Data processing method and device, computer equipment, medium and product

Abstract

The invention discloses a data processing method and device, computer equipment, a medium and a product, and the method comprises the steps: obtaining a training sample data set comprising transaction information, object information, behavior feature information and reference category labels of N sample objects, carrying out the derivative processing of the transaction information of each sample object in the N sample objects, and obtaining a training sample data set; derivation transaction information of each sample object is obtained; screening the transaction information, the derivative transaction information, the object information and the behavior characteristic information of each sample object to obtain target in-mold data of each sample object; and performing model training according to the target modeling data and the reference category labels of the N sample objects to obtain a target object evaluation model, the target object evaluation model being used for determining an evaluation result of the to-be-processed object according to the transaction information, the object information and the behavior feature information of the to-be-processed object. According to the invention, an object evaluation model with higher accuracy and higher universality can be trained, so that the accuracy of evaluation is improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a data processing method, a data processing apparatus, a computer device, a computer-readable storage medium, and a computer program product. Background Art

[0002] During the transaction process, how to detect the transaction object and promptly find out whether the transaction object is a non-standard user is currently a major research hotspot.

[0003] At present, most of them use training models to evaluate and manage transaction objects, which mainly include the following two types of model training methods. The first type is to train models based on basic information of transaction objects (such as gender, age, etc.) and characteristics such as transaction time and location (such as the proportion of night transactions, the proportion of transactions in different regions, etc.). The accuracy of models trained by this type of method is not high enough. The second type is to train models for a specific application scenario. This type of method is too targeted and is not applicable to other application scenarios.

[0004] Therefore, how to train an object evaluation model with high accuracy and strong versatility has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The embodiments of the present application provide a data processing method, apparatus, computer equipment, medium, and product, which can train an object evaluation model with higher accuracy and greater versatility, thereby improving the accuracy of the evaluation of transaction objects.

[0006] On the one hand, an embodiment of the present application discloses a data processing method, the method comprising:

[0007] Acquire a training sample data set, wherein the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, wherein the transaction information includes one or more of a transaction method, a transaction scenario, a transaction time, and statistical data, where N is an integer greater than or equal to 1;

[0008] Performing derivative processing on the transaction information of each sample object in the N sample objects to obtain derivative transaction information of each sample object;

[0009] Screening the transaction information, derivative transaction information, object information and behavior characteristic information of each sample object to obtain target modeling data of each sample object;

[0010] Model training is performed based on the target input data and reference category labels of the N sample objects to obtain a target object evaluation model, which is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral characteristic information of the object to be processed.

[0011] In one aspect, an embodiment of the present application discloses a data processing device, the device comprising:

[0012] The acquiring unit acquires a training sample data set, wherein the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, wherein the transaction information includes one or more of a transaction method, a transaction scenario, a transaction time, and statistical data, and N is an integer greater than or equal to 1;

[0013] a processing unit, configured to perform derivative processing on the transaction information of each sample object among the N sample objects to obtain the derivative transaction information of each sample object;

[0014] The processing unit is further used to screen the transaction information, derivative transaction information, object information and behavior feature information of each sample object to obtain target modeling data of each sample object;

[0015] The processing unit is also used to perform model training based on the target input data and reference category labels of the N sample objects to obtain a target object evaluation model, and the target object evaluation model is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral characteristic information of the object to be processed.

[0016] On the one hand, an embodiment of the present application discloses a computer device, which includes a processor suitable for implementing one or more computer programs; and a computer-readable storage medium, wherein the computer-readable storage medium stores one or more computer programs, and the one or more computer programs are suitable for being loaded by the processor and executed with the above-mentioned data processing method.

[0017] In one aspect, the present application discloses a computer-readable storage medium storing one or more computer programs, wherein the one or more computer programs are suitable for being loaded by a processor and executing the above-mentioned data processing method.

[0018] The present application discloses a computer program product, which includes a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the above-mentioned data processing method.

[0019] In the embodiment of the present application, when training the target object evaluation, a training sample data set including the transaction information, object information, behavior feature information and reference category labels of N sample objects is obtained, and the transaction information of each sample object in the N sample objects is derived to obtain the derived transaction information of each sample object; then the transaction information, derived transaction information, object information and behavior feature information of each sample object are screened to obtain the target model input data of each sample object; finally, the model is trained according to the target model input data and reference category labels of the N sample objects to obtain the target object evaluation model, which is used to determine the evaluation result of the object to be processed according to the transaction information, object information and behavior feature information of the object to be processed. Before the model training, the model training method performs feature derivation on the transaction information, and then uses the object information and behavior feature information as auxiliary information, so that the transaction data of each sample object is more comprehensive, and the transaction of the object can be analyzed more clearly, so that the versatility of the trained model is higher. At the same time, the importance of various information is determined by screening the transaction information, derived transaction information, object information and behavior feature information of each sample object, retaining important information and eliminating meaningless information, so that the accuracy of the trained model is higher. When the target object evaluation model is used to evaluate the object to be processed, the evaluation result obtained is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 It is a schematic diagram of a network architecture of a data processing system disclosed in an embodiment of the present application;

[0022] Figure 2 It is a flowchart of a data processing method disclosed in an embodiment of the present application;

[0023] Figure 3 is a schematic diagram of a derivation process disclosed in an embodiment of the present application;

[0024] Figure 4 It is a flowchart of another data processing method disclosed in an embodiment of the present application;

[0025] Figure 5 It is an overall framework diagram of an object evaluation process disclosed in an embodiment of the present application;

[0026] Figure 6 is a structural schematic diagram of a data processing device disclosed in an embodiment of the present application;

[0027] Figure 7 It is a structural schematic diagram of a computer device disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions 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.

