CN111160745A

CN111160745A - User account data processing method and device

Info

Publication number: CN111160745A
Application number: CN201911334802.5A
Authority: CN
Inventors: 刘永波; 唐啸; 肖雷; 曾凡麟
Original assignee: China Construction Bank Corp; CCB Finetech Co Ltd
Current assignee: China Construction Bank Corp
Priority date: 2019-12-23
Filing date: 2019-12-23
Publication date: 2020-05-15

Abstract

The invention discloses a method and a device for processing user account data, wherein the method comprises the following steps: acquiring user account data, wherein the user account data comprises: user industrial and commercial information, account basic information, account fund information and judicial information; inputting user account data into a pre-trained risk identification model to output account risk information; and executing early warning operation according to the account risk information. By the method and the device, whether the account has risks or not can be quickly predicted, so that the authorized authorities can be efficiently assisted to handle cases.

Description

User account data processing method and device

Technical Field

The invention relates to the field of data processing, in particular to a method and a device for processing user account data.

Background

With the increase of the work intensity of the state for fighting against the illegal crimes of the telecommunication network, the personal account opening and the account transaction are strictly controlled, and the following technical problems mainly exist: a large amount of details of fund transaction of involved accounts are exported through an EXCEL (electronic form), and the relation of the accounts is statistically analyzed by writing macros, so that time and labor are consumed.

At present, a bank system establishes an account-fund transaction-enterprise relation through a knowledge graph technology, a bank assists an authorized authority to handle cases, account fund transaction and account information are mainly analyzed through a statistical analysis tool, and the method is long in time consumption and high in error rate.

Disclosure of Invention

In view of the above, the present invention provides a method and an apparatus for processing user account data to solve at least one of the above-mentioned problems.

According to a first aspect of the present invention, there is provided a method for processing user account data, the method comprising: obtaining user account data, the user account data comprising: user industrial and commercial information, account basic information, account fund information and judicial information; inputting the user account data into a pre-trained risk identification model to output account risk information; and executing early warning operation according to the account risk information.

According to a second aspect of the present invention, there is provided an apparatus for processing user account data, the apparatus comprising: a data obtaining unit, configured to obtain user account data, where the user account data includes: user industrial and commercial information, account basic information, account fund information and judicial information; the risk information output unit is used for inputting the user account data into a pre-trained risk identification model so as to output account risk information; and the early warning unit is used for executing early warning operation according to the account risk information.

According to a third aspect of the present invention, there is provided an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method for processing user account data when executing the program.

According to a fourth aspect of the present invention, the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the above-described method of processing user account data.

According to the technical scheme, the acquired user account data is input into the pre-trained risk identification model to predict the risk of the account, and corresponding early warning operation is executed according to the risk information.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative work.

FIG. 1 is a flow diagram of a method of processing user account data according to an embodiment of the invention;

FIG. 2 is a schematic diagram of risk identification model construction according to an embodiment of the present invention;

FIG. 3 is a detailed flowchart of the account involvement prediction probability according to an embodiment of the invention;

FIG. 4 is a block diagram of a processing device for user account data according to an embodiment of the present invention;

FIG. 5 is a block diagram illustrating the detailed structure of a device for processing user account data according to an embodiment of the present invention;

FIG. 6 is a block diagram of the structure of the model training unit 46 according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of an electronic device according to an embodiment of the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

At present, for false tracking of public accounts, a bank system already establishes association information of 'account-fund transaction-enterprise relationship', however, the association information is lack of association information of cases and authorized authorities, and the false upper and lower relationships of the public accounts are not fully expressed, so that time is consumed for assisting the authorized authorities to handle cases, and the error rate is high. Moreover, because the public enterprise relation data is in the industrial and commercial institutions, the account fund transaction and the case information collaborated and checked are in the bank system, the authorized organ can not analyze the overall relation of the case related entity, and the authorized organ comprises: public security, high hospital, supervision committee, certificate supervision, bank insurance supervision of province, city and county, etc. Based on this, the embodiment of the present invention provides a method for processing user account data, so as to overcome the above problems.

An embodiment of the present invention provides a method for processing user account data, where fig. 1 is a flowchart of the method, and as shown in fig. 1, the method includes:

step 101, obtaining user account data, wherein the user account data comprises: user industrial and commercial information, account basic information, account fund information and judicial information.

