model, multiple simulated ordinary orders in the simulated resource transfer scenario can be generated to simulate real ordinary orders in the real resource transfer scenario.

[0071] The resource transfer scenario model is trained based on a virtual ordinary order and is used to generate a simulated ordinary order with the same order characteristics as the virtual ordinary order. The virtual ordinary order is constructed based on the first system dynamics model, which is used to indicate the characteristics of an ordinary resource transfer order (that is, a real ordinary order in a real resource transfer scenario). Therefore, the simulated ordinary order generated by the resource transfer scenario model also has the characteristics of an ordinary resource transfer order.

[0072] In a possible implementation manner of the present application, the resource transfer scenario model may be obtained by training through the following steps:

[0073] A1: Establish the first system dynamics model.

[0074] In an embodiment of the present application, the first system dynamics model refers to a structural model established by applying the system dynamics principles to analyze the resource transfer process in the real resource transfer scenario, the structure of the ordinary resource transfer order, the resource transfer behavior corresponding to the ordinary resource transfer order, etc. in the real resource transfer scenario in order to indicate the characteristics of the ordinary resource transfer order in the real resource transfer scenario, and simulating the real resource transfer scenario and the dynamic changes of the ordinary resource transfer order.

[0075] By applying system dynamics to analyze ordinary resource transfer orders, a first system dynamics model can be established to construct a virtual ordinary order with the same order characteristics as the ordinary resource transfer order.

[0076] In a possible implementation manner of the present application, the first system dynamics model may specifically include a first resource transfer process model and a first resource transfer variable model.

[0077] Among them, the first resource transfer process model is used to indicate the characteristics of the resource transfer process in the real resource transfer scenario, and is used to constrain the resource transfer process corresponding to the virtual ordinary order. It can avoid the constructed virtual ordinary order and the ordinary resource transfer order being too different, and enhance the similarity between the virtual ordinary order and the ordinary resource transfer order.

[0078] The first resource transfer variable model is used to indicate the variable characteristics of ordinary resource transfer orders in real resource transfer scenarios, and is used to constrain order variables in virtual ordinary orders. It can also avoid excessive differences between the constructed virtual ordinary orders and ordinary resource transfer orders, and further enhance the similarity between virtual ordinary orders and ordinary resource transfer orders.

[0079] As an example, the first resource transfer process model and the first resource transfer variable model can both be expressed in the mathematical form of constraint equations. That is, the first resource transfer process model can be expressed as a first resource transfer process equation, and the first resource transfer process equation is used to indicate the constraint conditions of the resource transfer process based on which the virtual ordinary order is constructed. The first resource transfer variable model can be expressed as a first resource transfer variable equation, and the first resource transfer variable equation is used to indicate the constraint conditions of the resource transfer variables based on which the virtual ordinary order is constructed.

[0080] A2: Construct multiple virtual ordinary orders according to the first resource transfer process model and the first resource transfer variable model.

[0081] According to the first resource transfer process model and the first resource transfer variable model included in the first system dynamics model, a plurality of virtual ordinary orders having the same order characteristics as ordinary resource transfer orders are constructed.

[0082] A3: Train the resource transfer scenario model based on virtual ordinary orders so that the resource transfer scenario model can learn the characteristics of ordinary resource transfer orders.

[0083] A virtual ordinary order has the same order characteristics as an ordinary resource transfer order. Therefore, training the resource transfer scenario model based on the virtual ordinary order can enable the resource transfer scenario model to learn the characteristics of an ordinary resource transfer order and generate multiple simulated ordinary orders with the same order characteristics as an ordinary resource transfer order.

[0084] It can be seen that the first system dynamics model for indicating the characteristics of ordinary resource transfer orders is established by applying the system dynamics principles. It can construct multiple virtual ordinary orders with the characteristics of ordinary resource transfer orders, so that the resource transfer scenario model can learn the characteristics of ordinary resource transfer orders based on virtual ordinary orders, and then generate multiple simulated ordinary orders in the simulated resource transfer scenario. This not only avoids obtaining ordinary resource transfer orders in real resource transfer scenarios and infringing on relevant privacy information, but also can simulate the characteristics of ordinary resource transfer orders in real and complex resource transfer scenarios.

[0085] In a possible implementation manner of the present application, the process of constructing multiple virtual ordinary orders according to the first resource transfer process model and the first resource transfer variable model in A2 above can be further subdivided into the following steps:

[0086] A21: Determine a virtual resource transfer node in the virtual ordinary order according to the resource transfer nodes included in the first resource transfer process model.

[0087] In an embodiment of the present application, the first resource transfer process model may include multiple resource transfer nodes. A resource transfer node refers to a node in the resource transfer process. Taking the resource transfer business as an online transaction business as an example, the resource transfer nodes included in the first resource transfer process model may be a transaction initiator, a transaction recipient, a payment institution, a payment request, a payment result, and an abnormal risk, etc., which is not limited in the present application. The first resource transfer process model is used to indicate the characteristics of an ordinary resource transfer order, so the abnormal risk therein is a value less than a preset abnormal risk threshold.

[0088] The virtual resource transfer node refers to a virtual node corresponding to the resource transfer node, which is used to constrain the virtual ordinary order to avoid problems in the resource transfer process and to avoid large differences from the real ordinary resource transfer order. For example, if the resource transfer node included in the first resource transfer process model is the transaction initiator, then the virtual resource transfer node is also the transaction initiator in the virtual ordinary order. The virtual resource transfer node will only implement the functions corresponding to the transaction initiator (initiating a transaction request, etc.), avoiding unreasonable situations such as the transaction receiver initiating a transaction request in the virtual ordinary order.