[0029] In order to improve the accuracy of the evaluation of the transaction object during the transaction process, the embodiment of the present application provides a data processing method, including: ① Obtaining a training sample data set, the training sample data set includes transaction information, object information, behavioral feature information and reference category labels of N sample objects, and the transaction information includes one or more of transaction methods, transaction scenarios, transaction times and statistical data, and N is an integer greater than or equal to 1. ② Performing derivative processing on the transaction information of each sample object in the N sample objects to obtain the derived transaction information of each sample object, and performing feature derivation on the transaction information to make the transaction information more comprehensive, so that the transaction of the object can be analyzed more clearly, thereby making the trained model more versatile. ③ Screening the transaction information, derived transaction information, object information and behavioral feature information of each sample object to obtain the target model input data of each sample object, and by screening the transaction information, derived transaction information, object information and behavioral feature information of each sample object, the importance of various information is determined and important information is retained. ④ Perform model training based on the target input data and reference category labels of N sample objects to obtain a target object evaluation model, which is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral feature information of the object to be processed. Therefore, the target object evaluation model determined by this method has higher versatility and higher accuracy, and any transaction object can be evaluated more accurately through the target object evaluation model.

[0030] In one possible implementation, the data processing method provided in the embodiment of the present application can be implemented based on artificial intelligence technology (AI), which refers to the theory, method, technology and application system of using digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science, which attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that the machines have the functions of perception, reasoning and decision-making. AI is a comprehensive discipline that involves a wide range of fields, including both hardware-level technology and software-level technology. AI basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. AI software technology mainly includes several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning; and the expression migration method provided in the embodiment of the present application mainly involves machine learning in AI technology. Specifically, during model training, models such as xgboost (eXtreme Gradient Boosting) and random forest are used for training.

[0031] In a possible implementation, the data processing method provided in the embodiment of the present application can also be implemented based on cloud technology. Cloud technology refers to a hosting technology that unifies a series of resources such as hardware, software, and network in a wide area network or a local area network to realize data calculation, storage, processing, and sharing. Cloud technology is a general term for network technology, information technology, integration technology, management platform technology, application technology, etc. based on the cloud computing business model application, which can form a resource pool, which is used on demand and is flexible and convenient. Cloud computing technology will become an important support. Cloud storage is a new concept extended and developed from the concept of cloud computing. A distributed cloud storage system (hereinafter referred to as a storage system) refers to a storage system that uses cluster applications, grid technology, and distributed storage file systems to bring together a large number of different types of storage devices (storage devices are also called storage nodes) in the network through application software or application interfaces to work together and provide data storage and business access functions to the outside world. In short, a database can be regarded as an electronic file cabinet-a place to store electronic files. Users can add, query, update, delete, and other operations on the data in the file. The so-called "database" is a collection of data that is stored together in a certain way, can be shared with multiple users, has as little redundancy as possible, and is independent of the application program. In the embodiments of the present application, it may specifically involve one or more of cloud storage, cloud database, and big data in cloud technology. For example, the data required to execute the data processing method (such as a training sample data set) is obtained from the cloud database.

[0032] In a possible implementation, the data processing method provided in the embodiment of the present application can also be implemented based on the blockchain technology. The blockchain is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, encryption algorithm, etc. It is essentially a decentralized database, a string of data blocks generated by cryptographic methods, each of which contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. In the embodiment of the present application, the data required to execute the data processing method can be stored in the form of blocks on the blockchain; the data generated by executing the data processing method (such as the target object evaluation model) can also be stored in the form of blocks on the blockchain; in addition, the data processing device that executes the data processing method can be a node device in the blockchain network.

[0033] See also Figure 1, is a schematic diagram of the architecture of a data processing system disclosed in an embodiment of the present application, and the data processing system may include multiple terminal devices 101 and a server 102. Each terminal device 101 corresponds to a user, and the user corresponding to the terminal device 101 may be a transaction object. The transaction object initiates a transaction through the terminal device 101, and the terminal device 101 is responsible for recording the transaction data of each transaction object, including transaction information, object information, and behavior characteristic information. The server 102 is mainly used for model training and model prediction. Model training refers to training the model through the transaction information, object information, behavior characteristic information and reference category labels of a batch of sample objects to obtain a target object evaluation model; model prediction is for any transaction object, and the transaction information, object information and behavior characteristic information of the object to be processed are processed through the target object evaluation model to obtain the scoring result of the object.

[0034] In one embodiment, the terminal device 101 includes, but is not limited to: smart phones, tablet computers, smart wearable devices, smart voice interaction devices, smart home appliances, personal computers, vehicle-mounted terminals, smart cameras, virtual reality devices (such as AR (Augmented Reality) devices), etc., and this application does not limit this. The server 102 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or 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 (Content Delivery Network), and big data and artificial intelligence platforms, etc., and this embodiment of the application does not limit this. The terminal device 101 and the server 102 can establish a direct communication connection through wired communication, or can establish an indirect communication connection through wireless communication, and this embodiment of the application does not limit this.

[0035] In an embodiment of the present application, in combination with the data processing system, the data processing method may include: the server 102 obtains training sample data, the training sample data set includes transaction information, object information, behavior feature information and reference category labels of N sample objects; the transaction information of each sample object in the N sample objects is subjected to derivative processing to obtain the derivative transaction information of each sample object; the transaction information, derivative transaction information, object information and behavior feature information of each sample object are screened to obtain the target model input data of each sample object; the model is trained according to the target model input data and reference category labels of the N sample objects to obtain the target object evaluation model. After determining the target object evaluation model, the server 102 obtains the transaction information, object information and behavior feature information of the object to be processed from the terminal device 101, and then the server 102 scores the transaction information, object information and behavior feature information of the object to be processed through the target object evaluation model to obtain the evaluation result of the object to be processed.

[0036] In a feasible implementation, the data processing method provided in the embodiment of the present application can be applied as a general solution in various application scenarios, such as medical scenarios, vehicle network scenarios, etc., mainly used to explore the importance of variables and use important variables for model training.

[0037] It should be noted that in this application, the relevant data involved in the data processing process, such as the training sample data set in the training process, the transaction information, object information and behavioral characteristic information of the object to be processed in the prediction process. When the above embodiments of this application are applied to specific products or technologies, user permission or consent must be obtained, and the relevant data collection, use and processing process must comply with relevant laws, regulations and standards, comply with the principles of legality, legitimacy and necessity, and do not involve the acquisition of data types prohibited or restricted by laws and regulations. In some optional embodiments, the relevant data involved in the embodiments of this application are obtained after the object is authorized separately. In addition, when obtaining the separate authorization of the object, the purpose of the relevant data involved is indicated to the object.

[0038] The data processing method proposed in the embodiment of the present application is introduced in detail below.