The user business information may be business information of an enterprise.

The basic account information includes: business name, business number, unified credit code, unit address, registration time, registration place, etc.

The account fund information includes: fund change information of the account, and account information related to the fund change.

The judicial information includes: the account is queried by the authority for times, frozen times, scratch times, risk level, risk status, etc.

And 102, inputting the user account data into a pre-trained risk identification model to output account risk information.

And 103, executing early warning operation according to the account risk information.

The acquired user account data is input into a pre-trained risk identification model to predict the risk of the account, and corresponding early warning operation is executed according to the risk information.

In practice, a knowledge graph of the user account data may be constructed based on a graph data structure. And a knowledge graph of enterprise-account-authorized authority-fund is constructed according to the user industrial and commercial information, the account basic information, the account fund information and the judicial information in the user account data, so that the upper and lower relations of the case are more comprehensive, and the analysis of the authorized authority is facilitated.

After step 103, the knowledge-graph of the user account data may be updated according to the pre-warning operation. Specifically, when the risk identification model identifies a false account and performs judicial investigation and freezing deduction information to back investigate the associated account, the knowledge graph is updated.

For example, when a certain authorized authority makes a query (or freezing or scratching) service on an account, the bank collaborating system sends a distributed message, the system notifies to update the knowledge graph, and the knowledge graph system adds 1 to the attribute 'the number of queries by the authorized authority' of the account after receiving the message.

In one embodiment, when the banking system maintains user (or client) and account information and generates a fund flow, the banking system issues a distributed message. And the background timing task asynchronously subscribes the message and simultaneously updates the knowledge graph.

For example, when a bank account system performs a transaction of transferring from one account to another account, the bank account system sends a distributed message to notify that the knowledge graph is updated, and after receiving the message, the knowledge graph system adds the borrowing relationship between the account and another account, and assigns borrowing relationship attributes such as amount, number of strokes, time period, amount ratio and number ratio.

The risk identification model described above may be trained by: acquiring a plurality of historical user account data, and extracting characteristic information of each historical user account data; generating a training set, a verification set and a test set according to the characteristic information; and training the risk recognition model according to the training set.

In actual operation, a plurality of recognition models can be selected in advance, the recognition models are trained according to the generated training set, the trained recognition models are predicted and verified according to the test set and the verification set, and the recognition model with the optimal effect is selected from the prediction and verification results as the risk recognition model.

In one embodiment, the risk identification model may also be tuned according to predetermined rules. The predetermined rule here may be expert law or expert experience.

The embodiment of the present invention is described in detail below with reference to the risk identification model building diagram shown in fig. 2, based on an example of a bank-to-public account.

As shown in fig. 2, the risk identification model construction includes the following steps:

step 21: and constructing a knowledge graph of the enterprise account fund authority.

(1) A graph database is used to store entities and relationships between entities, such as businesses (i.e., banks versus public customers), bank accounts, authorities, and the like. Specifically, information such as enterprise business information, enterprise account information, fund transaction information, authorized department judicial inquiry frozen deduction information and the like is stored.

The bank public customer information comprises: customer name, customer number, uniform credit code, unit address, registration time, registration place, inquiry by authorized authorities, risk level, risk status, etc.

The bank account information includes: the number of queries, the number of freezes, the number of ticks, the risk level, the risk status, etc. by the authorized authority.

The relationships between the above entities are as follows:

entity relationship entity

To public client → sub-name account → account number

To public client → legal person → to public client

To public client → investment → to public client

To the public client → guarantee → to the public client

To public client → branch → to public client

Account → loan → account

Account → credit → account

Account → common terminal → account

Account → common IP → Account

Account → common MAC → Account

The graph database stores the entities and the entity attributes, and the relationships, the relationships and the relationship attributes between the entities, wherein:

(1) entities and attributes of entities are as follows:

the attributes of the entity-bank to the public customer include: customer name, customer number, universal credit code, unit address, registration time, place of registration, inquiry by authorized authorities, risk level, risk status, etc.

Attributes of the entity-account include: the number of queries, the number of freezes, the number of ticks, the risk level, the risk status, etc. by the authority.

