CN115018627A - A credit risk assessment method and device, storage medium and electronic equipment - Google Patents

Abstract

The present invention provides a credit risk evaluation method and device, a storage medium and an electronic device, which can be applied to the financial field or other fields. The method includes: in response to a credit risk evaluation instruction, determining a user corresponding to the credit risk evaluation instruction; Obtain the credit behavior data of the user; perform feature extraction on the credit behavior data of the user through the first feature extraction module in the pre-built credit risk evaluation model to obtain the first credit behavior feature of the credit behavior data; The second feature extraction module in the credit risk evaluation model performs feature extraction on the credit behavior data of the user to obtain the second credit behavior feature of the credit behavior data; through the output module in the credit risk evaluation model based on The first credit behavior characteristic and the second credit behavior characteristic are used to obtain the credit risk evaluation result of the user. It can improve the evaluation accuracy and evaluation efficiency of credit risk.

Description

A credit risk assessment method and device, storage medium and electronic equipment

technical field

The invention relates to the technical field of data processing, and in particular, to a credit risk evaluation method and device, a storage medium and an electronic device.

Background technique

At present, with the development of economy and society, the Internet financial products of banks are developing rapidly. When banks provide credit services on the Internet, they need to conduct credit risk evaluations on the information of credit customers, so as to reduce the default risk of customers.

The existing credit risk evaluation method is based on the characteristics of the applicants designed by experts. The evaluation is highly subjective, and it is impossible to replicate the experience in batches, and the model development workload is large, resulting in low credit risk evaluation accuracy and evaluation efficiency.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to provide a credit risk evaluation method, which can improve the evaluation accuracy and evaluation efficiency of credit risk.

The present invention also provides a credit risk evaluation device to ensure the realization and application of the above method in practice.

A credit risk assessment method, including:

In response to the credit risk evaluation instruction, determining the user corresponding to the credit risk evaluation instruction;

Obtain the credit behavior data of the user;

Perform feature extraction on the credit behavior data of the user by using the first feature extraction module in the pre-built credit risk evaluation model to obtain the first credit behavior feature of the credit behavior data;

Perform feature extraction on the user's credit behavior data by the second feature extraction module in the credit risk evaluation model, to obtain the second credit behavior feature of the credit behavior data;

The credit risk evaluation result of the user is obtained based on the first credit behavior characteristic and the second credit behavior characteristic through the output module in the credit risk evaluation model.

In the above-mentioned method, optionally, the obtaining the credit risk evaluation result of the user based on the first credit behavior feature and the second credit behavior feature through the output module in the credit risk evaluation model includes:

Perform feature fusion on the first credit behavior feature and the second credit behavior feature through the feature fusion layer of the output module in the credit risk evaluation model to obtain a fusion feature;

Through the output layer of the output module in the credit risk evaluation model, the probability value between the user and each preset candidate tag is determined based on the fusion feature; the target probability with the largest value in each of the probability values is determined The candidate label to which the value belongs is used as the credit risk evaluation result of the user.

The above method, optionally, the process of constructing the credit risk evaluation model includes:

Obtain a training data set and an initial credit risk evaluation model; the training data set includes a plurality of training samples, and each of the training samples includes historical credit behavior data of a historical user with a sample label; the sample label represents the Credit risk evaluation results of historical users;

Use each training sample in the training data set to train the initial credit risk evaluation model until the initial credit risk evaluation model satisfies preset training conditions;

An initial credit risk evaluation model that satisfies the training conditions is determined as a credit risk evaluation model.

In the above method, optionally, the first feature extraction module in the pre-built credit risk evaluation model performs feature extraction on the credit behavior data of the user, and obtains the first credit behavior feature of the credit behavior data, including:

Through the convolutional neural network in the first feature extraction module of the credit risk evaluation model, feature extraction is performed on the credit behavior data to obtain the initial characteristics of the first credit behavior;

Through the multi-head attention mechanism layer in the first feature extraction module of the credit risk evaluation model, the initial feature of the first credit behavior is processed to obtain the first credit behavior feature of the credit behavior data.

The above method, optionally, after obtaining the credit risk evaluation result of the user, further comprising:

In the case that the credit risk evaluation result of the user satisfies the preset credit risk warning condition, output the credit risk warning information for the user.