[0039] See also Figure 2 , is a flow chart of a data processing method disclosed in an embodiment of the present application. The data processing method can be executed by a computer device, and the computer device can be Figure 1 The server 102 shown in the figure mainly describes the training process of the target object evaluation model, including but not limited to the following steps:

[0040] S201: Obtain a training sample data set, where the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, where the transaction information includes one or more of transaction methods, transaction scenarios, transaction times, and statistical data, where N is an integer greater than or equal to 1.

[0041] In order to train an object evaluation model that accurately evaluates user transaction data, a training sample data set is first obtained. The training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects. Transaction information includes one or more of transaction methods, transaction scenarios, transaction time, and statistical data.

[0042] Among them, transaction methods include change payment, bank card payment, etc.; transaction scenarios include red envelope scenarios, transfer scenarios, face-to-face payment scenarios, and commercial payment scenarios, etc.; transaction time specifically refers to a time period, which is obtained by segmenting a period of time of the sample object, such as early morning, 1 day, 7 days, and 30 days, etc.; statistical data is a dimension-based division of transaction data during the transaction process, such as transaction amount, number of transactions, and average transaction amount (average transaction amount for a period of time and average transaction amount for multiple transactions, etc.). Object information mainly refers to the static information of the transaction object, such as the user name, age, and other basic information of the object. Behavioral feature information is the transaction behavior of the transaction object over a period of time. According to the behavioral feature information, it can be determined whether the transaction behavior of the transaction object is normal. For example, the behavioral feature information is the transaction location information, the number of terminal devices used by the corresponding account used by the transaction object, and the behavior of the transaction object for the account (registering an account and canceling an account, etc.). The label of each sample object is summarized based on the sample object's easy information, object information, and behavioral feature information. This process can be determined manually through set rules.

[0043] S202: Performing derivative processing on the transaction information of each sample object among the N sample objects to obtain derivative transaction information of each sample object.

[0044] In order to more effectively analyze the transaction status of users, after obtaining the transaction information of N sample objects, the embodiment of the present application first performs derivative processing on the transaction information of each sample object. Compared with a single consideration method that only considers large-amount and night transactions, etc., it can more clearly grasp the transaction status of the object and can also more effectively deduce irregular transactions.

[0045] The transaction information includes one or more of the transaction method, transaction scenario, transaction time and statistical data. The specific transaction information has been introduced in step S201. In the derivation process, the transaction information of each sample object may be connected to one or more of the transaction method, transaction scenario, transaction time and statistical data. The data to be connected may be selected according to the actual situation. For the specific derivation process, please refer to Figure 3 , accurately identify the transaction of the object through transaction methods, transaction scenarios, transaction time and statistical data. Among them, the middle x represents the variable connection. For example, if you need to view the transaction status of the object's red envelopes, according to the transaction method, transaction time and statistical data, according to this variable derivation method, you can get the amount of red envelopes sent and received by the user in the early morning, the average amount of red envelopes, how many red envelopes were received, whether they were sent and received with change or bank cards; you can also view the amount and number of red envelopes sent and received by the user on 1 / 7 / 30 days.

[0046] It should be noted that the object information and the behavior feature information do not need to be derived, and both can be used as auxiliary information for each sample object to strengthen the input data of each sample object.

[0047] S203: Filter the transaction information, derivative transaction information, object information and behavior feature information of each sample object to obtain target modeling data for each sample object.

[0048] After obtaining the derivative transaction information, object information and behavioral characteristic information of each sample object, not all of this information is effective. Some information will have side effects on the final model. Therefore, before that, it is necessary to evaluate the effectiveness of these variables, which is a screening process.

[0049] In one possible implementation, the weight of evidence and information value are used in the screening process, among which WEO is mainly used in the following business applications:

[0050] 1. Handling missing values: When the data source does not have 100% coverage, there will be missing values. In this case, null can be used as a separate bin. This is very useful when modeling data sources, and can effectively utilize data sources with a coverage rate of only 20%.

[0051] 2. Handling outliers: When there are outliers in the data, they can be discretized by binning to improve the robustness (anti-interference ability) of the variable. For example, if age has an outlier of 200, it can be classified into the bin "age>60" to eliminate the impact.

[0052] 3. Business Explanation: We are used to linearly judging the effect of variables. When x increases, y increases. However, there is often a nonlinear relationship between x and y. In this case, WOE transformation can be used.

[0053] IV is an indicator closely related to WOE, and is often used to evaluate the predictive ability of variables. Therefore, it can be used to quickly screen variables. Variables with relatively small IV and Woe values ​​can be eliminated according to the actual business situation; at the same time, some more important variables can be obtained through IV and Woe, which can better explain "historical samples and future samples follow the same overall distribution", so that mathematical and statistical laws can be summarized from historical samples to predict the performance of future samples.

[0054] In the specific implementation process, the process of processing the transaction information, derivative transaction information, object information and behavioral characteristic information of each sample object using the evidence weight method includes: ① Grouping N sample objects according to each of the transaction information, derivative transaction information, object information and behavioral characteristic information to obtain multiple groups corresponding to each information; obtaining a first number of sample objects with a reference category label of the first type in each of the multiple groups, and obtaining a second number of sample objects with a reference category label of the second type in each group, where the reference type label is used to indicate the quality of each sample object; obtaining the total number of sample objects with a reference category label of the first type in the N sample objects, and obtaining the total number of sample objects with a reference category label of the second type in the N sample objects; according to; according to the first number, the second number, the total number of sample objects with a reference category label of the first type in the N sample objects, and the total number of sample objects with a reference category label of the second type in the N sample objects, determine the first weight value of each group, and the first weight value of each group is the woe of each group. Specifically, this process is to first determine a first ratio based on the first number and the total number of sample objects with reference category labels of the first type in N sample objects, determine a second ratio based on the second number and the total number of sample objects with reference category labels of the second type in N sample objects, and then determine a first weight value for each group based on the first ratio and the second ratio.

[0055] Assume that there are 100 sample objects, of which 40 have the reference category label of the first type, and 60 have the reference category label of the second type. We analyze the information of "age" as an example. First, the ages of the 100 sample objects are grouped. Assume that 5 groups are obtained, and the woe of each group needs to be calculated. Taking the first group as an example, there are 20 sample objects in the first group, of which 14 have the reference category label of the first type, and 6 have the reference category label of the second type. Correspondingly, the woe of the first group = ln((14 / 40) / (6 / 60)) = 1.2528.