[0089] According to the resource transfer nodes included in the first resource transfer process model, a virtual resource transfer node for constraining the virtual ordinary order can be determined.

[0090] A22: Determine a virtual resource transfer variable in a virtual ordinary order according to the resource transfer variable information included in the first resource transfer variable model.

[0091] In the embodiment of the present application, the first resource transfer variable model may include a plurality of resource transfer variable information. Resource transfer variable information refers to variable information that may change in a common resource transfer order in a real resource transfer scenario. Virtual resource transfer variables refer to variables that conform to resource transfer variable information.

[0092] According to the resource transfer variable information included in the first resource transfer variable model, a virtual resource transfer variable for constraining a virtual ordinary order may be determined.

[0093] Specifically, taking the resource transfer business as an online transaction business as an example, the resource transfer variable information included in the first resource transfer process model can be: transaction amount, transaction time, transaction direction, transaction type, credit rating of transaction participants (i.e., transaction initiator, transaction recipient, and payment institution, etc.), and abnormal risk and other horizontal variable information, or it can be the number of payment requests, payment request success rate (for example, statistics on whether the payment requests of multiple ordinary resource transfer orders are successfully requested), payment request failure rate, payment request processing time, transaction recipient response time, payment institution processing time, payment institution processing success rate (for example, statistics on whether the payment of multiple ordinary resource transfer orders is successfully obtained), payment institution processing failure rate, payment institution processing volume, transaction recipient processing time, transaction recipient processing success rate (for example, statistics on whether the transaction recipients corresponding to multiple ordinary resource transfer orders have successfully received them), transaction recipient processing failure rate, transaction recipient processing volume, number of abnormal reports (for example, obtained by counting the abnormal risks of multiple ordinary resource transfer orders), these auxiliary variables, can also be constants such as payment institution processing capacity, transaction recipient processing capacity and abnormal risk threshold.

[0094] The horizontal variables are state variables that change over time in real resource transfer scenarios; the auxiliary variables are intermediate variables used to describe the resource transfer process; the constants are variables that do not change over time in resource transfer scenarios and are fixed variables. Constants can be pre-set fixed values ​​or fixed values ​​obtained by counting multiple common resource transfer orders in real resource transfer scenarios.

[0095] A23: Construct a virtual ordinary order based on the virtual resource transfer node and the virtual resource transfer variable.

[0096] According to the above virtual resource transfer nodes and virtual resource transfer variables, a virtual ordinary order having the same order characteristics as an ordinary resource transfer order can be constructed.

[0097] It can be seen that the constructed virtual ordinary order is constrained by the resource transfer nodes in the first resource transfer process model and the resource transfer variable information in the first resource transfer variable model, so that its corresponding virtual resource transfer nodes and virtual resource transfer nodes conform to the actual resource transfer scenario.

[0098] S202: Generate multiple simulated abnormal orders in a simulated resource transfer scenario through an abnormal resource transfer model.

[0099] In an embodiment of the present application, a simulated abnormal order is used to simulate an abnormal order in a real resource transfer scenario, and an abnormal order refers to an order that will affect the security of resource transfer. Taking the resource transfer business as an online transaction business as an example, an abnormal order can be an illegal transaction order, and the present application does not limit this. For example, the simulated abnormal order can be a simulated online transaction abnormal order, and the simulated online transaction abnormal order can include the transaction party, the transaction recipient, the payment institution, the transaction time, and the transaction amount, etc., and the present application does not limit this.

[0100] Through the abnormal resource transfer model, multiple simulated abnormal orders in the simulated resource transfer scenario can be generated to simulate the real abnormal orders in the real resource transfer scenario.

[0101] The above abnormal resource transfer model is trained based on virtual abnormal orders and is used to generate simulated abnormal orders with the same order characteristics as the virtual abnormal orders. The virtual abnormal orders are constructed based on the second system dynamics model, which is used to indicate the characteristics of abnormal resource transfer orders (that is, real abnormal orders in real resource transfer scenarios). Therefore, the simulated abnormal orders generated by the abnormal resource transfer model also have the characteristics of abnormal resource transfer orders.

[0102] Taking the resource transfer business as an online transaction business as an example, the anomaly in the online transaction business may be that the account application time of the transaction recipient corresponding to the order is short, but the number of transactions is too large; it may also be that the account of the transaction recipient corresponding to the order has been reported by other accounts; it may also be that the transaction initiator initiates the transaction request too frequently, etc. This application does not limit this.

[0103] In a possible implementation manner of the present application, the above abnormal resource transfer model may be obtained by training through the following steps:

[0104] B1: Based on historical abnormal trading orders, establish the second system dynamics model;

[0105] In the embodiment of the present application, the second system dynamics model refers to a structural model established by applying the system dynamics principle to analyze the resource transfer process in the real resource transfer scenario, the structure of the abnormal resource transfer order, the resource transfer behavior corresponding to the abnormal resource transfer order, etc., in order to indicate the characteristics of the abnormal resource transfer order in the real resource transfer scenario, and simulating the real resource transfer scenario and the dynamic changes of the abnormal resource transfer order. The historical abnormal transaction order refers to the order that has been judged as an abnormal transaction in the real resource transfer scenario.

[0106] Based on the historical abnormal transaction orders, the system dynamics analysis is applied to the historical abnormal transaction orders, and a second system dynamics model can be established to construct a virtual abnormal order with the same order characteristics as the abnormal resource transfer order.

[0107] In a possible implementation manner of the present application, the second system dynamics model may specifically include a second resource transfer process model and a second resource transfer variable model.

