Object category identification method and device, program product and storage medium

Abstract

The embodiment of the invention provides an object category identification method and device, a program product and a storage medium, and the method comprises the steps: obtaining a plurality of groups of target training data, each group of target training data comprising first data, second data and third data; multiple sets of target training data are used for training an initial comprehensive analysis model to obtain a target comprehensive analysis model, the initial comprehensive analysis model comprises a first component, a second component and a third component, and for each set of training data, the first component is used for generating behavior associated data of an object based on the first data and the second data, and the third component is used for generating behavior associated data of the object based on the second data; the second component is used for generating risk assessment data of the object based on the first data, the third component is used for generating object grouped data of the object based on the first data, and the initial comprehensive analysis model is used for outputting an object category of the object based on the behavior association data, the risk assessment data and the object grouped data.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computers, and more specifically, to a method and device for identifying an object category, a program product, and a storage medium. Background Art

[0002] As the digital transformation of the financial industry accelerates, corporate long-tail customers, as one of the important customer groups of banks, have a direct impact on the overall performance and competitiveness of banks. Corporate long-tail customers have become the focus of banks due to their large size, significant growth, and outstanding performance contribution.

[0003] However, due to the large number of customers and the scattered and complex data, traditional customer management methods face many challenges in customer identification, service efficiency and risk control. Therefore, the relevant technologies have the problems of low customer identification accuracy and limited risk control capabilities. Summary of the invention

[0004] The embodiments of the present application provide a method and device for identifying an object category, a program product, and a storage medium, so as to at least solve the problem of low accuracy in identifying a customer in the related art.

[0005] According to an embodiment of the present application, a method for identifying an object category is provided, comprising: acquiring multiple groups of target training data, wherein each group of target training data comprises: first data, second data and third data, the first data comprises object information and transaction information of the object, the second data is behavior information of the object, and the third data is category information of the object; using the multiple groups of target training data to train an initial comprehensive analysis model to obtain a target comprehensive analysis model, wherein the initial comprehensive analysis model comprises a first component, a second component and a third component, the first component is a component comprising an association analysis algorithm, the second component is a component comprising a classification algorithm, and the third component is a component comprising a clustering algorithm, and for each group of the training data, the first component is used to generate behavior association data of the object based on the first data and the second data, the second component is used to generate risk assessment data of the object based on the first data, and the third component is used to generate object grouping data of the object based on the first data, and the initial comprehensive analysis model is used to output the object category of the object based on the behavior association data, the risk assessment data and the object grouping data.

[0006] In an exemplary embodiment, before obtaining multiple groups of target training data, the method further includes: obtaining multiple groups of training data; performing data processing operations on the multiple groups of training data to obtain multiple groups of processed training data, wherein the data processing operations include removing abnormal data from the multiple groups of standardized training data after performing data standardization operations on the multiple groups of training data; performing feature extraction operations on the multiple groups of processed training data to obtain the multiple groups of target training data, wherein the feature extraction operations include extracting transaction features from the multiple groups of processed training data, and obtaining behavior change features of the objects based on time series, wherein the transaction features include at least one of the following: transaction frequency, transaction amount, transaction product features, and risk level features of transaction products.

[0007] In an exemplary embodiment, the initial comprehensive analysis model is trained using the above-mentioned multiple groups of target training data to obtain the target comprehensive analysis model, including: training the above-mentioned initial comprehensive analysis model for the kth time through the following steps, wherein k is a positive integer; obtaining the kth group of training data used for the kth training of the above-mentioned initial comprehensive analysis model from the above-mentioned multiple groups of target training data; inputting the first data and the above-mentioned second data in the above-mentioned kth group of training data into the above-mentioned first component to obtain the above-mentioned behavior association data output by the above-mentioned first component, wherein the above-mentioned first component includes multiple first sub-components, and the multiple first sub-components are respectively used to use different types of association analysis algorithms to perform behavior association on the above-mentioned first data and the above-mentioned second data. For association analysis, multiple association analysis results are obtained, or multiple first subcomponents are respectively used to use different analysis parameters in the same type of association analysis algorithm to perform behavioral association analysis on the first data and the second data to obtain multiple association analysis results, and the behavioral association data includes multiple association analysis results; the first data in the kth group of training data is input into the second component to obtain the risk assessment data output by the second component, wherein the second component includes multiple second subcomponents, and the multiple second subcomponents are respectively used to use different types of classification algorithms to perform risk assessment on the first data to obtain multiple risk assessment results, or multiple second subcomponents are respectively used to use the same type of association analysis algorithm Different classification parameters in a classification algorithm of a type are used to perform risk assessment on the first data to obtain multiple risk assessment results, and the risk assessment data includes multiple risk assessment results; the first data in the kth group of training data is input into the third component to obtain the object grouping data output by the third component, wherein the third component includes multiple third sub-components, and the multiple third sub-components are respectively used to use different types of clustering algorithms to perform grouping processing on the first data to obtain multiple grouping results, or the multiple third sub-components are respectively used to use different grouping parameters in the same type of clustering algorithm to perform grouping processing on the first data to obtain multiple grouping results, and the object grouping data Including multiple grouping results; based on the behavior association data, the risk assessment data and the object grouping data, determining the object category output by the initial comprehensive analysis model of the k-th training; determining the loss value output by the loss function of the initial comprehensive analysis model of the k-th training according to the object category and the known category information of the k-th training included in the third data; when the loss value output by the loss function of the k-th training satisfies the preset loss value, ending the training to obtain the target comprehensive analysis model; when the loss value output by the loss function of the k-th training does not satisfy the preset loss value, adjusting the values ​​of the parameters in the initial comprehensive analysis model of the k-th training;The initial comprehensive analysis model trained for the kth time is used for the k+1th training until the loss value output by the loss function satisfies the preset loss value, and the training is terminated to obtain the target comprehensive analysis model. ;