[0056] Furthermore, for each type of information, after determining the first weight value of each group, the information value corresponding to each type of information is determined according to the first weight value of each group. In the specific implementation process, the information value of each group may be calculated first, and then the information values ​​of each group may be added together to obtain the information value corresponding to each type of information. Among them, for the information value of any group, the first ratio between the first number and the total number of sample objects with a reference category label of the first type in the N sample objects may be determined first, and the second ratio between the second number and the total number of sample objects with a reference category label of the second type in the N sample objects may be determined; the second ratio is subtracted from the first ratio to obtain the second weight value of each group, and the second weight value of each group and the first weight value of each group are used to determine the information value of each group. Specifically, the second weight value of each group and the first weight value of each group may be multiplied to obtain the information value of each group.

[0057] In a possible implementation, after determining the information value of each type of information, the information value corresponding to each type of information is compared with the information value threshold, which is a summarized value. As shown in Table 1, it is a prediction table of information value. The greater the information value, the stronger the prediction ability corresponding to the information. Table 1 is exemplary, and there may be other standards, which are not limited here.

[0058] Table 1

[0059] Information value range Prediction effect Less than 0.02 Almost no 0.02～0.1 weak 0.1～0.3 medium 0.3～0.5 powerful Greater than 0.5 Unbelievable, need further confirmation

[0060] In a possible implementation, the information value threshold can be set to 0.0.2, that is, the target information whose corresponding information value is greater than the information value threshold is determined from the transaction information, derived transaction information, object information, and behavior feature information. Then, the target information of each sample object is spliced ​​to obtain the target input data of each sample object.

[0061] Among them, it should be noted that for some data with too much information value, for example, information with an information value greater than 0.5, you can first make a manual judgment on this information to determine whether there is an error in the information. If there is no error, you can retain this information.

[0062] Although some important variables can be obtained through IV and WEO, these variables can better explain "historical samples and future samples follow the same overall distribution", and can also summarize mathematical and statistical laws from historical samples to predict the performance of future samples. However, according to business experience, these variables are highly explanatory, simple, and have fewer types. If these important features are directly entered into the machine learning model together with other variables, the effect is not ideal and it is impossible to effectively fit future samples. Therefore, it is necessary to enhance these samples. The enhancement method adopted in this solution is mainly to bin the important features again, and then add the bins to the original variables to increase the weight. Among them, there is also a part of the information that does not need to be binned. The information that is not binned refers to some variables that have few values ​​or strong significance, so such variables can be directly entered into the model.

[0063] Specifically, if the information value of the target variable is greater than a preset value, and the number of possible values ​​of the target variable reaches a preset threshold, the target variable is binned, and then the binned information and the non-binned information are concatenated to obtain the target input data. The preset value can be the same as the information value threshold, or a value greater than the information value threshold.

[0064] For example, the process of binning the user's transaction amount is as follows: it is a continuous variable. The user's transaction amount may be 100, 500, 2000, or 100000. If this amount data is directly input into the model, the model will not learn the specific meaning of this variable. In order to better let the model learn the meaning of this variable, we can divide the transaction amount into four intervals: [0.1000], [1000, 5000], [5000, 20000], and [20000, +], and re-mark them, such as using [1, 2, 3, 4] to map the values ​​of specific intervals.

[0065] S204: Perform model training based on the target input data and reference category labels of N sample objects to obtain a target object evaluation model, which is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral feature information of the object to be processed.

[0066] In a possible implementation, the initial object evaluation model is called to process the target input data of each sample object in N sample objects to obtain the predicted category label of each sample object. The initial object evaluation model can be a binary classification model. When processing the target input data of each sample object, the output data is 0 or 1, representing different categories respectively; the initial object evaluation model can be a binary classification model, and the initial object evaluation model can be a scoring model. When processing the target input data of each sample object, the output data is a score. When it is greater than a certain threshold, it corresponds to one category, and when it is less than or equal to a certain threshold, it corresponds to another category. Then, determine the third number of sample objects whose corresponding predicted category labels and reference category labels are consistent among the N sample objects, and determine the fourth number of sample objects whose predicted category labels are of the first type among the N sample objects; determine the ratio between the third number and the N value as the accuracy of the initial object evaluation model, and determine the ratio between the fourth number and the number of sample objects whose reference category labels are of the first type among the N sample objects as the recall of the initial object evaluation model. If the accuracy is less than the first threshold, or the recall is less than the second threshold, the model parameters of the initial object evaluation model are adjusted according to the predicted category labels of the N sample objects and the reference category labels; if the accuracy is greater than or equal to the first threshold, and the recall is greater than or equal to the second threshold, the initial object evaluation model is used as the target object evaluation model. Among them, the first threshold and the second threshold are both values ​​less than 1. If the first threshold and the second threshold are infinitely close to 1, the model is generally overfitting and cannot effectively identify unknown problems. This conflicts with the ability of this application to discover unknown problems. Therefore, in combination with the actual business situation, the appropriate first threshold and second threshold are selected, thereby training a model that can both better fit historical data and predict unknown problems in a timely manner. For example, the first threshold can be set to 0.85, and the second threshold can be set to 0.93.

[0067] Among them, the initial object evaluation model can adopt the xgboost model, or it can adopt random forest, lightGBM (Light Gradient Boosting Machine, which is a fast, distributed, high-performance gradient boosting framework based on the decision tree algorithm), etc.

[0068] In the embodiment of the present application, the training process of the target object evaluation model is mainly described. A training sample data set including transaction information, object information, behavioral feature information and reference category labels of N sample objects is obtained, and the transaction information of each sample object in the N sample objects is derived to obtain the derived transaction information of each sample object; then the transaction information, derived transaction information, object information and behavioral feature information of each sample object are screened to obtain the target input data of each sample object; finally, the model is trained according to the target input data and reference category labels of the N sample objects to obtain the target object evaluation model. In the process of model training, the model training method, on the one hand, performs feature derivation on the transaction information, and then uses the object information and behavioral feature information as auxiliary information, so that the transaction data of each sample object is more comprehensive, and the transaction of the object can be analyzed more clearly, so that the versatility of the trained model is higher. On the other hand, the importance of various information is determined by screening the transaction information, derived transaction information, object information and behavioral feature information of each sample object, retaining important information and eliminating meaningless information, so that the accuracy of the trained model is higher.