[0108] Among them, the second resource transfer process model is used to indicate the characteristics of the resource transfer process in the real resource transfer scenario, and is used to constrain the resource transfer process corresponding to the virtual abnormal order. It can avoid the constructed virtual abnormal order and the abnormal resource transfer order being too different, and enhance the similarity between the virtual abnormal order and the abnormal resource transfer order.

[0109] The second resource transfer variable model is used to indicate the variable characteristics of abnormal resource transfer orders in real resource transfer scenarios, and is used to constrain order variables in virtual abnormal orders. It can also avoid the constructed virtual abnormal orders and abnormal resource transfer orders being too different, thereby improving the similarity between virtual abnormal orders and abnormal resource transfer orders.

[0110] As an example, the second resource transfer process model and the second resource transfer variable model can both be expressed in the mathematical form of a constraint equation. That is, the second resource transfer process model can be expressed as a second resource transfer process equation, and the second resource transfer process equation is used to indicate the constraint conditions of the resource transfer process based on which the virtual abnormal order is constructed. The second resource transfer variable model can be expressed as a second resource transfer variable equation, and the second resource transfer variable equation is used to indicate the constraint conditions of the resource transfer variables based on which the virtual abnormal order is constructed.

[0111] B2: Construct multiple virtual abnormal orders according to the second resource transfer process model and the second resource transfer variable model.

[0112] According to the second resource transfer process model and the second resource transfer variable model included in the second system dynamics model, the resource transfer process and resource transfer variables are constrained to construct a plurality of virtual abnormal orders having the same order characteristics as the abnormal resource transfer order.

[0113] B3: Train the abnormal resource transfer model based on virtual abnormal orders so that the abnormal resource transfer model learns the characteristics of abnormal resource transfer orders.

[0114] It can be seen from the above embodiments that the virtual abnormal order has the same order characteristics as the abnormal resource transfer order. By training the abnormal resource transfer model based on the virtual abnormal order, the abnormal resource transfer model can learn the characteristics of the virtual abnormal order, that is, learn the characteristics of the abnormal resource transfer order in the real resource transfer scenario, so that it can generate multiple simulated abnormal orders with the same order characteristics as the abnormal resource transfer order.

[0115] It can be seen that the system dynamics is also applied to analyze the abnormal resource transfer orders in the real resource transfer scenario, and a second system dynamics model for indicating the characteristics of abnormal resource transfer orders is established, so that it can construct multiple virtual abnormal orders with the characteristics of abnormal resource transfer orders. Based on the training of the abnormal resource transfer model based on virtual abnormal orders, it can learn the characteristics of abnormal resource transfer orders, thereby generating multiple simulated abnormal orders with the characteristics of abnormal resource transfer orders in the simulated resource transfer scenario. The privacy of abnormal resource transfer orders in the real resource transfer scenario is protected, and the characteristics of abnormal resource transfer orders in real complex resource transfer scenarios are also simulated.

[0116] In a possible implementation manner of the present application, the process of constructing multiple virtual abnormal orders according to the second resource transfer process model and the second resource transfer variable model in B2 above can be further subdivided into the following steps:

[0117] B21: Determine the virtual resource transfer node in the virtual abnormal order according to the resource transfer nodes included in the second resource transfer process model.

[0118] In an embodiment of the present application, a resource transfer node refers to a node in the resource transfer process. Taking the resource transfer business as an online transaction business as an example, the resource transfer nodes included in the second resource transfer process model may include a transaction initiator, a transaction recipient, a payment institution, a payment request, a payment result, and an abnormal risk, etc., which are not limited in this application. The second resource transfer process model is used to indicate the characteristics of an abnormal resource transfer order, so the abnormal risk therein is a value greater than or equal to a preset abnormal risk threshold. It should be noted that this application does not limit the number of resource transfer nodes included in the second resource transfer process model.

[0119] The virtual resource transfer node refers to a virtual node corresponding to the resource transfer node, which is used to constrain the virtual abnormal order to avoid problems in the resource transfer process, resulting in excessive differences from the abnormal resource transfer order.

[0120] According to the resource transfer nodes included in the second resource transfer process model, a virtual resource transfer node for constraining the virtual abnormal order can be determined.

[0121] B22: Determine the virtual resource transfer variable in the virtual abnormal order according to the resource transfer variable information included in the second resource transfer variable model.

[0122] In an embodiment of the present application, the second resource transfer variable model may include a plurality of resource transfer variable information. Resource transfer variable information refers to variable information that may change in an abnormal resource transfer order in a real resource transfer scenario. Virtual resource transfer variables refer to variables that conform to resource transfer variable information.

[0123] According to the resource transfer variable information included in the second resource transfer variable model, a virtual resource transfer variable for constraining the virtual abnormal order can be determined.

[0124] B23: Construct a virtual abnormal order based on the virtual resource transfer node and the virtual resource transfer variable.

[0125] According to the above virtual resource transfer nodes and virtual resource transfer variables, a virtual abnormal order having the same order characteristics as the abnormal resource transfer order can be constructed.

[0126] It can be seen that a virtual abnormal order is constructed by constraining the resource transfer nodes in the second resource transfer process model and the resource transfer variable information in the second resource transfer variable model, wherein the virtual resource transfer nodes and virtual resource transfer variables are reasonable in the real resource transfer scenario.

[0127] S203: Based on simulated common orders and simulated abnormal orders, an order recognition model is trained to recognize order types in a simulated resource scenario.

[0128] Based on multiple simulated ordinary orders and multiple simulated abnormal orders in the simulated resource transfer scenario, the order recognition model is trained to identify the order types in the simulated resource scenario, so that it can learn to distinguish between the two different order types, simulated ordinary orders and simulated abnormal orders in the simulated resource scenario.