A credit risk assessment device, comprising:

a determining unit, configured to, in response to the credit risk evaluation instruction, determine the user corresponding to the credit risk evaluation instruction;

an obtaining unit, used to obtain the credit behavior data of the user;

a first feature extraction unit, configured to perform feature extraction on the credit behavior data of the user through the first feature extraction module in the pre-built credit risk evaluation model, to obtain the first credit behavior feature of the credit behavior data;

a second feature extraction unit, which performs feature extraction on the credit behavior data of the user through the second feature extraction module in the credit risk evaluation model to obtain the second credit behavior feature of the credit behavior data;

An output unit, configured to obtain a credit risk evaluation result of the user based on the first credit behavior feature and the second credit behavior feature through an output module in the credit risk evaluation model.

The above-mentioned device, optionally, the output unit includes:

a feature fusion subunit, configured to perform feature fusion on the first credit behavior feature and the second credit behavior feature through the feature fusion layer of the output module in the credit risk evaluation model to obtain a fusion feature;

The output subunit is used to determine the probability value between the user and each preset candidate tag based on the fusion feature through the output layer of the output module in the credit risk evaluation model; each of the probability values The candidate tag to which the target probability value with the largest median value belongs is used as the credit risk evaluation result of the user.

The above-mentioned device, optionally, further includes: a model training unit; the model training unit is used for:

Obtain a training data set and an initial credit risk evaluation model; the training data set includes a plurality of training samples, and each of the training samples includes historical credit behavior data of a historical user with a sample label; the sample label represents the Credit risk evaluation results of historical users;

Use each training sample in the training data set to sequentially train the initial credit risk evaluation model until the initial credit risk evaluation model satisfies preset training conditions;

An initial credit risk evaluation model that satisfies the training conditions is determined as a credit risk evaluation model.

In the above device, optionally, the first feature extraction unit is configured as:

Through the convolutional neural network in the first feature extraction module of the credit risk evaluation model, feature extraction is performed on the credit behavior data to obtain the initial characteristics of the first credit behavior;

Through the multi-head attention mechanism layer in the first feature extraction module of the credit risk evaluation model, the initial feature of the first credit behavior is processed to obtain the first credit behavior feature of the credit behavior data.

For the above device, optionally, the credit risk assessment device further includes: an alarm unit;

The alarming unit is configured to output credit risk alarm information for the user when the user's credit risk evaluation result satisfies a preset credit risk alarm condition.

A storage medium, the storage medium comprising a storage instruction, wherein when the instruction is executed, a device where the storage medium is located is controlled to execute the above-mentioned credit risk assessment method.

An electronic device includes a memory, and one or more instructions, wherein the one or more instructions are stored in the memory and configured to perform, by one or more processors, a credit risk assessment method as described above.

Compared with the prior art, the present invention includes the following advantages:

The present invention provides a credit risk evaluation method and device, a storage medium and an electronic device. The method includes: in response to a credit risk evaluation instruction, determining a user corresponding to the credit risk evaluation instruction; acquiring credit behavior data of the user; The first feature extraction module in the pre-built credit risk evaluation model performs feature extraction on the credit behavior data of the user to obtain the first credit behavior feature of the credit behavior data; through the first feature extraction module in the credit risk evaluation model The second feature extraction module performs feature extraction on the credit behavior data of the user, and obtains the second credit behavior feature of the credit behavior data; the output module in the credit risk evaluation model is based on the first credit behavior feature and all the The second credit behavior characteristic is obtained, and the credit risk evaluation result of the user is obtained. By applying the method provided by the embodiments of the present invention, the user's credit risk evaluation can be performed by extracting features of different levels of credit behavior data, which can improve the evaluation accuracy and evaluation efficiency of credit risk.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only It is an embodiment of the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to the provided drawings without creative work.