[0069] See also Figure 4 , is a flow chart of another data processing method disclosed in an embodiment of the present application, the data processing method combines model training and model prediction. Similarly, the data processing method can also be executed by a computer device, which can be Figure 1 The server 102 shown, correspondingly, the data processing method includes but is not limited to the following steps:

[0070] S401: Obtain a training sample data set, where the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, where the transaction information includes one or more of transaction methods, transaction scenarios, transaction times, and statistical data, where N is an integer greater than or equal to 1.

[0071] S402: Performing derivative processing on the transaction information of each sample object among the N sample objects to obtain derivative transaction information of each sample object.

[0072] S403: Filter the transaction information, derived transaction information, object information and behavior feature information of each sample object to obtain target modeling data for each sample object.

[0073] S404: Perform model training based on the target input data and reference category labels of the N sample objects to obtain a target object evaluation model.

[0074] The implementation process of the above S401-S404 can be found in Figure 2The relevant description of the implementation process of S201-S204 in the illustrated embodiment is not repeated here. The determined target object evaluation model can be stored in a local server, a cloud server, or a blockchain.

[0075] S405: Acquire transaction information, object information, and behavior characteristic information of the object to be processed, where the transaction information includes one or more of transaction method, transaction scenario, transaction time, and statistical data.

[0076] In a possible implementation, the evaluation of the transaction object can be triggered by a timer, that is, a batch of transaction objects are automatically evaluated at intervals; it can also be triggered by a merchant, that is, the merchant sends a request to the server to evaluate a certain user. Before calling the target object evaluation model to evaluate the object to be processed, the transaction information, object information, and behavioral feature information of the object to be processed are obtained first. The transaction information includes one or more of the transaction method, transaction scenario, transaction time, and statistical data. The object to be processed can be one or more.

[0077] S406: Perform derivative processing on the transaction information of the object to be processed to obtain derivative transaction information of the object to be processed.

[0078] The derivative processing process here is the same as that in step S202, and the transaction information of the object to be processed includes one or more of the transaction method, transaction scenario, transaction time and statistical data, and the derivative transaction information of the object to be processed is obtained.

[0079] S407: The derivative transaction information, object information and behavior feature information of the object to be processed are integrated and processed to obtain the target input model data of the object to be processed.

[0080] In a possible implementation, the fusion processing may be to splice the derivative transaction information, object information, and behavior feature information of the object to be processed.

[0081] S408: Calling the target object evaluation model to process the target input data of the object to be processed to obtain an evaluation result of the object to be processed.

[0082] After determining the target input data of the object to be processed, the target object evaluation model is directly called to process the target input data of the object to be processed to obtain the evaluation result of the object to be processed. The evaluation result of the object to be processed can be a classification result or a score. If it is a classification result, the quality of the object to be processed can be directly determined; if it is a score, the quality of the object to be processed is determined according to the score and the score threshold.

[0083] In a possible implementation, if it is determined based on the evaluation results that the subject to be processed is a user with irregular transactions, the merchant may be informed, or a prompt message may be directly output to the user to inform the transaction subject of the potential problem.

[0084] The above steps S405-S408 describe how to evaluate the object to be processed. The overall framework can be as follows: Figure 5 As shown, firstly, the transaction data of the object to be processed (including transaction information, object information, and behavioral characteristic information) is obtained; then, these transaction data are derived and integrated to obtain the target input data of the object to be processed; the target object evaluation model is called to process the target input data; and the scoring result of the object to be processed is output.

[0085] The embodiment of the present application describes a data processing method in combination with model prediction. After training a target object evaluation model, the target object evaluation model can be directly called to process any object to be processed, thereby determining the evaluation result of each object to be processed. By using the target object evaluation model trained by the present application to evaluate the object to be processed, a more accurate evaluation result can be obtained.

[0086] The method of the embodiment of the present application is described in detail above. In order to facilitate better implementation of the above scheme of the embodiment of the present application, the device of the embodiment of the present application is provided below accordingly.

[0087] See also Figure 6 , Figure 6 It is a structural diagram of a data processing device provided in an embodiment of the present application. The data processing device can be set in the computer device provided in an embodiment of the present application. Figure 6 The data processing device shown may be a computer program (including program code) running in a computer device, and the flow determination device may be used to execute Figure 2 , Figure 4 as well as Figure 5 Some or all of the steps in the method embodiment shown. Figure 6 , the data processing device may include the following units:

[0088] An acquisition unit 601 is used to acquire a training sample data set, wherein the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, wherein the transaction information includes one or more of a transaction method, a transaction scenario, a transaction time, and statistical data, and N is an integer greater than or equal to 1;

[0089] The processing unit 602 is used to perform derivative processing on the transaction information of each sample object among the N sample objects to obtain the derived transaction information of each sample object; screen the transaction information, derived transaction information, object information and behavioral characteristic information of each sample object to obtain the target input data of each sample object; perform model training based on the target input data of the N sample objects and reference category labels to obtain a target object evaluation model, and the target object evaluation model is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral characteristic information of the object to be processed.

[0090] In a possible implementation, the processing unit 602 screens the transaction information, the derived transaction information, the object information, and the behavior characteristic information of each sample object to obtain the target input data of each sample object, and is specifically configured to perform the following steps:

[0091] According to each of the transaction information, the derivative transaction information, the object information, and the behavior characteristic information, the N sample objects are grouped to obtain a plurality of groups corresponding to each of the information;

[0092] Determine a first weight value of each group in a plurality of groups corresponding to each type of information;

[0093] Determine the information value corresponding to each type of information according to the first weight value of each group corresponding to each type of information;

[0094] The target input data of each sample object is determined according to the information value and information value threshold corresponding to each information.

[0095] In a possible implementation, when the determining unit 603 determines the first weight value of each group in the multiple groups corresponding to each type of information, it is specifically configured to perform the following steps:

[0096] Obtaining a first number of sample objects whose reference category label is a first type in each of the plurality of groups, and obtaining a second number of sample objects whose reference category label is a second type in each of the plurality of groups;

[0097] The first weight value of each group is determined according to the first number, the second number, the total number of sample objects whose reference category labels are the first type among the N sample objects, and the total number of sample objects whose reference category labels are the second type among the N sample objects.