[0129] When the training end condition is met, the training of the order recognition model can be ended to obtain a trained order recognition model. The order recognition model can identify the type of real resource transfer orders in actual resource transfer scenarios, that is, it can identify abnormal resource transfer orders from multiple real resource transfer orders.

[0130] This application does not limit the conditions for the end of training. For example, it can be that the number of training times for the order recognition model reaches a preset number of training times; or the model performance of the order recognition model reaches a preset requirement, such as the order recognition model reaches an accuracy threshold for the order category recognition accuracy of simulated abnormal orders and simulated ordinary orders. This application does not limit this.

[0131] In a possible implementation manner of the present application, the training method of the above order recognition model, that is, S203, can be further subdivided into the following steps:

[0132] C1: Use the order recognition model to identify the prediction type corresponding to the training simulation order.

[0133] In an embodiment of the present application, the training simulation order is a simulated ordinary order generated by the above-mentioned resource transfer scenario model, or a simulated abnormal order generated by the above-mentioned abnormal resource transfer model.

[0134] The training simulated orders are input into the order recognition model so that it can recognize the training simulated orders and output the predicted type corresponding to the training simulated orders, that is, the type of the training simulated orders is identified to represent whether it is a simulated ordinary order or a simulated abnormal order.

[0135] C2: Adjust the model parameters of the order recognition model according to the predicted type corresponding to the training simulation order and the actual type of the training simulation order.

[0136] According to the prediction type output by the order recognition type and the actual type of the training simulation order, the model parameters of the order recognition model are adjusted until the training end conditions are met.

[0137] It can be seen that the order recognition model is trained by simulating ordinary orders and simulated abnormal orders, so that it can learn to distinguish between the two different order types of simulated ordinary orders and simulated abnormal orders. Simulated ordinary orders have the characteristics of ordinary resource transfer orders in real resource transfer scenarios, and simulated abnormal orders have the characteristics of abnormal resource transfer orders in real resource transfer scenarios. Therefore, the trained order recognition model can be used to identify whether the real resource transfer order is an ordinary resource transfer order or an abnormal resource transfer order in the actual resource transfer scenario.

[0138] In order to further understand the application scenarios of the model training method provided in the embodiment of the present application, next, the order identification method provided in the embodiment of the present application will be specifically introduced with the server as the execution entity.

[0139] See also Figure 2b , which is a flow chart of an order identification method provided by an embodiment of the present application. Figure 2b As shown, the order identification method may specifically include:

[0140] S21: Obtain the target resource transfer order.

[0141] In the embodiment of the present application, the target resource transfer order refers to any resource transfer order that needs to identify the order type. For example, the target resource transfer order can be a resource transfer order in an actual resource transfer scenario. In order to ensure the security of resource transfer, it is necessary to identify whether it is a normal order or an abnormal order with an abnormality.

[0142] This application does not limit the method of obtaining the target resource transfer order. For example, the target resource transfer order can be obtained when the transaction initiator initiates the transaction request. It can also be obtained when the institution verifies the resource transfer. This application does not limit this.

[0143] It should be emphasized that the acquisition and processing of target resource transfer orders in this application should comply with national laws and regulations. Before obtaining the target resource transfer order, the order processing rules should be informed and the target object’s separate consent should be sought. The target resource transfer order should be processed strictly in accordance with the requirements of laws and regulations and personal information processing rules, and technical measures should be taken to ensure the security of relevant data.

[0144] S22: Determine the order type corresponding to the target resource transfer order through the order identification model.

[0145] It should be noted that the order recognition model is trained by the model training method described in the above embodiments. The order recognition model can accurately and reliably identify ordinary orders and abnormal orders in real resource transfer scenarios, so the order type corresponding to the target resource transfer order can be determined.

[0146] The target resource transfer order obtained above is input into the order recognition model so that it predicts and outputs the order type corresponding to the target resource transfer order.

[0147] S23: Execute a target operation on the target resource transfer order according to the order type corresponding to the target resource transfer order.

[0148] It is understandable that when the target resource transfer order corresponds to different order types, different target operations need to be performed on the target resource transfer order.

[0149] In the embodiment of the present application, the target operation refers to an operation performed on the target resource transfer order based on the order type corresponding to the target resource transfer order, which can ensure the security of resource transfer.

[0150] Therefore, in a possible implementation of the present application, S23 may specifically be: when the order type corresponding to the target resource transfer order indicates that the target resource transfer order is an abnormal resource transfer order, a target alarm operation is performed on a related object of the target resource transfer order.

[0151] If the order type corresponding to the target resource transfer order indicates that the target resource transfer order is an abnormal resource transfer order, a target alarm operation needs to be performed on the related objects of the target resource transfer order to block the target resource transfer order and ensure the security of resource transfer.

[0152] In the embodiment of the present application, the related objects of the target resource transfer order refer to the objects involved in the resource transfer process of the target resource transfer order, for example, they may include the resource transfer initiator, the resource transfer recipient, and the resource transfer intermediary, etc., which are not limited in the present application.

[0153] The target alarm operation refers to an operation used to ensure the security of resource transfer. For example, the target alarm operation may be to send an alarm prompt to the object that initiates the resource transfer request, or to send an alarm prompt to the payment institution so that it can directly intercept the target resource transfer order, etc. This application does not limit this.