Attributes of entity-authority include: types of authorized authorities (public security, high school, supervision committee, certificate and police, silver insurance and supervision), regions (province, city and county), etc.

(2) The relationships between entities are as follows:

entity relationship entity

To public client → sub-name account → account number

To public client → legal person → to public client

To public client → investment → to public client

To the public client → guarantee → to the public client

To public client → branch → to public client

Account → loan → account

Account → credit → account

Account → common terminal → account

Account → common IP → Account

Account → common MAC → Account

(3) Relationships and relationship attributes are as follows:

relationship-under-name Account

Relation-legal person

Relationship-investment, attributes: investment proportion, investment amount, investment date and currency

Relationship-guarantee, attribute: the guaranteed amount and currency

Branch mechanism

Relationship-borrow, attribute: amount, number of strokes, time period, ratio of amount to number of strokes

Relationship-credit, attribute: amount, number of strokes, time period, ratio of amount to number of strokes

Relationship-shared terminal, attribute: number of times

Relationship-shared IP, attribute: number of times

Relationship-common MAC, attribute: number of times

After the knowledge graph is completed based on the information, the knowledge graph is updated when account information or funds change, an authorized organization checks the account and the like.

Step 22: and (5) constructing a characteristic project.

And establishing a feature library of the public accounts, extracting features for each public account, and performing feature representation on the public accounts according to the extracted features.

Specifically, the related transaction information, IP (Internet Protocol) information, MAC (Media Access Control Address) information, and the like, and the information on the enterprise-related staff such as the legal representative are inquired from the known case-related account.

According to the inquired information, the personnel and the enterprises which have real control, high management, legal person, investment and guarantee relationship with the personnel and the enterprises, the addresses obtained by the sources such as IP sections used by online channel login, registration and the like are continuously inquired, and the exploration of the multi-degree relationship and the related information including the age, the native place, the registration time of the enterprises and the like of the related people is expanded.

Step 23: and (5) constructing a model.

(1) And extracting sample characteristics, carrying out sample characteristic labeling and generating a training sample.

(2) And (5) carrying out model training. Specifically, multiple machine learning models can be adopted for training, so as to obtain an evaluation index of each model, perform model tuning according to model effects, and select a better false-to-public user model (i.e., the risk identification model).

In an embodiment of the present invention, the machine learning model may be:

the XGBoost (eXtreme Gradient Boosting) model is an improvement of the GBDT (Gradient Boosting decision tree), and can be used for classification and regression.

The LR (Logistic Regression) model is a 0/1 classification model that is learned from features and uses linear combinations of features as arguments.

An RF (Random Forest) model is a classifier that trains and predicts a sample using a plurality of trees.

The evaluation indexes of the model include the following three types: precision, Recall, F1score (F1 value), where:

precision: the proportion of the correct regular data to the regular data is used for predicting;

recall: the proportion of the data predicted to be the positive example to the actual positive example data;

f1-score: an index for comprehensively considering precision value and call value. When the multi-class classification is performed, there are macro-average (macro-average) and micro-average (micro-average).

And for the model trained by the training set sample, carrying out model prediction through a verification set, calculating the evaluation index of the model according to the prediction result, readjusting the variable of the model if the evaluation index does not reach the expected value, and calculating the evaluation index value of the model after training and verification. This is repeated until the evaluation index value of the model reaches the expected data.

(3) And preprocessing characteristic projects of the accounts according to the public account information of the actual involved cases and the account information associated with the involved cases, predicting and outputting the account involved case probability through a model, feeding back the account involved case condition through an expert audit analysis result, and adjusting model parameters to obtain an optimal model.

In practice, the account involved in the public account opening is an account which is provided by a public security organization and is already definitely involved, but the account involved in the public account opening is generally a group account opening. After a public security organization gives an involved account, the financial institution is required to provide the associated suspected account. In the relation diagram of enterprise-account-right authority-fund constructed above, the screened relation level value is set through the relation of the famous account number, the legal person, the contact way, the actual controller, the suspected stock right actual controller, the high management, the investment, the guarantee, the branch organization, the borrowing, the lending, the shared terminal, the shared IP, the shared MAC, the paring and the relatives, and if any relation exists between the relation level value and the case-related account within the relation level range, the account is considered as the case-related account.