[0008] In an exemplary embodiment, based on the above-mentioned behavior association data, the above-mentioned risk assessment data and the above-mentioned object grouping data, the object category output by the above-mentioned initial comprehensive analysis model trained for the kth time is determined, including: according to the voting mechanism in the above-mentioned initial comprehensive analysis model, a target association analysis result is determined from the above-mentioned behavior association data including multiple association analysis results, a target risk assessment result is determined from the above-mentioned risk assessment data including multiple risk assessment results, and a target grouping result is determined from the above-mentioned object grouping data including multiple grouping results; based on the above-mentioned target association analysis result, the above-mentioned target risk assessment result and the above-mentioned target grouping result, the above-mentioned object category is generated.

[0009] In an exemplary embodiment, when the loss value output by the loss function of the k-th training does not meet the preset loss value, the values ​​of the parameters in the initial comprehensive analysis model of the k-th training are adjusted, including: when the loss value output by the loss function of the k-th training does not meet the preset loss value, the voting mechanism is adjusted to obtain the initial comprehensive analysis model of the k-th training.

[0010] In an exemplary embodiment, after the initial comprehensive analysis model is trained using the above-mentioned multiple groups of target training data to obtain the target comprehensive analysis model, the above-mentioned method further includes: outputting the target data to the above-mentioned target comprehensive analysis model to obtain the target object category output by the above-mentioned target comprehensive analysis model, wherein the above-mentioned target data includes the target object information and target behavior information of the above-mentioned target object; generating a target service plan based on the above-mentioned target object category; wherein the target service plan generated based on the above-mentioned target object category includes: obtaining the target risk indicators of the above-mentioned target object in real time according to the above-mentioned target object category and through the risk warning model, wherein the above-mentioned target risk indicators include at least one of the following: financial indicators, transaction behavior indicators, credit risk indicators, and industry risk indicators; in the case where the above-mentioned target risk indicators include abnormal indicators, generating the above-mentioned target service plan, wherein the above-mentioned target service plan is used to perform warning operations and risk control operations on the above-mentioned target object, and the above-mentioned risk control operations include at least one of the following: adjusting the credit limit of the above-mentioned target object and adjusting the intensity of credit monitoring on the above-mentioned target object.

[0011] According to another embodiment of the present application, an object category recognition device is provided, including: an acquisition module, used to acquire multiple groups of target training data, wherein each group of target training data includes: first data, second data and third data, the first data includes object information and transaction information of the object, the second data is behavior information of the object, and the third data is category information of the object; a training module, used to train an initial comprehensive analysis model using the multiple groups of target training data to obtain a target comprehensive analysis model, wherein the initial comprehensive analysis model includes a first component, a second component and a third component, the first component is a component including an association analysis algorithm, the second component is a component including a classification algorithm, and the third component is a component including a clustering algorithm, for each group of the training data, the first component is used to generate behavior association data of the object based on the first data and the second data, the second component is used to generate risk assessment data of the object based on the first data, and the third component is used to generate object grouping data of the object based on the first data, and the initial comprehensive analysis model is used to output the object category of the object based on the behavior association data, the risk assessment data and the object grouping data.

[0012] According to another embodiment of the present application, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0013] According to another embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any one of the above method embodiments when running.

[0014] According to another embodiment of the present application, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0015] Through the present application, after obtaining multiple groups of target training data, by inputting the multiple groups of target training data into different algorithms, and using the multiple groups of target training data to input into different parameters of the same type of algorithm, the behavior association data, risk assessment data and object grouping data of the object are generated, and then the category of the object is further determined, so as to realize the fusion of the algorithm results of multiple different algorithms or different parameters of the same algorithm. After that, the category of the object is compared with the category information of the existing object, and it is further determined that when the loss value output by the loss function meets the preset loss value, the training is terminated to obtain the above-mentioned target comprehensive analysis model. Therefore, the problem of low customer recognition accuracy in the related technology can be solved, and the effect of accurately identifying customer categories and improving the service quality for customers can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic diagram of a hardware environment of an object category recognition method according to an embodiment of the present application;

[0017] Figure 2 is a flow chart of a method for identifying an object category according to an embodiment of the present application;

[0018] Figure 3 is a structural block diagram of an initial comprehensive analysis model according to an embodiment of the present application;

[0019] Figure 4 is a structural schematic diagram of an object category recognition device according to a specific embodiment of the present application;

[0020] Figure 5 It is a structural block diagram of an object category recognition device according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] The embodiments of the present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0022] It should be noted that the terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0023] The method embodiments provided in the embodiments of the present application can be executed in a server device or a similar computing device. Taking running on a server device as an example, Figure 1 FIG. 1 is a schematic diagram of a hardware environment of an object category recognition method according to an embodiment of the present application. Figure 1 As shown, the server device may include one or more ( Figure 1Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the server device may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above server device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0024] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as a computer program corresponding to a method for identifying an object category in an embodiment of the present application. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to a server device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0025] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the server device. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0026] In this embodiment, a method for identifying an object category is provided. Figure 2 is a flow chart of a method for identifying an object category according to an embodiment of the present application, such as Figure 2 As shown, the process includes the following steps:

[0027] Step S202, obtaining multiple groups of target training data, wherein each group of target training data includes: first data, second data and third data, the first data includes object information and transaction information of the object, the second data is behavior information of the object, and the third data is category information of the object;

[0028] Step S204, using multiple groups of target training data to train the initial comprehensive analysis model to obtain a target comprehensive analysis model, wherein the initial comprehensive analysis model includes a first component, a second component and a third component, the first component is a component including an association analysis algorithm, the second component is a component including a classification algorithm, and the third component is a component including a clustering algorithm. For each group of training data, the first component is used to generate behavior association data of the object based on the first data and the second data, the second component is used to generate risk assessment data of the object based on the first data, and the third component is used to generate object grouping data of the object based on the first data. The initial comprehensive analysis model is used to output the object category of the object based on the behavior association data, the risk assessment data and the object grouping data.