[0154] In addition, when the order type corresponding to the target resource transfer order indicates that the target resource transfer order is a common resource transfer order, the target pass operation can also be performed on the relevant objects of the target resource transfer order. Specifically, a prompt of order pass can be issued to the object initiating the resource transfer request, or a prompt of order pass can be issued to the payment institution so that it can directly pass the target resource transfer order, etc. This application does not limit this.

[0155] It can be seen that the order identification method provided in the embodiment of the present application applies the order identification model trained by the model training method in the above-mentioned embodiments. The order identification model can accurately and reliably identify the order type corresponding to the target resource transfer order in the real resource transfer scenario, and then perform the target operation on the target resource transfer order based on its order type, thereby ensuring the security of resource transfer.

[0156] To facilitate understanding, the application of the order recognition model, that is, the order recognition method, is described in detail below in combination with actual online transaction business.

[0157] See also Figure 3 , which is a signaling diagram of an online transaction process provided by an embodiment of the present application. Figure 3 As shown, the online transaction process may specifically include the following steps:

[0158] S301: receiving an online transaction initiation operation from a user;

[0159] S302: In response to the user's online transaction initiation operation, the transaction initiator sends a transaction request to the payment institution;

[0160] S303: The payment institution calls the order identification model in this application to identify the type of transaction order corresponding to the transaction request;

[0161] S304: If the type of the transaction order indicates that the transaction order is a common transaction order, the payment institution sends a transaction request to the transaction recipient and executes S306;

[0162] S305: If the type of the transaction order indicates that the transaction order is an abnormal transaction order, the payment institution sends a warning message to the transaction initiator and intercepts the transaction order;

[0163] S306: The transaction recipient calls the order identification model in this application to identify the type of transaction order corresponding to the transaction request;

[0164] S307: If the type of the transaction order indicates that the transaction order is a common transaction order, it is normal for the payment recipient to send a transaction request to the transaction structure, and S309 is executed;

[0165] S308: If the type of the transaction order indicates that the transaction order is an abnormal transaction order, the payment receiving direction sends a warning message to the transaction initiator via the transaction structure and intercepts the transaction order.

[0166] S309: The payment institution performs transaction processing on the payment request;

[0167] S310: After the transaction is completed, the payment institution sends a successful transaction result to the transaction initiator.

[0168] It can be seen that the order recognition model trained by the model training method provided in the embodiment of the present application is called twice in the above-mentioned online transaction process to identify the type of transaction orders. The order recognition model can accurately and reliably identify abnormal transaction orders in online transaction scenarios, send early warning information in a timely manner and intercept orders, thereby ensuring the security of online transactions.

[0169] To further understand the model training method provided in the embodiments of the present application, please refer to Figure 4 , which is a flowchart of a specific model training method provided in an embodiment of the present application. Figure 4 As shown, firstly, a virtual ordinary order is constructed based on a first system dynamics model including a first resource transfer process model and a first resource transfer variable model; then, a resource transfer scenario model is trained based on the virtual ordinary order to learn the characteristics of the ordinary resource transfer order; then, multiple simulated ordinary orders in the simulated resource transfer scenario are generated through the resource transfer scenario model. Similarly, firstly, a virtual abnormal order is constructed based on a second system dynamics model including a second resource transfer process model and a second resource transfer variable model; then, an abnormal resource transfer model is trained based on the virtual abnormal order to learn the characteristics of the abnormal resource transfer order; then, multiple simulated abnormal orders in the simulated resource transfer scenario are generated through the abnormal resource transfer model. Then, based on the simulated ordinary order and the simulated abnormal order, the order recognition model is trained to recognize the order type in the simulated resource scenario, and the order recognition model that finally meets the training end condition can be used to recognize the type of the real resource transfer order in the actual resource transfer scenario.

[0170] In addition, the model training method and order recognition method provided in the embodiments of the present application are specifically based on Figure 5 The technical architecture diagram for identifying abnormal orders is shown in the figure. Figure 5 As shown, it includes abnormal types, abnormal risk prevention and control technology and big data technology. The abnormal types of abnormal resource transfer orders include but are not limited to market risk abnormalities, credit risk abnormalities and operational risk abnormalities, etc. Figure 5 As shown, the model training method provided in this application mainly involves "machine learning" in abnormal risk prevention and control technology. Specifically, the model training of the resource transfer scenario model, abnormal resource transfer model and order identification model mentioned in the model training method involves model management technology, and the target operation performed by the target resource transfer order in the order identification method provided in this application involves risk monitoring and decision-making as well as risk disposal technologies. In this application, the acquisition and subsequent processing of target resource transfer orders and historical abnormal transaction orders also involve data protection in big data technology, access, processing and storage of data, and variable and model calculation technology. Therefore, it can be achieved through Figure 5The technical architecture for identifying abnormal orders shown implements the model training method and order identification method provided in this application.

[0171] In addition, in order to demonstrate that the model training method provided by this application can improve the accuracy of detecting abnormal orders and ensure the security of resource transfer. Specifically, taking the resource transfer business as an online transaction business as an example, the payment security level after applying the order recognition model trained in this application in the actual transaction scenario can be calculated. For details, please refer to Figure 6 As shown in FIG. 1 , this figure is a schematic diagram of calculating the payment security level provided by an embodiment of the present application. Figure 6 As shown, the current payment security level increment can be calculated based on the talent quantity function, the payment security level function and the payment security level growth table function, and the current payment security level can be calculated based on the current payment security level increment and the initial payment security level.