[0098] In a possible implementation, when the determining unit 603 determines the information value corresponding to each type of information according to the first weight value of each group corresponding to each type of information, it is specifically configured to perform the following steps:

[0099] Determine a first ratio between the first number and a total number of sample objects whose reference category labels are the first type in the N sample objects;

[0100] determining a second ratio between the second number and a total number of sample objects whose reference category labels are the second type in the N sample objects;

[0101] Determine the information value of each group according to the first ratio, the second ratio, and the first weight value of each group;

[0102] The information value corresponding to each type of information is determined according to the information values ​​of the multiple groups.

[0103] In a possible implementation, when the determination unit 603 determines the target input data of each sample object according to the information value and the information value threshold corresponding to each type of information, it is specifically configured to perform the following steps:

[0104] Comparing the information value corresponding to each type of information with the information value threshold;

[0105] Determine, from the transaction information, the derivative transaction information, the object information, and the behavior characteristic information, target information whose corresponding information value is greater than the information value threshold;

[0106] The target information of each sample object is spliced ​​to obtain the target modeling data of each sample object.

[0107] In a possible implementation, the processing unit 602 performs derivative processing on the transaction information of each sample object in the N sample objects to obtain the derivative transaction information of each sample object, specifically for executing the following steps:

[0108] The transaction information of each sample object includes one or more of the transaction method, transaction scenario, transaction time and statistical data, and connection processing is performed to obtain the derivative transaction information of each sample object.

[0109] In a possible implementation, the processing unit 602 performs model training according to the target input data and reference category labels of the N sample objects to obtain the target object evaluation model, which is specifically used to perform the following steps:

[0110] Calling the initial object evaluation model to process the target input data of each sample object in the N sample objects to obtain a predicted category label for each sample object;

[0111] Determine a third number of sample objects whose corresponding predicted category labels are consistent with the reference category labels among the N sample objects, and determine a fourth number of sample objects whose predicted category labels are of the first type among the N sample objects;

[0112] Determining the accuracy and recall of the initial object evaluation model according to the third number, the fourth number, and the total number N of sample objects;

[0113] If the accuracy is less than a first threshold, or the recall is less than a second threshold, adjusting the model parameters of the initial object evaluation model according to the predicted category labels and the reference category labels of the N sample objects;

[0114] If the accuracy is greater than or equal to the first threshold, and the recall is greater than or equal to the second threshold, the initial object evaluation model is used as the target object evaluation model.

[0115] In one implementation, the acquisition unit 601 is further used to acquire transaction information, object information, and behavior characteristic information of the object to be processed, wherein the transaction information includes one or more of a transaction method, a transaction scenario, a transaction time, and statistical data;

[0116] The processing unit 602 is also used to perform variable derivation processing on the transaction information of the object to be processed to obtain the derived transaction information of the object to be processed; fuse the derived transaction information, object information and behavioral characteristic information of the object to be processed to obtain the target input data of the object to be processed; call the target object evaluation model to process the target input data of the object to be processed to obtain the evaluation result of the object to be processed.

[0117] According to another embodiment of the present application, Figure 6 The various units in the data processing device shown can be respectively or all combined into one or several other units to constitute, or some of the units (some) can also be split into multiple smaller units in function to constitute, which can achieve the same operation without affecting the realization of the technical effects of the embodiments of the present application. The above-mentioned units are divided based on logical functions. In practical applications, the function of a unit can also be implemented by multiple units, or the function of multiple units can be implemented by one unit. In other embodiments of the present application, the data processing device may also include other units. In practical applications, these functions can also be implemented with the assistance of other units, and can be implemented by the collaboration of multiple units.

[0118] According to another embodiment of the present application, the program can be executed by running on a general computing device such as a computer including a central processing unit (CPU), a random access memory medium (RAM), a read-only memory medium (ROM), and other processing elements and storage elements. Figure 2 , Figure 4 or Figure 5 A computer program (including program code) for each step involved in part or all of the method shown in the figure is used to construct a computer program (including program code) for each step involved in part or all of the method shown in the figure. Figure 6 The data processing device shown in the embodiment of the present application is implemented by the data processing device. The computer program can be recorded on a computer-readable storage medium, for example, and loaded into the above-mentioned computing device through the computer-readable storage medium and run therein.

[0119] In the embodiment of the present application, a training sample data set including transaction information, object information, behavior feature information and reference category labels of N sample objects is obtained, and the transaction information of each sample object in the N sample objects is derived to obtain the derived transaction information of each sample object; then the transaction information, derived transaction information, object information and behavior feature information of each sample object are screened to obtain the target model input data of each sample object; finally, model training is performed based on the target model input data and reference category labels of the N sample objects to obtain the target object evaluation model, which is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavior feature information of the object to be processed. Before model training, the model training method performs feature derivation on the transaction information, and then uses the object information and behavior feature information as auxiliary information, so that the transaction data of each sample object is more comprehensive, and the transaction of the object can be analyzed more clearly, so that the versatility of the trained model is higher. At the same time, the importance of various information is determined by screening the transaction information, derived transaction information, object information and behavior feature information of each sample object, retaining important information and eliminating meaningless information, so that the accuracy of the trained model is higher.

[0120] Based on the above method and device embodiments, the present application embodiment provides a computer device. Figure 7 , Figure 7 is a schematic diagram of a computer device provided in an embodiment of the present application. The computer device may be Figure 1 Server 102 is shown. Figure 7 The computer device shown includes at least a processor 701, an input interface 702, an output interface 703, and a computer-readable storage medium 704. The processor 701, the input interface 702, the output interface 703, and the computer-readable storage medium 704 may be connected via a bus or other means.

[0121] The computer-readable storage medium 704 may be stored in a memory of a computer device. The computer-readable storage medium 704 is used to store a computer program. The computer program includes computer instructions. The processor 701 is used to execute the program instructions stored in the computer-readable storage medium 704. The processor 701 (or CPU (Central Processing Unit)) is the computing core and control core of the computer device, which is suitable for implementing one or more computer instructions, and is specifically suitable for loading and executing one or more computer instructions to implement a corresponding method flow or a corresponding function.