[0029] Optionally, in addition to the basic object information and transaction information of the object, the behavior information of the object, and the category of the object, the multiple groups of target data in this embodiment may also include the object's social media information, external credit information, etc., for example, the object's Weibo credit points, Sesame Credit points, etc.

[0030] Optionally, the second data in the target training data in the present application is the behavior information of the object, and the second data includes but is not limited to: the object's behavior of purchasing goods, the object's product usage behavior, and the object's search behavior on the platform.

[0031] Optionally, in this embodiment, the first component uses an association analysis algorithm to mine the associations between different behaviors of the object, wherein the association analysis algorithm includes but is not limited to: Eclat algorithm, Apriori algorithm, FP-Growth algorithm, and association rule discovery algorithm.

[0032] Optionally, the second component in this embodiment uses a classification algorithm to determine whether the object belongs to a specific category (such as a high-risk customer, a high-quality customer, etc.), and further determines risk assessment data such as the object's churn risk, wherein the classification algorithm includes but is not limited to: decision tree, random forest, gradient boosting tree.

[0033] Optionally, in this embodiment, when predicting more complex object behavior patterns and potential values, the second component also needs to introduce deep learning algorithms, for example, citing recurrent neural networks and long short-term memory networks, to capture the temporal dependencies and nonlinear relationships in object behaviors, and further improve the accuracy of determining the loss risk of objects.

[0034] Optionally, the third component in this embodiment uses a clustering algorithm to divide objects into different groups. In order to better understand the characteristics and behavior patterns of the object groups, clustering algorithms include but are not limited to: hierarchical clustering, K-means algorithm, DBSCAN algorithm.

[0035] Optionally, when training the target comprehensive analysis model, this embodiment can combine multiple financial institutions to conduct joint training through federated learning technology while protecting the privacy of the object, so as to achieve the purpose of sharing model parameters without sharing the original data, and can further improve the generalization ability and accuracy of the target comprehensive analysis model.

[0036] Through this embodiment, after obtaining multiple groups of target training data, the multiple groups of target training data are input into different components to generate object behavior association data, risk assessment data and object grouping data, and the category of the object is further determined based on the above three types of data, and the target comprehensive analysis model is further determined. Therefore, the problem of low customer recognition accuracy in related technologies can be solved, and the effect of accurately identifying customer categories and improving customer service quality can be achieved.

[0037] In an exemplary embodiment, before obtaining multiple groups of target training data, the method further includes: obtaining multiple groups of training data; performing data processing operations on the multiple groups of training data to obtain multiple groups of processed training data, wherein the data processing operations include removing abnormal data from the multiple groups of standardized training data after performing data standardization operations on the multiple groups of training data; performing feature extraction operations on the multiple groups of processed training data to obtain the multiple groups of target training data, wherein the feature extraction operations include extracting transaction features from the multiple groups of processed training data, and obtaining behavior change features of the objects based on time series, wherein the transaction features include at least one of the following: transaction frequency, transaction amount, transaction product features, and risk level features of transaction products.

[0038] Optionally, the multiple sets of training data in this embodiment include but are not limited to those obtained from databases in multiple financial customer systems and information disclosed in the customer's social media.

[0039] Optionally, in addition to the basic object information and transaction information of the object, the behavior information of the object, and the category of the object, the multiple sets of training data in this embodiment may also include the social media data of the object, external credit information, etc. For example, the object information of the object includes but is not limited to: name, gender, date of birth, contact information (telephone, email, address). The transaction information of the object generally refers to the transaction records and data related to a specific object (e.g., individual, company, product, service, etc.) in finance, business, investment or other fields. This information may include but is not limited to: transaction date and time, transaction parties, transaction amount, transaction type, transaction conditions. The behavior information of the object refers to data on the purchase behavior, preferences, habits, etc. of the object, which is of great significance for enterprises to understand the needs of the object, optimize product services, and improve the satisfaction and loyalty of the object. The behavior information of the object includes but is not limited to: purchase history, purchase frequency, purchase amount, purchase channel, product preference, price sensitivity. The category of the object refers to the classification of the object according to certain characteristics of the object, so as to more effectively perform market segmentation and personalized services. Methods of classifying objects include, but are not limited to: classifying according to the objects’ purchasing behavior, usage habits, brand loyalty and other behavioral characteristics; classifying according to the stage of the objects in the product life cycle, such as new customers, mature customers, declining customers, etc.; classifying according to the objects’ income level, consumption capacity and purchasing power; classifying according to the objects’ purchasing motivations, such as price-sensitive, quality-sensitive, brand-loyal, etc.; classifying according to the objects’ preferred purchasing channels, such as online shoppers, offline shoppers, etc. External credit information includes, but is not limited to: the objects’ Weibo credit points, Sesame Credit points, etc.

[0040] Optionally, the operation of removing abnormal data in the data processing operation performed in this embodiment includes but is not limited to removing duplicate, erroneous, and missing data. When it is detected that the acquired multiple sets of training data include missing data, the information of this object is first obtained from other platforms to complete the missing data, and the missing data is removed after confirming that the missing data cannot be completed.