[0172] Specifically, it can be expressed by the following formula 1 and formula 2:

[0173] R5(t)=L5(t)×A5(L9(t))(Formula 1)

[0174] L5(t T )=L5(t0)×R5(t T )(Formula 2)

[0175] Among them, t represents the time, t0 represents the initial time, t T represents the current moment, L9(t) represents the talent quantity function, L5(t) represents the payment security function, A5(L9(t)) represents the payment security growth table function, R5(t T ) represents the current payment security increment, L5(t0) represents the initial payment security, L5(t T ) indicates the current payment security level.

[0176] It should be noted that the above-mentioned talent quantity function, payment security function and payment security growth table function are all examples, and this application does not limit them.

[0177] Combination Figure 6 The above formulas 1 and 2 can be used to calculate the payment security level after the order recognition model trained in this application is applied in the actual transaction scenario, which can be improved compared to the initial payment security level. Therefore, the order recognition model trained by the model training method provided in this application can accurately and reliably identify ordinary orders and abnormal orders in real resource transfer scenarios, improve the detection accuracy of abnormal orders, and ensure the security of resource transfer.

[0178] Based on the model training method provided in the above embodiment, the present application also provides a model training device accordingly. The model training device provided in the embodiment of the present application will be specifically introduced from the perspective of functional modularization.

[0179] See also Figure 7a , which is a schematic diagram of the structure of a model training device provided in an embodiment of the present application. Figure 7a As shown, the model training device may specifically include:

[0180] A first order generation module 710 is used to generate a plurality of simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model; the resource transfer scenario model is trained based on virtual common orders, and the virtual common orders are constructed according to a first system dynamics model for indicating characteristics of common resource transfer orders;

[0181] A second order generation module 720 is used to generate multiple simulated abnormal orders in a simulated resource transfer scenario through an abnormal resource transfer model; the abnormal resource transfer model is trained based on virtual abnormal orders, and the virtual abnormal orders are constructed according to a second system dynamics model for indicating characteristics of abnormal resource transfer orders;

[0182] Model training module 730 is used to train the order recognition model to recognize order types in simulated resource scenarios based on simulated ordinary orders and simulated abnormal orders; the order recognition model that meets the training end conditions is used to recognize the type of real resource transfer orders in actual resource transfer scenarios.

[0183] As an implementation mode, the resource transfer scenario model may be trained by the following units:

[0184] A first model building unit is used to build a first system dynamics model; the first system dynamics model includes a first resource transfer process model and a first resource transfer variable model;

[0185] A first order building unit, used to build a plurality of virtual common orders according to the first resource transfer process model and the first resource transfer variable model;

[0186] The first model training unit is used to train the resource transfer scenario model based on a virtual ordinary order, so that the resource transfer scenario model learns the characteristics of the ordinary resource transfer order.

[0187] As an implementation manner, the first order building unit is specifically used to:

[0188] Determine a virtual resource transfer node in a virtual ordinary order according to the resource transfer nodes included in the first resource transfer process model;

[0189] Determining a virtual resource transfer variable in a virtual ordinary order according to the resource transfer variable information included in the first resource transfer variable model;

[0190] A virtual ordinary order is constructed according to the virtual resource transfer node and the virtual resource transfer variable.

[0191] As an implementation mode, the above abnormal resource transfer model may be specifically trained by the following units:

[0192] A second model building unit is used to build a second system dynamics model based on historical abnormal transaction orders; the second system dynamics model includes a second resource transfer process model and a second resource transfer variable model;

[0193] A second order construction unit, configured to construct a plurality of virtual abnormal orders according to the second resource transfer process model and the second resource transfer variable model;

[0194] The second model training unit is used to train the abnormal resource transfer model based on the virtual abnormal order, so that the abnormal resource transfer model learns the characteristics of the abnormal resource transfer order.

[0195] As an implementation manner, the second order building unit is specifically used to:

[0196] Determine a virtual resource transfer node in the virtual abnormal order according to the resource transfer nodes included in the second resource transfer process model;

[0197] Determining a virtual resource transfer variable in a virtual abnormal order according to the resource transfer variable information included in the second resource transfer variable model;

[0198] A virtual abnormal order is constructed according to the virtual resource transfer node and the virtual resource transfer variable.

[0199] As an implementation method, the model training module 730 may specifically include:

[0200] A type identification unit is used to identify the prediction type corresponding to the training simulation order through the order identification model; the training simulation order is a simulated ordinary order or a simulated abnormal order;

[0201] The model parameter adjustment unit is used to adjust the model parameters of the order recognition model according to the predicted type corresponding to the training simulation order and the actual type of the training simulation order.

[0202] Based on the order recognition method provided in the above embodiment, the present application also provides an order recognition device. The order recognition device provided in the embodiment of the present application is specifically introduced below.

[0203] See also Figure 7b, which is a schematic diagram of the structure of an order identification device provided in an embodiment of the present application. Figure 7b As shown, the order identification device specifically includes:

[0204] An order acquisition module 71 is used to acquire a target resource transfer order;

[0205] An order recognition model 72 is used to determine the order type corresponding to the target resource transfer order through an order recognition model; the order recognition model is trained by the model training method described in the above embodiments;

[0206] The order processing module 73 is used to perform a target operation on the target resource transfer order according to the order type corresponding to the target resource transfer order.

[0207] As an implementation manner, the order processing module 73 may be specifically configured to execute a target alarm operation on a related object of the target resource transfer order when the order type corresponding to the target resource transfer order indicates that the target resource transfer order is an abnormal resource transfer order.

[0208] The embodiment of the present application also provides a computer device, which may specifically be a terminal device or a server. The terminal device and the server provided in the embodiment of the present application will be introduced below from the perspective of hardware entity.