Fig. 1 is a method flow chart of a credit risk assessment method provided by the present invention;

2 is a flowchart of a process for obtaining the credit risk evaluation result of the user provided by the present invention;

3 is a flowchart of a process for constructing a credit risk evaluation model provided by the present invention;

4 is an exemplary diagram of a process for obtaining a first credit behavior feature of credit behavior data provided by the present invention;

5 is a schematic structural diagram of a credit risk evaluation model provided by the present invention;

6 is a schematic structural diagram of a credit risk assessment device provided by the present invention;

FIG. 7 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

In this application, the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device comprising a list of elements includes not only those elements, but also no Other elements expressly listed, or which are also inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

An embodiment of the present invention provides a credit risk evaluation method, which can be applied to electronic equipment. The method flowchart of the method is shown in FIG. 1 , and specifically includes:

S101: In response to a credit risk evaluation instruction, determine a user corresponding to the credit risk evaluation instruction.

In this embodiment, the credit risk evaluation instruction may be an instruction triggered when a credit risk evaluation needs to be performed for the user, for example, it may be an instruction triggered in the case of handling a loan business, a credit leasing business, etc. for the user, or an instruction triggered by An instruction that is automatically triggered when the user meets the preset credit evaluation conditions.

S102: Acquire the credit behavior data of the user.

In this embodiment, the user's new user behavior data may be acquired in a preset database based on the user's identifier, or the user's credit behavior data may be acquired from the instruction information of the credit risk evaluation instruction.

Optionally, the credit behavior data may include the user's performance behavior information, transaction behavior information, and other user behavior information that affects credit.

S103: Perform feature extraction on the credit behavior data of the user by using a first feature extraction module in a pre-built credit risk evaluation model to obtain a first credit behavior feature of the credit behavior data.

In this embodiment, the first feature extraction module can extract important local features for the user's credit risk evaluation, the local features are the first credit behavior features, and the global features of the credit behavior data can be weighted processed.

S104: Perform feature extraction on the credit behavior data of the user by using the second feature extraction module in the credit risk evaluation model to obtain a second credit behavior feature of the credit behavior data.

In this embodiment, the second credit behavior feature may be a global feature of the credit behavior data.

S105: Obtain a credit risk evaluation result of the user based on the first credit behavior feature and the second credit behavior feature through the output module in the credit risk evaluation model.

In this embodiment, the output module may fuse the first credit behavior feature and the second credit behavior feature to obtain the fused feature, and obtain the user's credit risk evaluation result according to the fused feature.

Optionally, the credit risk evaluation result may be a credit risk rating, or a credit risk score.

An embodiment of the present invention provides a credit risk evaluation method, including: in response to a credit risk evaluation instruction, determining a user corresponding to the credit risk evaluation instruction; acquiring credit behavior data of the user; and using a pre-built credit risk evaluation model The first feature extraction module in the credit behavior data of the user performs feature extraction to obtain the first credit behavior feature of the credit behavior data; through the second feature extraction module in the credit risk evaluation model, the user is Perform feature extraction on the credit behavior data, and obtain the second credit behavior feature of the credit behavior data; through the output module in the credit risk evaluation model, based on the first credit behavior feature and the second credit behavior feature, obtain The credit risk evaluation result of the user. By applying the method provided by the embodiment of the present invention, the user's credit risk evaluation can be performed by extracting features of different levels of credit behavior data, which can improve the evaluation accuracy and evaluation efficiency of credit risk.

In an embodiment provided by the present invention, based on the above implementation process, optionally, the output module in the credit risk evaluation model is based on the first credit behavior feature and the second credit behavior feature, Obtain the credit risk evaluation result of the user, as shown in Figure 2, including:

S201: Perform feature fusion on the first credit behavior feature and the second credit behavior feature through a feature fusion layer of an output module in the credit risk evaluation model to obtain a fusion feature.

In this embodiment, the first credit behavior feature and the second credit behavior feature can be spliced to obtain a fusion feature. The splicing method can be splicing the first credit behavior feature after the second credit behavior feature, or the second The credit behavior feature is spliced after the first credit behavior feature. By fusing the first credit behavior feature and the second credit behavior feature, the fusion feature is obtained, and the features that have an important impact on the credit evaluation can be obtained, and the overall situation is reduced. loss of information.

S202: Determine the probability value between the user and each preset candidate tag based on the fusion feature through the output layer of the output module in the credit risk evaluation model; determine the probability value with the largest value among the probability values The candidate label to which the target probability value belongs is used as the credit risk evaluation result of the user.

In this embodiment, the candidate tags can represent the user's risk classification, and different candidate tags represent users with different risk classifications. The output layer can determine the probability value between each candidate tag according to the fusion feature. The probability value of each candidate tag determines the user's credit risk evaluation result, and specifically, the candidate tag with the largest probability value may be used as the user's credit risk evaluation result.