[0122] The embodiment of the present application also provides a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space that stores the operating system of the computer device. In addition, one or more computer instructions suitable for being loaded and executed by the processor are also stored in the storage space. These computer instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory, or a non-volatile memory (Non-VolatileMemory), such as at least one disk storage; optionally, it can also be at least one computer-readable storage medium located away from the aforementioned processor.

[0123] In some embodiments, the processor 701 may load and execute one or more computer instructions stored in the computer-readable storage medium 804 to implement the above-mentioned Figure 2 , Figure 4 as well as Figure 5 In a specific implementation, the computer instructions in the computer-readable storage medium 704 are loaded by the processor 701 and execute the following steps:

[0124] Acquire a training sample data set, wherein the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, wherein the transaction information includes one or more of a transaction method, a transaction scenario, a transaction time, and statistical data, where N is an integer greater than or equal to 1;

[0125] Performing derivative processing on the transaction information of each sample object in the N sample objects to obtain derivative transaction information of each sample object;

[0126] Screening the transaction information, derivative transaction information, object information and behavior characteristic information of each sample object to obtain target modeling data of each sample object;

[0127] Model training is performed based on the target input data and reference category labels of the N sample objects to obtain a target object evaluation model, which is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral characteristic information of the object to be processed.

[0128] In a possible implementation, the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to screen the transaction information, derivative transaction information, object information, and behavior feature information of each sample object, and when the target input data of each sample object is obtained, the computer instructions are specifically used to perform the following steps:

[0129] According to each of the transaction information, the derivative transaction information, the object information, and the behavior characteristic information, the N sample objects are grouped to obtain a plurality of groups corresponding to each of the information;

[0130] Determine a first weight value of each group in a plurality of groups corresponding to each type of information;

[0131] Determine the information value corresponding to each type of information according to the first weight value of each group corresponding to each type of information;

[0132] The target input data of each sample object is determined according to the information value and information value threshold corresponding to each information.

[0133] In a possible implementation, when the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to determine the first weight value of each group in the multiple groups corresponding to each type of information, the computer instructions are specifically used to perform the following steps:

[0134] Obtaining a first number of sample objects whose reference category label is a first type in each of the plurality of groups, and obtaining a second number of sample objects whose reference category label is a second type in each of the plurality of groups;

[0135] The first weight value of each group is determined according to the first number, the second number, the total number of sample objects whose reference category labels are the first type among the N sample objects, and the total number of sample objects whose reference category labels are the second type among the N sample objects.

[0136] In a possible implementation, when the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to determine the information value corresponding to each type of information according to the first weight value of each group corresponding to each type of information, the computer instructions are specifically used to perform the following steps:

[0137] Determine a first ratio between the first number and a total number of sample objects whose reference category labels are the first type in the N sample objects;

[0138] determining a second ratio between the second number and a total number of sample objects whose reference category labels are the second type in the N sample objects;

[0139] Determine the information value of each group according to the first ratio, the second ratio, and the first weight value of each group;

[0140] The information value corresponding to each type of information is determined according to the information values ​​of the multiple groups.

[0141] In a possible implementation, when the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to determine the target model input data of each sample object according to the information value and the information value threshold corresponding to each type of information, the computer instructions are specifically used to perform the following steps:

[0142] Comparing the information value corresponding to each type of information with the information value threshold;

[0143] Determine, from the transaction information, the derivative transaction information, the object information, and the behavior characteristic information, target information whose corresponding information value is greater than the information value threshold;

[0144] The target information of each sample object is spliced ​​to obtain the target modeling data of each sample object.

[0145] In a possible implementation, the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to perform derivative processing on the transaction information of each sample object in the N sample objects, and when the derivative transaction information of each sample object is obtained, the computer instructions are specifically used to perform the following steps:

[0146] The transaction information of each sample object includes one or more of the transaction method, transaction scenario, transaction time and statistical data, and connection processing is performed to obtain the derivative transaction information of each sample object.

[0147] In a possible implementation, the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to perform model training according to the target input data and reference category labels of the N sample objects to obtain the target object evaluation model, specifically for performing the following steps:

[0148] Calling the initial object evaluation model to process the target input data of each sample object in the N sample objects to obtain a predicted category label for each sample object;

[0149] Determine a third number of sample objects whose corresponding predicted category labels are consistent with the reference category labels among the N sample objects, and determine a fourth number of sample objects whose predicted category labels are of the first type among the N sample objects;

[0150] Determining the accuracy and recall of the initial object evaluation model according to the third number, the fourth number, and the total number N of sample objects;

[0151] If the accuracy is less than a first threshold, or the recall is less than a second threshold, adjusting the model parameters of the initial object evaluation model according to the predicted category labels and the reference category labels of the N sample objects;

[0152] If the accuracy is greater than or equal to the first threshold, and the recall is greater than or equal to the second threshold, the initial object evaluation model is used as the target object evaluation model.

[0153] In one possible implementation, the computer instructions in the computer-readable storage medium 704 are loaded by the processor 701 and are further used to perform the following steps:

[0154] Acquire transaction information, object information, and behavior characteristic information of the object to be processed, wherein the transaction information includes one or more of transaction method, transaction scenario, transaction time, and statistical data;

[0155] Performing variable derivation processing on the transaction information of the object to be processed to obtain derived transaction information of the object to be processed;

[0156] The derivative transaction information, object information and behavior characteristic information of the object to be processed are integrated to obtain the target model input data of the object to be processed;

[0157] The target object evaluation model is called to process the target input data of the object to be processed to obtain an evaluation result of the object to be processed.

[0158] In the embodiment of the present application, a training sample data set including transaction information, object information, behavior feature information and reference category labels of N sample objects is obtained, and the transaction information of each sample object in the N sample objects is derived to obtain the derived transaction information of each sample object; then the transaction information, derived transaction information, object information and behavior feature information of each sample object are screened to obtain the target model input data of each sample object; finally, model training is performed based on the target model input data and reference category labels of the N sample objects to obtain the target object evaluation model, which is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavior feature information of the object to be processed. Before model training, the model training method performs feature derivation on the transaction information, and then uses the object information and behavior feature information as auxiliary information, so that the transaction data of each sample object is more comprehensive, and the transaction of the object can be analyzed more clearly, so that the versatility of the trained model is higher. At the same time, the importance of various information is determined by screening the transaction information, derived transaction information, object information and behavior feature information of each sample object, retaining important information and eliminating meaningless information, so that the accuracy of the trained model is higher.