[0209] See also Figure 8 , Figure 8 Schematic diagram of the structure of a terminal device provided in an embodiment of the present application. Figure 8 For the sake of convenience, only the parts related to the embodiments of the present application are shown. For specific technical details not disclosed, please refer to the method part of the embodiments of the present application. The terminal can be any terminal device including a mobile phone, a tablet computer, a personal digital assistant, a point of sales (POS), a car computer, etc., taking the terminal as a computer as an example:

[0210] Figure 8 FIG. 1 is a block diagram showing a partial structure of a computer related to a terminal provided in an embodiment of the present application. Figure 8 The computer includes: a radio frequency (RF) circuit 1210, a memory 1220, an input unit 1230 (including a touch panel 1231 and other input devices 1232), a display unit 1240 (including a display panel 1241), a sensor 1250, an audio circuit 1260 (which can be connected to a speaker 1261 and a microphone 1262), a wireless fidelity (WiFi) module 1270, a processor 1280, and a power supply 1290. Those skilled in the art can understand that Figure 8 The computer structure shown in the figure does not constitute a limitation of the computer, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0211] The memory 1220 can be used to store software programs and modules. The processor 1280 executes various functional applications and data processing of the computer by running the software programs and modules stored in the memory 1220. The memory 1220 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the computer (such as audio data, a phone book, etc.), etc. In addition, the memory 1220 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0212] The processor 1280 is the control center of the computer, which uses various interfaces and lines to connect various parts of the entire computer, and executes various functions of the computer and processes data by running or executing software programs and / or modules stored in the memory 1220, and calling data stored in the memory 1220. Optionally, the processor 1280 may include one or more processing units; preferably, the processor 1280 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 1280.

[0213] In the embodiment of the present application, the processor 1280 included in the terminal also has the following functions:

[0214] Generate multiple simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model; the resource transfer scenario model is trained based on virtual common orders, and the virtual common orders are constructed according to a first system dynamics model for indicating characteristics of common resource transfer orders;

[0215] Generate multiple simulated abnormal orders in the simulated resource transfer scenario through an abnormal resource transfer model; the abnormal resource transfer model is trained based on virtual abnormal orders, and the virtual abnormal orders are constructed according to a second system dynamics model for indicating characteristics of abnormal resource transfer orders;

[0216] Based on the simulated ordinary orders and the simulated abnormal orders, the order recognition model is trained to identify the order type in the simulated resource scenario; the order recognition model that meets the training end conditions is used to identify the type of real resource transfer orders in the actual resource transfer scenario.

[0217] In the embodiment of the present application, the processor 1280 included in the terminal also has the following functions:

[0218] Obtain target resource transfer order;

[0219] Determine the order type corresponding to the target resource transfer order through an order recognition model; the order recognition model is obtained by training using the model training method described in the above embodiments;

[0220] According to the order type corresponding to the target resource transfer order, a target operation is performed on the target resource transfer order.

[0221] Optionally, the processor 1280 is also used to execute the steps of any implementation of the model training method provided in the embodiments of the present application, or the processor 1280 is also used to execute the steps of any implementation of the order identification method provided in the embodiments of the present application.

[0222] See also Fig. 9 , Fig. 9 A schematic diagram of the structure of a server provided for an embodiment of the present application. The server 1300 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 1322 (for example, one or more processors) and a memory 1332, and one or more storage media 1330 (for example, one or more mass storage devices) storing application programs 1342 or data 1344. Among them, the memory 1332 and the storage medium 1330 may be short-term storage or permanent storage. The program stored in the storage medium 1330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Furthermore, the central processing unit 1322 can be configured to communicate with the storage medium 1330 to execute a series of instruction operations in the storage medium 1330 on the server 1300.

[0223] The server 1300 may also include one or more power supplies 1326, one or more wired or wireless network interfaces 1350, one or more input and output interfaces 1358, and / or one or more operating systems, such as Windows Server 2003. TM , Mac OS X TM , Unix TM,Linux TM , FreeBSD TM etc.

[0224] The steps performed by the server in the above embodiment can be based on the Fig. 9 The server structure shown.

[0225] The CPU 1322 is used to execute the following steps:

[0226] Generate multiple simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model; the resource transfer scenario model is trained based on virtual common orders, and the virtual common orders are constructed according to a first system dynamics model for indicating characteristics of common resource transfer orders;

[0227] Generate multiple simulated abnormal orders in the simulated resource transfer scenario through an abnormal resource transfer model; the abnormal resource transfer model is trained based on virtual abnormal orders, and the virtual abnormal orders are constructed according to a second system dynamics model for indicating characteristics of abnormal resource transfer orders;

[0228] Based on the simulated ordinary orders and the simulated abnormal orders, the order recognition model is trained to identify the order type in the simulated resource scenario; the order recognition model that meets the training end conditions is used to identify the type of real resource transfer orders in the actual resource transfer scenario.

[0229] CPU 1322 is also used to execute the following steps:

[0230] Obtain target resource transfer order;

[0231] Determine the order type corresponding to the target resource transfer order through an order recognition model; the order recognition model is obtained by training using the model training method described in the above embodiments;

[0232] According to the order type corresponding to the target resource transfer order, a target operation is performed on the target resource transfer order.

[0233] Optionally, CPU1322 can also be used to execute the steps of any implementation method of the model training method provided in the embodiments of the present application, or CPU1322 can also be used to execute the steps of any implementation method of the order identification method provided in the embodiments of the present application.

[0234] An embodiment of the present application also provides a computer-readable storage medium for storing a computer program, which is used to execute any implementation of a model training method described in the aforementioned embodiments, or the computer program is used to execute the steps of any implementation method of the order identification method provided in the embodiment of the present application.