In an embodiment provided by the present invention, based on the above-mentioned implementation process, optionally, the process of constructing the credit risk evaluation model, as shown in FIG. 3 , includes:

S301: Obtain a training data set and an initial credit risk evaluation model; the training data set includes a plurality of training samples, each of which includes historical credit behavior data of a historical user and a sample label; the sample label represents the Credit risk evaluation results of historical users.

In this embodiment, the sample tag, that is, the above-mentioned candidate tag, can tag the credit risk evaluation result of the user's historical user.

S302: Use each training sample in the training data set to train the initial credit risk evaluation model until the initial credit risk evaluation model satisfies preset training conditions.

In this embodiment, a residual module can be set for the first feature extraction module of the initial credit risk evaluation model, and the initial credit risk evaluation model with residual modules can be trained by training samples in the training data. Input the sample into the initial credit risk evaluation model to obtain the evaluation result corresponding to the training sample output by the initial credit risk evaluation model; calculate the loss function value from the evaluation result and the sample label of the training sample; adjust the initial credit risk according to the loss function value Network parameters for risk assessment models.

Optionally, the training condition may be that the number of training times of the initial credit risk evaluation model is greater than the preset number of training thresholds, or that the prediction accuracy of the initial credit risk evaluation model is greater than the preset accuracy threshold, or the initial credit risk Evaluate the convergence of the loss function of the model.

S303: Determine the initial credit risk evaluation model that satisfies the training condition as a credit risk evaluation model.

Wherein, the initial credit risk evaluation model without the residual module that satisfies the training conditions can be determined as the credit risk evaluation model.

In an embodiment provided by the present invention, based on the above implementation process, optionally, the first feature extraction module in a pre-built credit risk evaluation model performs feature extraction on the credit behavior data of the user, and obtains the The first credit behavior feature of the credit behavior data, as shown in Figure 4, includes:

S401: Perform feature extraction on the credit behavior data through a convolutional neural network in a first feature extraction module of the credit risk evaluation model to obtain initial features of a first credit behavior.

In this embodiment, the convolutional neural network in the first feature extraction module includes a plurality of convolutional layers, and each convolutional layer processes the credit behavior data to obtain the initial characteristics of the first credit behavior.

S402: Process the initial features of the first credit behavior through the multi-head attention mechanism layer in the first feature extraction module of the credit risk evaluation model to obtain the first credit behavior feature of the credit behavior data.

In this embodiment, the multi-head attention mechanism layer includes a plurality of parallel attention mechanism layers, and the first credit behavior feature is obtained by processing the initial feature of the first credit behavior through the multi-head attention mechanism layer.

In an embodiment provided by the present invention, based on the above-mentioned implementation process, optionally, after obtaining the credit risk evaluation result of the user, the method further includes:

In the case that the credit risk evaluation result of the user satisfies the preset credit risk warning condition, output the credit risk warning information for the user.

In this embodiment, the credit risk evaluation result may be a credit risk score or a credit risk classification; if the credit risk evaluation result is a credit risk score, if the credit risk score is greater than a preset risk threshold, the user's risk evaluation is determined The result satisfies the credit risk warning condition; when the credit risk evaluation result is the credit risk classification, if the credit risk classification is greater than the preset alarm classification, it is determined that the user's risk evaluation result satisfies the credit risk warning condition.

Referring to FIG. 5, it is a schematic structural diagram of a credit risk evaluation model provided by an embodiment of the present invention, including that the credit risk evaluation model includes a first feature extraction module, a second feature extraction module, and an output module, and the first feature extraction module may include a volume The second feature extraction module can include a convolutional neural network with at least one convolutional layer; the output module can include a feature fusion layer and an output layer, and the feature fusion layer and the output layer can also be Including the middle layer.

The multi-head attention mechanism layer in this embodiment can focus on the focus of the application scenario, while ignoring some unimportant factors. It can be regarded as a combination function, by calculating the probability distribution of attention, to highlight the impact of a key input on the output.