[0159] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions 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 executes the data processing method provided in the above-mentioned various optional modes.

[0160] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, some steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0161] In the several embodiments provided in this application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, and the division of the modules described above is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0162] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A data processing method, characterized in that: The method comprises: Acquire a training sample data set, wherein the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, wherein the transaction information includes one or more of a transaction method, a transaction scenario, a transaction time, and statistical data, where N is an integer greater than or equal to 1; Performing derivative processing on the transaction information of each sample object in the N sample objects to obtain derivative transaction information of each sample object; Screening the transaction information, derivative transaction information, object information and behavior characteristic information of each sample object to obtain target modeling data of each sample object; Model training is performed based on the target input data and reference category labels of the N sample objects to obtain a target object evaluation model, which is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral characteristic information of the object to be processed.

2. The method according to claim 1, characterized in that The screening of the transaction information, the derivative transaction information, the object information and the behavior characteristic information of each sample object to obtain the target modeling data of each sample object includes: According to each of the transaction information, the derivative transaction information, the object information, and the behavior characteristic information, the N sample objects are grouped to obtain a plurality of groups corresponding to each of the information; Determine a first weight value of each group in a plurality of groups corresponding to each type of information; Determine the information value corresponding to each type of information according to the first weight value of each group corresponding to each type of information; The target input data of each sample object is determined according to the information value and information value threshold corresponding to each information.

3. The method according to claim 2, characterized in that The determining a first weight value of each of the multiple groups corresponding to each type of information includes: Obtaining a first number of sample objects whose reference category label is a first type in each of the plurality of groups, and obtaining a second number of sample objects whose reference category label is a second type in each of the plurality of groups; The first weight value of each group is determined according to the first number, the second number, the total number of sample objects whose reference category labels are the first type among the N sample objects, and the total number of sample objects whose reference category labels are the second type among the N sample objects.

4. The method according to claim 3, characterized in that The determining the information value corresponding to each type of information according to the first weight value of each group corresponding to each type of information includes: Determine a first ratio between the first number and a total number of sample objects whose reference category labels are the first type in the N sample objects; determining a second ratio between the second number and a total number of sample objects whose reference category labels are the second type in the N sample objects; Determine the information value of each group according to the first ratio, the second ratio, and the first weight value of each group; The information value corresponding to each type of information is determined according to the information values ​​of the multiple groups.

5. The method according to claim 2, characterized in that: The step of determining the target model input data of each sample object according to the information value and the information value threshold corresponding to each type of information includes: Comparing the information value corresponding to each type of information with the information value threshold; Determine, from the transaction information, the derivative transaction information, the object information, and the behavior characteristic information, target information whose corresponding information value is greater than the information value threshold; The target information of each sample object is spliced ​​to obtain the target modeling data of each sample object.

6. The method according to claim 1, characterized in that The performing derivation processing on the transaction information of each sample object in the N sample objects to obtain the derived transaction information of each sample object includes: The transaction information of each sample object includes one or more of the transaction method, transaction scenario, transaction time and statistical data, and connection processing is performed to obtain the derivative transaction information of each sample object.

7. The method according to any one of claims 1 to 6, characterized in that: The method of performing model training according to the target input data and reference category labels of the N sample objects to obtain a target object evaluation model includes: Calling the initial object evaluation model to process the target input data of each sample object in the N sample objects to obtain a predicted category label for each sample object; Determine a third number of sample objects whose corresponding predicted category labels are consistent with the reference category labels among the N sample objects, and determine a fourth number of sample objects whose predicted category labels are of the first type among the N sample objects; Determining the accuracy and recall of the initial object evaluation model according to the third number, the fourth number, and the total number N of sample objects; If the accuracy is less than a first threshold, or the recall is less than a second threshold, adjusting the model parameters of the initial object evaluation model according to the predicted category labels and the reference category labels of the N sample objects; If the accuracy is greater than or equal to the first threshold, and the recall is greater than or equal to the second threshold, the initial object evaluation model is used as the target object evaluation model.

8. The method according to claim 1, characterized in that The method further comprises: Acquire transaction information, object information, and behavior characteristic information of the object to be processed, wherein the transaction information includes one or more of transaction method, transaction scenario, transaction time, and statistical data; Performing variable derivation processing on the transaction information of the object to be processed to obtain derived transaction information of the object to be processed; The derivative transaction information, object information and behavior characteristic information of the object to be processed are integrated to obtain the target model input data of the object to be processed; The target object evaluation model is called to process the target input data of the object to be processed to obtain an evaluation result of the object to be processed.

9. A data processing device, characterized in that: The device comprises: The acquiring unit acquires a training sample data set, wherein the training sample data set includes transaction information, object information, behavior feature information, and reference category labels of N sample objects, wherein the transaction information includes one or more of a transaction method, a transaction scenario, a transaction time, and statistical data, and N is an integer greater than or equal to 1; a processing unit, configured to perform derivative processing on the transaction information of each sample object among the N sample objects to obtain the derivative transaction information of each sample object; The processing unit is further used to screen the transaction information, derivative transaction information, object information and behavior feature information of each sample object to obtain target modeling data of each sample object; The processing unit is also used to perform model training based on the target input data and reference category labels of the N sample objects to obtain a target object evaluation model, and the target object evaluation model is used to determine the evaluation result of the object to be processed based on the transaction information, object information and behavioral characteristic information of the object to be processed.

10. A computer device, characterized in that: The computer device comprises: a processor adapted to implement one or more computer programs; and, A computer-readable storage medium, wherein the computer-readable storage medium stores one or more computer programs, wherein the one or more computer programs are suitable for being loaded by the processor and executing the data processing method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more computer programs, and the one or more computer programs are suitable for being loaded by a processor and executing the data processing method according to any one of claims 1 to 8.

12. A computer program product, characterized in that The computer program product includes a computer program, which is stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the data processing method according to any one of claims 1 to 8.

Images