[0041] Optionally, in this embodiment, when performing the feature extraction operation, the extracted features are transaction features of the object, which are features that can reflect the financial qualifications of the object. The key transaction features include, but are not limited to, transaction frequency, transaction amount, transaction product features, and risk level features of transaction products. Among them, transaction product features include, but are not limited to, the types of transaction products, the channels for product transactions, and the time of transactions.

[0042] Optionally, in this embodiment, when performing feature extraction, a time series is also introduced to obtain the behavior change characteristics of the object, wherein the time series is the time when the behavior information of the object is obtained synchronously.

[0043] Through this embodiment, data processing operations are performed on the multiple sets of training data obtained, including removing duplicate, erroneous and information-missing data, and performing feature extraction operations on the processed data, thereby screening the multiple sets of training data obtained, improving the quality and accuracy of the training data, and ensuring the reliability of subsequent modeling.

[0044] In an exemplary embodiment, the initial comprehensive analysis model is trained using the above-mentioned multiple groups of target training data to obtain the target comprehensive analysis model, including: training the above-mentioned initial comprehensive analysis model for the kth time through the following steps, wherein k is a positive integer; obtaining the kth group of training data used for the kth training of the above-mentioned initial comprehensive analysis model from the above-mentioned multiple groups of target training data; inputting the first data and the above-mentioned second data in the above-mentioned kth group of training data into the above-mentioned first component to obtain the above-mentioned behavior association data output by the above-mentioned first component, wherein the above-mentioned first component includes multiple first sub-components, and the multiple first sub-components are respectively used to use different types of association analysis algorithms to perform behavior association on the above-mentioned first data and the above-mentioned second data. For association analysis, multiple association analysis results are obtained, or multiple first subcomponents are respectively used to use different analysis parameters in the same type of association analysis algorithm to perform behavioral association analysis on the first data and the second data to obtain multiple association analysis results, and the behavioral association data includes multiple association analysis results; the first data in the kth group of training data is input into the second component to obtain the risk assessment data output by the second component, wherein the second component includes multiple second subcomponents, and the multiple second subcomponents are respectively used to use different types of classification algorithms to perform risk assessment on the first data to obtain multiple risk assessment results, or multiple second subcomponents are respectively used to use the same type of association analysis algorithm Different classification parameters in a classification algorithm of a type are used to perform risk assessment on the first data to obtain multiple risk assessment results, and the risk assessment data includes multiple risk assessment results; the first data in the kth group of training data is input into the third component to obtain the object grouping data output by the third component, wherein the third component includes multiple third sub-components, and the multiple third sub-components are respectively used to use different types of clustering algorithms to perform grouping processing on the first data to obtain multiple grouping results, or the multiple third sub-components are respectively used to use different grouping parameters in the same type of clustering algorithm to perform grouping processing on the first data to obtain multiple grouping results, and the object grouping data Including multiple grouping results; based on the behavior association data, the risk assessment data and the object grouping data, determining the object category output by the initial comprehensive analysis model of the k-th training; determining the loss value output by the loss function of the initial comprehensive analysis model of the k-th training according to the object category and the known category information of the k-th training included in the third data; when the loss value output by the loss function of the k-th training satisfies the preset loss value, ending the training to obtain the target comprehensive analysis model; when the loss value output by the loss function of the k-th training does not satisfy the preset loss value, adjusting the values ​​of the parameters in the initial comprehensive analysis model of the k-th training;The initial comprehensive analysis model trained for the kth time is used for the k+1th training until the loss value output by the loss function satisfies the preset loss value, and the training is terminated to obtain the target comprehensive analysis model. ;

[0045] Optionally, the association analysis results in this embodiment are used to represent the behavioral correlation between the first data and the second data of the object, wherein the second data is the behavioral information of the object, and the second data includes but is not limited to: the object's purchase behavior, the object's product usage behavior, and the object's search behavior on the platform.

[0046] For example, in a specific embodiment, the first data and the second data in the third set of training data are input into the first component. If the second data is the product usage behavior of the object and the search behavior of the object on the platform, the first component uses the association analysis algorithm to analyze the behavioral correlation between the product usage behavior of the object in the first data and the second data to obtain the first sub-result. At the same time, the first component uses the association analysis algorithm to analyze the behavioral correlation between the search behavior of the object in the first data and the second data on the platform to obtain the second sub-result, and performs a merge operation on the first sub-result and the second sub-result to obtain the association analysis result; wherein the first sub-result and the second sub-result can be displayed in the form of percentages, and the merge operation includes but is not limited to an average operation, a weighted operation, etc.

[0047] Optionally, the first data in this embodiment is object information and transaction information of the object, and the transaction information includes but is not limited to: transaction frequency, transaction amount, transaction product characteristics, and risk level characteristics of the transaction product.

[0048] For example, in a specific embodiment, the first data in the third set of training data is input into the second component to obtain risk assessment data output by the second component. If the transaction information in the first data includes the transaction frequency and transaction amount of the object and the characteristics of the transaction product, the second component uses a classification algorithm to analyze the first data to obtain the degree of loss risk of the object, and further outputs the risk assessment data of the object.

[0049] For example, in a specific embodiment, the first data in the third group of training data is input into the third component to obtain object grouping data output by the third component. If the transaction information in the first data includes the transaction frequency and transaction amount of the object and the characteristics of the transaction product, the second component uses a clustering algorithm to analyze the first data, obtains the object grouping result, and outputs it.

[0050] Optionally, the analysis results output by the third component in this embodiment include but are not limited to: the object belongs to an individual customer, a corporate customer, a high net worth customer, a credit risk customer, a clean customer with no credit record or financial history, a private banking customer, a retail customer, etc.

[0051] Optionally, in this embodiment, the object categories output by the initial comprehensive analysis model include but are not limited to: high-value customers, potential customers, and risky customers.