The attention mechanism can be described as mapping a query and a set of corresponding key-value pairs to the output, where the query is Q, the key is K, the value is V, and the output is output. These elements are all vectors. In general, key K and value V are equal. The relationship between them can be expressed by the following formula:

where d _k is the dimension of K that will affect the size of the dot product. softmax is an activation function that can be used to normalize weights, where Z is a K-dimensional vector and Z _j represents an element in this K-dimensional vector. With softmax, we can normalize the elements in the vector to be between 0 and 1, and let the sum of these elements be 1.

Specifically, the key and value are all the features in the credit behavior data extracted by the convolutional neural network, and the query corresponds to the weight matrix that the attention module needs to learn. The output is a weighted sum of values (features of the credit behavior data), where the weight assigned to each value is computed by querying the compatibility function with the corresponding key. Through the attention mechanism, the deep learning model can be more focused on the information that is more important for the precipitation prediction task.

The multi-head attention mechanism is composed of several parallel attention mechanism layers with different training parameters, and each head performs a linear transformation before the attention operation to project the three inputs to a lower dimension . The operation of each attention mechanism is performed independently, and then the final result is obtained by concatenating the outputs of each head. Specifically, the input to the multi-head attention layer is three sequences of vectors: Q, K, and V. For the ith head, the formula of the attention mechanism is as follows:

是用来将三个输入映射到具有较低纬度d_p的子空间的权重矩阵，这三个矩阵的参数是需要模型去学习的。然后多头注意机制的输出是将以上每个头的结果进行综合，多头注意力机制的公式如下：it's here,

is the weight matrix used to map the three inputs to the subspace with lower latitude _dp , and the parameters of these three matrices need to be learned by the model. Then the output of the multi-head attention mechanism is to synthesize the results of each of the above heads. The formula of the multi-head attention mechanism is as follows:

Multihead(Q,K,V)=Concat(head ₁ ,head ₂ ,...,head _h )W ^O

也是一个需要模型去学习的权重矩阵。Here, h is the number of heads,

It is also a weight matrix that the model needs to learn.

The advantage of multi-head attention is that it can learn relevant information in different subspaces. However, due to the prominence of important features, some secondary features may be lost, resulting in incomplete global information.

Since the depth of a simple stacked network cannot increase the learning expression effect of the network, this is because as the depth of the network increases, it will lead to gradient divergence and loss of information. Therefore, by introducing the residual module, the problem of gradient disappearance with the increase of network depth can be solved. Furthermore, residual connections avoid losing global features to ensure the integrity of the original information. Therefore, in this embodiment, the multi-head attention mechanism is combined with the residual connection to avoid information loss caused by the attention mechanism. The improved formula is as follows:

REAT(f,X)=X+f(X)

where f(X)=MultiHead(X,X,X). In the credit risk evaluation model, X is the feature of the credit behavior data extracted by the convolutional neural network.

The multi-head attention mechanism adopted in this embodiment is a self-multi-head attention mechanism, in which the key K, the value V and the query Q are all the same tensor sequence X. In this way, each row vector in the feature matrix is the dot product of all column vectors, and the first feature extraction module provided by the embodiment of the present invention can extract more comprehensive features, that is, fuse global features and local features.

The goal of the credit evaluation model is to extract the high-level features of personal credit through deep networks to achieve risk evaluation and prediction. In order to better capture the characteristics of personal credit, the model includes two channels: the first feature extraction module and the second feature extraction module. Among them, the first feature extraction module is used to extract high-level semantics for personal credit behavior, and the second feature extraction module is responsible for obtaining deeper feature expressions from the non-semantic features that have been obtained. In the channel of the first feature extraction module Taking advantage of self-multi-head attention and residual connections, the loss of global information is reduced while paying special attention to the factors that have an important impact on credit evaluation. Finally, the features extracted from these two channels are spliced and sent to the fully connected layer to obtain the credit evaluation output.