The account characteristic information established according to the characteristic engineering mainly comprises the following steps: the method comprises the steps of processing a transitional account, false common public account and false common public account, fictitious registration address, judicial investigation and control, a high-frequency fund collection account, case-related account affiliates, account opening time, account opening mechanisms, client age, contact information, transaction stroke number, transaction amount, loan proportion, counter-party account number, abstract type number, transaction time distribution, transaction IP distribution, transaction balance, counter-party account to public to private, counter-party account to foreign and the like, and processing characteristic information of the case-related account associated account according to a to-be-predicted collection format (such as account ID, characteristic 1 value and characteristic 2 value.) required by model prediction.

And (4) using the adjusted model, taking the set to be predicted as model input, and operating the model to output the case-involved probability of the account to be predicted. And the case detection expert feeds back whether the account is involved or not to the predicted result according to the actual occurrence result of the case, calculates the model evaluation index, adjusts the variable of the model according to the index data result and a certain step length, brings the actual involved account into the sample data, and retrains and verifies the model.

Fig. 3 is a detailed flow diagram of the account involvement prediction probability according to the embodiment of the present invention, and as shown in fig. 3, public account information and intra-row core risk database data of the public security historical involvement need to be acquired first, so as to be used in the following processes:

step 1, selecting data of the current year cycle according to an account original sample, wherein the data comprises the following data: account ID, origin characteristics, label, etc.

And 2, data exploration and pretreatment are carried out, and field and label quantity distribution is mainly subjected to statistical analysis.

And 3, performing characteristic engineering operation, and processing and selecting the characteristics of the data, wherein the method specifically comprises the following steps:

step 3.1, extracting features according to historical account transactions;

step 3.2, extracting basic static characteristics of the account;

step 3.3, extracting the risk characteristics of the core risk library, wherein the risk characteristics comprise: checking freezing buckle information and risk level;

and generating a feature table according to the feature engineering operation in the step 3.

And 4, generating final training sample data according to the data in the feature table.

And 5, segmenting the training set, the test set and the verification set data.

And 6, performing model training operation by adopting various models, and comparing, wherein the various models comprise: LR model, RF model, and XGBoost model.

Step 7, evaluating the effect of each model based on the evaluation indexes according to the test set, and carrying out tuning operation on the model adopted in the step 6 according to the effect;

step 8, obtaining an optimal model according to the model tuning operation in the step 7, and generating a model file;

after obtaining the optimal model, the optimal model can be used for carrying out account involvement prediction operation on public account information and inline account information of a new public security involvement, and the method specifically comprises the following steps:

step 9, inquiring and filtering the associated account information of the case-involved account;

and step 10, carrying out preprocessing and characteristic engineering operation on the inquired data, and then sending the data after preprocessing and characteristic engineering operation to the model file in the step 8 for prediction.

And step 11, predicting the account case-involved probability.

And step 12, carrying out manual rechecking operation.

From the above description, the optimal model can quickly predict lawless persons to create various case scenes by establishing the model and carrying out tuning processing on the model through a specialist method, so that various case types can be analyzed.

Based on similar inventive concepts, the embodiment of the present invention further provides a device for processing user account data, and preferably, the device is configured to implement the steps of the above method embodiment.

Fig. 4 is a block diagram of a structure of a device for processing user account data according to an embodiment of the present invention, as shown in fig. 4, the device including: a data acquisition unit 41, a risk information output unit 42, and an early warning unit 43, wherein:

a data obtaining unit 41, configured to obtain user account data, where the user account data includes: user worker and merchant information, account basic information, account fund information and judicial information;

a risk information output unit 42, configured to input the user account data into a risk recognition model trained in advance, so as to output account risk information;

and the early warning unit 43 is configured to perform an early warning operation according to the account risk information.

The risk information output unit 42 inputs the user account data acquired by the data acquisition unit 41 into a risk recognition model trained in advance to predict the risk of the account, and the early warning unit 43 executes corresponding early warning operation according to the risk information.

In practical operation, as shown in fig. 5, the apparatus further comprises: a knowledge-graph construction unit 44 and a knowledge-graph updating unit 45, wherein:

a knowledge graph construction unit 44, configured to construct a knowledge graph of the user account data based on a graph data structure.