[0052] Optionally, in this embodiment, after multiple sets of training data are input into the first component, the second component and the third component, the above components are executed simultaneously.

[0053] Through this embodiment, after obtaining multiple groups of target training data, by inputting the first data and the second data in the multiple groups of target training data into the first component, inputting the first data in the multiple groups of target training data into the second component, and inputting the first data in the multiple groups of target training data into the third component, the behavior association data, risk assessment data and object grouping data of the object are generated respectively, and then the category of the object is further determined, so as to realize the fusion of the algorithm results of multiple different algorithms or different parameters of the same algorithm, and then the category of the object is compared with the category information of the existing object, and it is further determined that when the loss value output by the loss function meets the preset loss value, the training is terminated to obtain the above-mentioned target comprehensive analysis model, which solves the problem of insufficient recognition accuracy of customers in related technologies and achieves the effect of accurately identifying customer types.

[0054] In an exemplary embodiment, based on the above-mentioned behavior association data, the above-mentioned risk assessment data and the above-mentioned object grouping data, the object category output by the above-mentioned initial comprehensive analysis model trained for the kth time is determined, including: according to the voting mechanism in the above-mentioned initial comprehensive analysis model, a target association analysis result is determined from the above-mentioned behavior association data including multiple association analysis results, a target risk assessment result is determined from the above-mentioned risk assessment data including multiple risk assessment results, and a target grouping result is determined from the above-mentioned object grouping data including multiple grouping results; based on the above-mentioned target association analysis result, the above-mentioned target risk assessment result and the above-mentioned target grouping result, the above-mentioned object category is generated.

[0055] Optionally, the initial comprehensive analysis model in this embodiment is mainly a mathematical or computational model used to understand, predict, optimize or explain the above object data. The initial comprehensive analysis model includes but is not limited to: a regression model, a classification model, a machine learning model, and an optimization model. For example, when the initial comprehensive analysis model is a machine learning model, such as Figure 3As shown, it is a structural block diagram of the initial comprehensive analysis model in this embodiment. The initial comprehensive analysis model includes base learner 1, base learner 2, and base learner 3 (corresponding to the first component, the second component and the third component mentioned above, respectively). Metadata (corresponding to the multiple sets of training data mentioned above) are input into base learner 1, base learner 2, and base learner 3, respectively, and the three output results are input into the meta learner. The voting mechanism in the meta learner selects the final result of each component from the results obtained by the three components, and generates the final prediction data, that is, the object category in this embodiment.

[0056] Optionally, the voting mechanism in this embodiment is hard voting, in which voting is performed according to preset voting rules, and the result with the largest number of votes is determined as the final result determined by the corresponding component, wherein the preset voting rules are determined based on the accuracy of the historical output results of the algorithm used by the corresponding component.

[0057] Through this embodiment, a more appropriate result is selected from the results output by different components using a voting mechanism, and then the category of the object is determined by the output results of the first component, the second component, and the third component. This makes the determination of the object category more rigorous and achieves the technical effect of accurately outputting the object category.

[0058] In an exemplary embodiment, when the loss value output by the loss function of the k-th training does not meet the preset loss value, the values ​​of the parameters in the initial comprehensive analysis model of the k-th training are adjusted, including: when the loss value output by the loss function of the k-th training does not meet the preset loss value, the voting mechanism is adjusted to obtain the initial comprehensive analysis model of the k-th training.

[0059] Optionally, in this embodiment, when the loss value output by the loss function of the kth training does not meet the preset loss value, the subcomponent that outputs this result is confirmed based on the result selected by the voting rule, and the parameters of the algorithm used in the subcomponent are adjusted. When the result output by a certain subcomponent increases the loss value output by the loss function multiple times, the voting mechanism is adjusted and this subcomponent is no longer selected for use.

[0060] Through this embodiment, when the loss value output by the loss function of the kth training does not meet the preset loss value, the voting mechanism is adjusted to further select a more suitable algorithm and its corresponding parameters, thereby achieving the technical effect of accurately identifying the object type.

[0061] In an exemplary embodiment, after the initial comprehensive analysis model is trained using the above-mentioned multiple groups of target training data to obtain the target comprehensive analysis model, the above-mentioned method further includes: outputting the target data to the above-mentioned target comprehensive analysis model to obtain the target object category output by the above-mentioned target comprehensive analysis model, wherein the above-mentioned target data includes the target object information and target behavior information of the above-mentioned target object; generating a target service plan based on the above-mentioned target object category; wherein the target service plan generated based on the above-mentioned target object category includes: obtaining the target risk indicators of the above-mentioned target object in real time according to the above-mentioned target object category and through the risk warning model, wherein the above-mentioned target risk indicators include at least one of the following: financial indicators, transaction behavior indicators, credit risk indicators, and industry risk indicators; in the case where the above-mentioned target risk indicators include abnormal indicators, generating the above-mentioned target service plan, wherein the above-mentioned target service plan is used to perform warning operations and risk control operations on the above-mentioned target object, and the above-mentioned risk control operations include at least one of the following: adjusting the credit limit of the above-mentioned target object and adjusting the intensity of credit monitoring on the above-mentioned target object.

[0062] Optionally, the target risk indicators in this embodiment include at least one of the following: financial indicators, transaction behavior indicators, credit risk indicators, and industry risk indicators; among which, financial indicators include debt-to-asset ratio, current ratio, quick ratio and other data used to evaluate the debt repayment ability of the object; transaction behavior indicators include abnormal transaction frequency, large capital flow, cross-border transactions and other data used to identify potential fraud risks; credit risk indicators include data based on historical credit records, repayment status, etc., which are used to evaluate the credit rating and default risk of the object; industry risk indicators include market dynamics, policy changes, competitive situation, etc. of the industry in which the object of attention is located, which are used to evaluate the impact of industry risks on the operation of this object.