[0235] The embodiment of the present application also provides a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. The processor of the computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes any one of the implementations of the model training method described in the above embodiments, or the computer device executes any one of the implementations of the order recognition method described in the above embodiments.

[0236] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0237] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0238] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0239] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0240] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store computer programs.

[0241] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0242] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A model training method, characterized in that: The method comprises: Generate multiple simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model; the resource transfer scenario model is trained based on virtual common orders, and the virtual common orders are constructed according to a first system dynamics model for indicating characteristics of real resource transfer orders; Generate multiple simulated abnormal orders in the simulated resource transfer scenario through an abnormal resource transfer model; the abnormal resource transfer model is trained based on virtual abnormal orders, and the virtual abnormal orders are constructed according to a second system dynamics model for indicating characteristics of abnormal resource transfer orders; Based on the simulated ordinary orders and the simulated abnormal orders, the order recognition model is trained to identify the order type in the simulated resource scenario; the order recognition model that meets the training end conditions is used to identify the type of real resource transfer orders in the actual resource transfer scenario.

2. The method according to claim 1, characterized in that The resource transfer scenario model is trained in the following way: Establishing the first system dynamics model; the first system dynamics model includes a first resource transfer process model and a first resource transfer variable model; Constructing a plurality of the virtual ordinary orders according to the first resource transfer process model and the first resource transfer variable model; The resource transfer scenario model is trained based on the virtual ordinary order so that the resource transfer scenario model learns the characteristics of ordinary resource transfer orders.

3. The method according to claim 2, characterized in that The step of constructing a plurality of virtual ordinary orders according to the first resource transfer process model and the first resource transfer variable model includes: Determine a virtual resource transfer node in the virtual ordinary order according to the resource transfer nodes included in the first resource transfer process model; Determining a virtual resource transfer variable in the virtual ordinary order according to the resource transfer variable information included in the first resource transfer variable model; The virtual ordinary order is constructed according to the virtual resource transfer node and the virtual resource transfer variable.

4. The method according to claim 1, characterized in that: The abnormal resource transfer model is trained in the following way: Based on the historical abnormal transaction orders, establishing the second system dynamics model; the second system dynamics model includes a second resource transfer process model and a second resource transfer variable model; Constructing a plurality of the virtual abnormal orders according to the second resource transfer process model and the second resource transfer variable model; The abnormal resource transfer model is trained based on the virtual abnormal order so that the abnormal resource transfer model learns the characteristics of the abnormal resource transfer order.

5. The method according to claim 4, characterized in that The step of constructing a plurality of virtual abnormal orders according to the second resource transfer process model and the second resource transfer variable model includes: Determine a virtual resource transfer node in the virtual abnormal order according to the resource transfer nodes included in the second resource transfer process model; Determining a virtual resource transfer variable in the virtual abnormal order according to the resource transfer variable information included in the second resource transfer variable model; The virtual abnormal order is constructed according to the virtual resource transfer node and the virtual resource transfer variable.

6. The method according to claim 1, characterized in that The training of the order recognition model to recognize the order type in the simulated resource scenario based on the simulated common order and the simulated abnormal order includes: Identify the prediction type corresponding to the training simulation order through the order recognition model; the training simulation order is the simulated ordinary order or the simulated abnormal order; The model parameters of the order recognition model are adjusted according to the predicted type corresponding to the training simulation order and the actual type of the training simulation order.

7. An order identification method, characterized in that: The method comprises: Obtain target resource transfer order; Determining the order type corresponding to the target resource transfer order through an order recognition model; the order recognition model is obtained by training using the method described in any one of claims 1 to 6; According to the order type corresponding to the target resource transfer order, a target operation is performed on the target resource transfer order.

8. The method according to claim 7, characterized in that The performing a target operation on the target resource transfer order according to the order type corresponding to the target resource transfer order includes: When the order type corresponding to the target resource transfer order indicates that the target resource transfer order is an abnormal resource transfer order, a target alarm operation is performed on a related object of the target resource transfer order.

9. A model training device, characterized in that: The device comprises: A first order generation module is used to generate a plurality of simulated common orders in a simulated resource transfer scenario through a resource transfer scenario model; the resource transfer scenario model is trained based on virtual common orders, and the virtual common orders are constructed according to a first system dynamics model for indicating characteristics of common resource transfer orders; A second order generation module is used to generate a plurality of simulated abnormal orders in the simulated resource transfer scenario through an abnormal resource transfer model; the abnormal resource transfer model is trained based on virtual abnormal orders, and the virtual abnormal orders are constructed according to a second system dynamics model for indicating characteristics of abnormal resource transfer orders; A model training module is used to train an order recognition model to recognize order types in the simulated resource scenario based on the simulated ordinary orders and the simulated abnormal orders; the order recognition model that meets the training end conditions is used to recognize the type of real resource transfer orders in actual resource transfer scenarios.

10. An order identification device, characterized in that: The device comprises: An order acquisition module is used to obtain target resource transfer orders; An order recognition model, used to determine the order type corresponding to the target resource transfer order through the order recognition model; the order recognition model is trained by the method according to any one of claims 1 to 6; The order processing module is used to perform a target operation on the target resource transfer order according to the order type corresponding to the target resource transfer order.

11. A computer device, characterized in that: The computer device includes a processor and a memory; The memory is used to store computer programs; The processor is used to execute the model training method described in any one of claims 1 to 6, or the order identification method described in any one of claims 7 to 8 according to the computer program.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store a computer program, and the computer program is used to execute the model training method described in any one of claims 1 to 6, or the order identification method described in any one of claims 7 to 8.

Images