Specifically, the knowledge graph constructing unit 44 constructs a knowledge graph of "enterprise-account-authorized institution-fund" according to the user business information, the account basic information, the account fund information, and the judicial information in the user account data, so that the context of the case is more comprehensive, and the analysis of the authorized institution is facilitated.

And the knowledge graph updating unit 45 is used for updating the knowledge graph of the user account data according to the early warning operation. That is, the knowledge-map updating unit 45 updates the knowledge-map when the risk identification model identifies a false account and performs judicial investigation of the information of freezing and deduction for the associated account.

In one embodiment, when a banking system maintains user (or called customer) and account information, generates a fund flow, and a judicial authority initiates a freeze-thaw service, the banking system issues a distributed message. The background timing task asynchronously subscribes to the message, and meanwhile, the knowledge graph updating unit 45 updates the knowledge graph.

With continued reference to fig. 5, the apparatus further comprises: a model training unit 46 for training the risk identification model.

Fig. 6 is a block diagram of the structure of the model training unit 46, and as shown in fig. 6, the model training unit 46 includes: a historical data acquisition module 461, a feature extraction module 462, a training set generation module 463, and a model training module 464, wherein:

a historical data acquiring module 461, configured to acquire a plurality of historical user account data;

a feature extraction module 462, configured to extract feature information of each historical user account data;

a training set generating module 463, configured to generate a training set according to the feature information;

a model training module 464, configured to train the risk identification model according to the training set.

Specifically, the model training module 464 includes: a verification test set generation submodule 4641, a model training submodule 4642, and a risk identification model determination submodule 4643, wherein:

a verification test set generation submodule 4641 configured to generate a verification set and a test set according to the feature information;

a model training submodule 4642, configured to train a plurality of preselected recognition models according to the training set, respectively;

a risk identification model determination submodule 4643, configured to determine the risk identification model from the trained multiple identification models according to the validation set and the test set. Specifically, the risk identification model determination submodule 4643 predicts and verifies the trained multiple identification models according to the test set and the verification set, and selects an identification model with the best effect as a risk identification model from the prediction and verification results.

With continued reference to fig. 5, the apparatus further comprises: and the tuning unit 47 is used for tuning the risk identification model according to a preset rule. The predetermined rule may be expert experience, and the risk identification model may be further optimized by the optimizing unit 47 to obtain a more accurate prediction.

For specific execution processes of the units, the modules, and the sub-modules, reference may be made to the description in the above method embodiment, and details are not described here again.

In practical operation, the units, the modules and the sub-modules may be combined or may be arranged singly, and the present invention is not limited thereto.

FIG. 7 is a schematic diagram of an electronic device according to an embodiment of the invention. The electronic device shown in fig. 7 is a general-purpose data processing apparatus comprising a general-purpose computer hardware structure including at least a processor 701 and a memory 702. The processor 701 and the memory 702 are connected by a bus 703. The memory 702 is adapted to store one or more instructions or programs that are executable by the processor 701. The one or more instructions or programs are executed by processor 701 to implement the steps in the method of processing user account data described above.

The processor 701 may be a stand-alone microprocessor or a combination of one or more microprocessors. Thus, the processor 701 implements the processing of data and the control of other devices by executing commands stored in the memory 702 to thereby execute the method flows of the embodiments of the present invention as described above. The bus 703 connects the above components together, as well as connecting the above components to the display controller 704 and the display device and input/output (I/O) device 705. Input/output (I/O) devices 705 may be a mouse, keyboard, modem, network interface, touch input device, motion sensing input device, printer, and other devices known in the art. Typically, input/output (I/O) devices 705 are connected to the system through an input/output (I/O) controller 706.

The memory 702 may store, among other things, software components such as an operating system, communication modules, interaction modules, and application programs. Each of the modules and applications described above corresponds to a set of executable program instructions that perform one or more functions and methods described in embodiments of the invention.

The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the steps of the processing method for user account data.