[0063] Optionally, the risk warning model used in this embodiment may utilize data mining and machine learning techniques, including but not limited to using reinforcement learning algorithms such as Q-learning and Deep Q Network.

[0064] Optionally, in this embodiment, the target service solutions generated based on the target object category include but are not limited to personalized service solutions and financial solutions, wherein the personalized service solutions include but are not limited to customized product recommendations, exclusive account manager services, customized credit policies, etc., to meet the diverse needs of the objects.

[0065] Optionally, if the target object category is a high-value object, the corresponding target service plan can be an exclusive financial solution, such as customized financial products, priority loan approval channels, exclusive object manager services, etc.

[0066] Optionally, if the target object category is a potential object, the corresponding target service plan can be to provide targeted market information and product recommendations, and at the same time establish a growth incentive plan, such as providing low-cost financing support, business consulting and training, etc., to help the target object grow rapidly.

[0067] Optionally, the early warning operation in this embodiment will classify the early warning into different levels (such as red, orange, yellow, etc.) according to the risk level and urgency of the abnormal indicators, so as to take corresponding response measures. Based on this, a detailed emergency response process is formulated to clarify the response measures, responsible persons and processing time limits for early warnings at each level to ensure that risks are controlled in a timely and effective manner.

[0068] Optionally, in this embodiment, a feedback mechanism can be constructed according to the target service plan to regularly evaluate the implementation effect of the business strategy. Through data comparison and other methods, effective strategies can be identified and continuously optimized, while invalid or inefficient strategies can be eliminated to ensure continuous optimization and iteration of business management.

[0069] In a specific embodiment, Figure 4 As shown, it is a structural schematic diagram of an object category identification device according to a specific embodiment of the present application, and the object category identification device includes a data collection and preprocessing module, a data mining and model building module, and a business strategy customization and optimization module. After the data collection and preprocessing module obtains multiple groups of data, it will perform data cleaning and feature extraction operations on them to improve the validity of the data, and then input the obtained feature data into the data mining and model building module. The data mining and model building module uses the association analysis algorithm, classification algorithm and clustering algorithm respectively included in the three components to generate the behavior association data, risk assessment data and grouping data of the object respectively, and then selects the result of the corresponding component output according to the voting mechanism in the model fusion, obtains the result of the target comprehensive analysis model output, and outputs the object category of the object according to the above result; then the business strategy customization and optimization module selects the corresponding service plan based on the category of the object, regularly evaluates and continuously optimizes the service plan based on the results of the implementation of the plan, and uses the risk warning model to monitor the risk indicators of the object in real time.

[0070] It should be noted that, through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0071] In the present embodiment, a device for identifying an object category is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware is also possible and conceived.

[0072] Figure 5 is a structural block diagram of an object category recognition device according to an embodiment of the present application, such as Figure 5 As shown, the device comprises:

[0073] An acquisition module 52 is used to acquire multiple sets of target training data, wherein each set of target training data includes: first data, second data and third data, the first data includes object information and transaction information of the object, the second data is behavior information of the object, and the third data is category information of the object;

[0074] The training module 54 is used to train the initial comprehensive analysis model using the above-mentioned multiple groups of target training data to obtain the target comprehensive analysis model, wherein the above-mentioned initial comprehensive analysis model includes a first component, a second component and a third component, the above-mentioned first component is a component including an association analysis algorithm, the above-mentioned second component is a component including a classification algorithm, and the above-mentioned third component is a component including a clustering algorithm. For each group of the above-mentioned training data, the above-mentioned first component is used to generate the behavior association data of the above-mentioned object based on the above-mentioned first data and the above-mentioned second data, the above-mentioned second component is used to generate the risk assessment data of the above-mentioned object based on the above-mentioned first data, and the above-mentioned third component is used to generate the object grouping data of the above-mentioned object based on the above-mentioned first data. The above-mentioned initial comprehensive analysis model is used to output the object category of the above-mentioned object based on the above-mentioned behavior association data, the above-mentioned risk assessment data and the above-mentioned object grouping data.

[0075] In an exemplary embodiment, the apparatus is further used to obtain multiple sets of training data; perform data processing operations on the multiple sets of training data to obtain multiple sets of processed training data, wherein the data processing operations include removing abnormal data from the multiple sets of standardized training data after performing data standardization operations on the multiple sets of training data; perform feature extraction operations on the multiple sets of processed training data to obtain the multiple sets of target training data, wherein the feature extraction operations include extracting transaction features from the multiple sets of processed training data, and obtaining behavior change features of the objects based on time series, wherein the transaction features include at least one of the following: transaction frequency, transaction amount, transaction product features, and risk level features of transaction products.