In summary, the embodiment of the present invention provides a processing scheme for user account data, and a knowledge graph of "enterprise-account-authorized institution-fund" is established based on judicial investigation of freezing and deduction information, enterprise account relationship information, and account fund information, so that the context relationship of a case is more comprehensive; moreover, by establishing a false public account identification model and adjusting the model based on a specialist law, the false public account can quickly respond to lawbreakers to create various case scenes; and a feature engineering based on an expert method and a machine learning algorithm is established, and the features of an enterprise-account-authorized authority-fund entity are extracted and expressed, so that the analysis of various case types can be satisfied. Compared with the prior art, the embodiment of the invention can assist the authorized authorities to effectively analyze the overall relation of case associated entities and more accurately predict the false public account.

The preferred embodiments of the present invention have been described above with reference to the accompanying drawings. The many features and advantages of the embodiments are apparent from the detailed specification, and thus, it is intended by the appended claims to cover all such features and advantages of the embodiments which fall within the true spirit and scope thereof. Further, since numerous modifications and changes will readily occur to those skilled in the art, it is not desired to limit the embodiments of the invention to the exact construction and operation illustrated and described, and accordingly, all suitable modifications and equivalents may be resorted to, falling within the scope thereof.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

The principle and the implementation mode of the invention are explained by applying specific embodiments in the invention, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, in light of the above description, the present invention should not be construed as limited to the embodiments and the application scope of the present invention.

Claims

1. A method for processing user account data, the method comprising:

obtaining user account data, the user account data comprising: user industrial and commercial information, account basic information, account fund information and judicial information;

inputting the user account data into a pre-trained risk identification model to output account risk information;

and executing early warning operation according to the account risk information.

2. The method for processing user account data according to claim 1, further comprising:

building a knowledge graph of the user account data based on a graph data structure.

3. The method for processing the user account data according to claim 2, wherein after performing an early warning operation according to the account risk information, the method further comprises:

and updating the knowledge graph of the user account data according to the early warning operation.

4. The method of processing user account data of claim 1, wherein the risk identification model is trained by:

acquiring a plurality of historical user account data, and extracting characteristic information of each historical user account data;

generating a training set according to the characteristic information;

and training the risk recognition model according to the training set.

5. The method for processing user account data according to claim 4, further comprising:

and carrying out tuning operation on the risk identification model according to a preset rule.

6. The method of processing user account data of claim 5, wherein training the risk identification model according to the training set comprises:

generating a verification set and a test set according to the characteristic information;

respectively training a plurality of preselected recognition models according to the training set;

determining the risk recognition model from the trained plurality of recognition models according to the validation set and the test set.

7. An apparatus for processing user account data, the apparatus comprising:

a data obtaining unit, configured to obtain user account data, where the user account data includes: user industrial and commercial information, account basic information, account fund information and judicial information;

the risk information output unit is used for inputting the user account data into a pre-trained risk identification model so as to output account risk information;

and the early warning unit is used for executing early warning operation according to the account risk information.

8. The apparatus for processing user account data according to claim 7, wherein the apparatus further comprises:

and the knowledge graph construction unit is used for constructing the knowledge graph of the user account data based on the graph data structure.

9. The apparatus for processing user account data according to claim 8, wherein the apparatus further comprises:

and the knowledge graph updating unit is used for updating the knowledge graph of the user account data according to the early warning operation.

10. The apparatus for processing user account data according to claim 7, wherein the apparatus further comprises: a model training unit for training the risk recognition model,

the model training unit includes:

the historical data acquisition module is used for acquiring a plurality of historical user account data;

the characteristic extraction module is used for extracting characteristic information of each historical user account data;

the training set generating module is used for generating a training set according to the characteristic information;

and the model training module is used for training the risk identification model according to the training set.

11. The apparatus for processing user account data according to claim 10, wherein the apparatus further comprises:

and the tuning unit is used for tuning the risk identification model according to a preset rule.

12. The apparatus for processing user account data according to claim 11, wherein the model training module comprises:

the verification test set generation submodule is used for generating a verification set and a test set according to the characteristic information;

the model training submodule is used for respectively training a plurality of preselected recognition models according to the training set;

and the risk identification model determining submodule is used for determining the risk identification model from the trained multiple identification models according to the verification set and the test set.

13. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the method for processing user account data according to any one of claims 1 to 6 when executing the program.

14. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method for processing user account data according to any one of claims 1 to 6.