[0076] In an exemplary embodiment, the above-mentioned device is also used to train the above-mentioned initial comprehensive analysis model for the kth time through the following steps, wherein k is a positive integer; obtaining the kth group of training data used for the kth training of the above-mentioned initial comprehensive analysis model from the above-mentioned multiple groups of target training data; inputting the first data and the above-mentioned second data in the above-mentioned kth group of training data into the above-mentioned first component to obtain the above-mentioned behavior association data output by the above-mentioned first component, wherein the above-mentioned first component includes multiple first sub-components, and the multiple first sub-components are respectively used to use different types of association analysis algorithms to perform behavior association analysis on the above-mentioned first data and the above-mentioned second data to obtain multiple association analysis results, or multiple above-mentioned first sub-components. A subcomponent is used to use different analysis parameters in the same type of association analysis algorithm to perform behavioral association analysis on the first data and the second data to obtain multiple association analysis results, and the behavioral association data includes multiple association analysis results; the first data in the kth group of training data is input into the second component to obtain the risk assessment data output by the second component, wherein the second component includes multiple second subcomponents, and the multiple second subcomponents are used to use different types of classification algorithms to perform risk assessment on the first data to obtain multiple risk assessment results, or the multiple second subcomponents are used to use different classification parameters in the same type of classification algorithm. Perform risk assessment on the first data to obtain multiple risk assessment results, and the risk assessment data includes multiple risk assessment results; input the first data in the k-th group of training data into the third component to obtain the object grouping data output by the third component, wherein the third component includes multiple third subcomponents, and the multiple third subcomponents are respectively used to use different types of clustering algorithms to perform grouping processing on the first data to obtain multiple grouping results, or the multiple third subcomponents are respectively used to use different grouping parameters in the same type of clustering algorithm to perform grouping processing on the first data to obtain multiple grouping results, and the object grouping data includes multiple sub Group results; based on the above-mentioned behavior association data, the above-mentioned risk assessment data and the above-mentioned object grouping data, determine the object category output by the above-mentioned initial comprehensive analysis model of the k-th training; according to the above-mentioned object category and the category information of the above-mentioned k-th training included in the known above-mentioned third data, determine the loss value output by the loss function of the above-mentioned initial comprehensive analysis model of the k-th training; when the loss value output by the above-mentioned loss function of the above-mentioned k-th training meets the preset loss value, end the training to obtain the above-mentioned target comprehensive analysis model; when the loss value output by the above-mentioned loss function of the above-mentioned k-th training does not meet the preset loss value, adjust the values ​​of the parameters in the above-mentioned initial comprehensive analysis model of the k-th training;The initial comprehensive analysis model trained for the kth time is used for the k+1th training until the loss value output by the loss function satisfies the preset loss value, and the training is terminated to obtain the target comprehensive analysis model. ;

[0077] In an exemplary embodiment, the above-mentioned device is also used to determine the object category output by the above-mentioned initial comprehensive analysis model trained for the kth time based on the above-mentioned behavior association data, the above-mentioned risk assessment data and the above-mentioned object grouping data, including: determining the target association analysis result from the above-mentioned behavior association data including multiple association analysis results, determining the target risk assessment result from the above-mentioned risk assessment data including multiple risk assessment results, and determining the target grouping result from the above-mentioned object grouping data including multiple grouping results according to the voting mechanism in the above-mentioned initial comprehensive analysis model; generating the above-mentioned object category based on the above-mentioned target association analysis result, the above-mentioned target risk assessment result and the above-mentioned target grouping result.

[0078] In an exemplary embodiment, the above-mentioned device is also used to adjust the values ​​of parameters in the initial comprehensive analysis model of the above-mentioned k-th training when the loss value output by the above-mentioned loss function of the above-mentioned k-th training does not meet the preset loss value, including: when the loss value output by the above-mentioned loss function of the above-mentioned k-th training does not meet the preset loss value, adjust the above-mentioned voting mechanism to obtain the initial comprehensive analysis model of the above-mentioned k-th training.

[0079] In an exemplary embodiment, the above-mentioned device is also used to train the initial comprehensive analysis model using the above-mentioned multiple groups of target training data. After obtaining the target comprehensive analysis model, the above-mentioned method also includes: outputting the target data to the above-mentioned target comprehensive analysis model to obtain the target object category output by the above-mentioned target comprehensive analysis model, wherein the above-mentioned target data includes the target object information and target behavior information of the above-mentioned target object; generating a target service plan based on the above-mentioned target object category; wherein, generating the target service plan based on the above-mentioned target object category includes: obtaining the target risk indicators of the above-mentioned target object in real time according to the above-mentioned target object category and through the risk warning model, wherein the above-mentioned target risk indicators include at least one of the following: financial indicators, transaction behavior indicators, credit risk indicators, industry risk indicators; in the case where the above-mentioned target risk indicators include abnormal indicators, generating the above-mentioned target service plan, wherein the above-mentioned target service plan is used to perform warning operations and risk control operations on the above-mentioned target object, and the above-mentioned risk control operations include at least one of the following: adjusting the credit limit of the above-mentioned target object, and adjusting the intensity of credit monitoring on the above-mentioned target object.

[0080] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0081] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0082] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0083] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0084] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0085] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any one of the above method embodiments are implemented.

[0086] An embodiment of the present application further provides another computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0087] The embodiments of the present application also provide a computer program, which includes computer instructions, which are stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps of any one of the above method embodiments.

[0088] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0089] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0090] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for identifying an object category, characterized in that: include: Acquire multiple groups of target training data, wherein each group of the target training data includes: first data, second data, and third data, the first data includes object information and transaction information of an object, the second data is behavior information of the object, and the third data is category information of the object; The initial comprehensive analysis model is trained using the multiple groups of target training data to obtain a target comprehensive analysis model, wherein the initial comprehensive analysis model includes a first component, a second component and a third component, the first component is a component including an association analysis algorithm, the second component is a component including a classification algorithm, and the third component is a component including a clustering algorithm. For each group of the training data, the first component is used to generate behavior association data of the object based on the first data and the second data, the second component is used to generate risk assessment data of the object based on the first data, and the third component is used to generate object grouping data of the object based on the first data. The initial comprehensive analysis model is used to output the object category of the object based on the behavior association data, the risk assessment data and the object grouping data.

2. The method according to claim 1, characterized in that Before acquiring multiple sets of target training data, the method further includes: Obtain multiple sets of training data; Performing data processing operations on the multiple sets of training data to obtain multiple sets of processed training data, wherein the data processing operations include, after performing data standardization operations on the multiple sets of training data, removing abnormal data from the multiple sets of standardized training data; Perform a feature extraction operation on the multiple groups of processed training data to obtain the multiple groups of target training data, wherein the feature extraction operation includes extracting transaction features from the multiple groups of processed training data, and obtaining behavior change features of the object based on a time series, and the transaction features include at least one of the following: transaction frequency, transaction amount, transaction product features, and risk level features of transaction products.

3. The method according to claim 1, characterized in that The initial comprehensive analysis model is trained using the multiple sets of target training data to obtain a target comprehensive analysis model, including: The initial comprehensive analysis model is trained for the kth time by the following steps, wherein k is a positive integer; Acquire a kth group of training data from the multiple groups of target training data for use in performing a kth training on the initial comprehensive analysis model; Inputting the first data and the second data in the kth group of training data into the first component to obtain the behavior association data output by the first component, wherein the first component includes a plurality of first subcomponents, and the plurality of first subcomponents are respectively used to perform behavior association analysis on the first data and the second data using different types of association analysis algorithms to obtain a plurality of association analysis results, or the plurality of first subcomponents are respectively used to perform behavior association analysis on the first data and the second data using different analysis parameters in the same type of association analysis algorithm to obtain a plurality of association analysis results, and the behavior association data includes a plurality of the association analysis results; Inputting the first data in the k-th group of training data into the second component to obtain the risk assessment data output by the second component, wherein the second component includes a plurality of second subcomponents, and the plurality of second subcomponents are respectively used to perform risk assessment on the first data using different types of classification algorithms to obtain a plurality of risk assessment results, or the plurality of second subcomponents are respectively used to perform risk assessment on the first data using different classification parameters in the same type of classification algorithm to obtain a plurality of risk assessment results, and the risk assessment data includes a plurality of the risk assessment results; Input the first data in the k-th group of training data into the third component to obtain the object grouping data output by the third component, wherein the third component includes a plurality of third subcomponents, and the plurality of third subcomponents are respectively used to perform grouping processing on the first data using different types of clustering algorithms to obtain a plurality of grouping results, or the plurality of third subcomponents are respectively used to perform grouping processing on the first data using different grouping parameters in the same type of clustering algorithm to obtain a plurality of grouping results, and the object grouping data includes a plurality of the grouping results; Determining the object category output by the initial comprehensive analysis model of the k-th training based on the behavior association data, the risk assessment data and the object grouping data; Determine a loss value output by a loss function of the initial comprehensive analysis model of the kth training according to the object category and the known category information of the kth training included in the third data; When the loss value output by the loss function of the k-th training meets the preset loss value, the training is terminated to obtain the target comprehensive analysis model; When the loss value output by the loss function of the k-th training does not meet the preset loss value, adjusting the values ​​of the parameters in the initial comprehensive analysis model of the k-th training; The initial comprehensive analysis model trained for the kth time is used to perform the k+1th training until the loss value output by the loss function meets the preset loss value, and the training is terminated to obtain the target comprehensive analysis model.

4. The method according to claim 3, characterized in that Determining the object category output by the initial comprehensive analysis model of the k-th training based on the behavior association data, the risk assessment data, and the object grouping data includes: Determine a target association analysis result from the multiple association analysis results included in the behavior association data, determine a target risk assessment result from the multiple risk assessment results included in the risk assessment data, and determine a target grouping result from the multiple grouping results included in the object grouping data according to the voting mechanism in the initial comprehensive analysis model; The object category is generated based on the target association analysis result, the target risk assessment result and the target grouping result.

5. The method according to claim 4, characterized in that When the loss value output by the loss function of the k-th training does not meet the preset loss value, adjusting the values ​​of the parameters in the initial comprehensive analysis model of the k-th training includes: When the loss value output by the loss function of the k-th training does not meet the preset loss value, the voting mechanism is adjusted to obtain the initial comprehensive analysis model of the k-th training.

6. The method according to claim 1, characterized in that After the initial comprehensive analysis model is trained using the multiple sets of target training data to obtain the target comprehensive analysis model, the method further includes: Outputting target data to the target comprehensive analysis model to obtain a target object category output by the target comprehensive analysis model, wherein the target data includes target object information and target behavior information of the target object; generating a target service solution based on the target object category; Wherein, generating a target service solution based on the target object category includes: According to the target object category, the target risk indicator of the target object is obtained in real time through the risk warning model, wherein the target risk indicator includes at least one of the following: financial indicator, transaction behavior indicator, credit risk indicator, industry risk indicator; In the case where the target risk indicators include abnormal indicators, the target service plan is generated, wherein the target service plan is used to perform early warning operations and risk control operations on the target object, and the risk control operations include at least one of the following: adjusting the credit limit of the target object, and adjusting the intensity of credit monitoring on the target object.

7. A device for identifying an object category, characterized in that: include: an acquisition module, configured to acquire multiple sets of target training data, wherein each set of target training data includes: first data, second data, and third data, wherein the first data includes object information and transaction information of an object, the second data is behavior information of the object, and the third data is category information of the object; A training module is used to train an initial comprehensive analysis model using the multiple groups of target training data to obtain a target comprehensive analysis model, wherein the initial comprehensive analysis model includes a first component, a second component and a third component, the first component is a component including an association analysis algorithm, the second component is a component including a classification algorithm, and the third component is a component including a clustering algorithm. For each group of the training data, the first component is used to generate behavior association data of the object based on the first data and the second data, the second component is used to generate risk assessment data of the object based on the first data, and the third component is used to generate object grouping data of the object based on the first data. The initial comprehensive analysis model is used to output the object category of the object based on the behavior association data, the risk assessment data and the object grouping data.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program implements the steps of the method described in any one of claims 1 to 6 when executed by a processor.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method described in any one of claims 1 to 6 are implemented.

Images