Corresponding to the method described in FIG. 1 , an embodiment of the present invention further provides a credit risk assessment apparatus for implementing the method in FIG. 1 . The credit risk assessment apparatus provided by the embodiment of the present invention can be applied to electronic equipment. , the schematic diagram of its structure is shown in Figure 6, which specifically includes:

A determination unit 601, configured to, in response to a credit risk evaluation instruction, determine a user corresponding to the credit risk evaluation instruction;

an obtaining unit 602, configured to obtain the credit behavior data of the user;

The first feature extraction unit 603 is configured to perform feature extraction on the credit behavior data of the user through the first feature extraction module in the pre-built credit risk evaluation model to obtain the first credit behavior feature of the credit behavior data;

The second feature extraction unit 604 is to perform feature extraction on the credit behavior data of the user through the second feature extraction module in the credit risk evaluation model to obtain the second credit behavior feature of the credit behavior data;

The output unit 605 is configured to obtain the credit risk evaluation result of the user based on the first credit behavior feature and the second credit behavior feature through the output module in the credit risk evaluation model.

By applying the device provided by the embodiment of the present invention, the user's credit risk evaluation can be performed by extracting features of different levels of credit behavior data, which can improve the evaluation accuracy and evaluation efficiency of credit risk.

In an embodiment provided by the present invention, based on the above solution, optionally, the output unit 605 includes:

a feature fusion subunit, configured to perform feature fusion on the first credit behavior feature and the second credit behavior feature through the feature fusion layer of the output module in the credit risk evaluation model to obtain a fusion feature;

The output subunit is used to determine the probability value between the user and each preset candidate tag based on the fusion feature through the output layer of the output module in the credit risk evaluation model; each of the probability values The candidate tag to which the target probability value with the largest median value belongs is used as the credit risk evaluation result of the user.

In an embodiment provided by the present invention, based on the above solution, optionally, it further includes: a model training unit; the model training unit is used for:

Obtain a training data set and an initial credit risk evaluation model; the training data set includes a plurality of training samples, and each of the training samples includes historical credit behavior data of a historical user with a sample label; the sample label represents the Credit risk evaluation results of historical users;

Use each training sample in the training data set to sequentially train the initial credit risk evaluation model until the initial credit risk evaluation model satisfies preset training conditions;

An initial credit risk evaluation model that satisfies the training conditions is determined as a credit risk evaluation model.

In an embodiment provided by the present invention, based on the above solution, optionally, the first feature extraction unit 603 is configured as:

Through the convolutional neural network in the first feature extraction module of the credit risk evaluation model, feature extraction is performed on the credit behavior data to obtain the initial characteristics of the first credit behavior;

Through the multi-head attention mechanism layer in the first feature extraction module of the credit risk evaluation model, the initial feature of the first credit behavior is processed to obtain the first credit behavior feature of the credit behavior data.

In an embodiment provided by the present invention, based on the above solution, optionally, the credit risk assessment apparatus further includes: an alarm unit;

The alarming unit is configured to output credit risk alarm information for the user when the user's credit risk evaluation result satisfies a preset credit risk alarm condition.

The specific principles and execution processes of each unit and module in the credit risk assessment apparatus disclosed in the above embodiments of the present invention are the same as the credit risk assessment methods disclosed in the above embodiments of the present invention. For reference, please refer to the credit risk assessment provided by the above embodiments of the present invention The corresponding part in the method will not be repeated here.

An embodiment of the present invention further provides a storage medium, where the storage medium includes stored instructions, wherein when the instructions are executed, a device where the storage medium is located is controlled to execute the above credit risk assessment method.

An embodiment of the present invention further provides an electronic device, the schematic structural diagram of which is shown in FIG. 7 , and specifically includes a memory 701 and one or more instructions 702 , wherein one or more instructions 702 are stored in the memory 701 and are processed through the memory 701 . The one or more instructions 702 are configured to be executed by one or more processors 703 to:

In response to the credit risk evaluation instruction, determining the user corresponding to the credit risk evaluation instruction;

Obtain the credit behavior data of the user;

Perform feature extraction on the credit behavior data of the user by using the first feature extraction module in the pre-built credit risk evaluation model to obtain the first credit behavior feature of the credit behavior data;

Perform feature extraction on the user's credit behavior data by the second feature extraction module in the credit risk evaluation model, to obtain the second credit behavior feature of the credit behavior data;

The credit risk evaluation result of the user is obtained based on the first credit behavior characteristic and the second credit behavior characteristic through the output module in the credit risk evaluation model.

It should be noted that a credit risk evaluation method and device, storage medium and electronic equipment provided by the present invention can be used in the field of artificial intelligence, blockchain, distributed, cloud computing, big data, Internet of Things, Mobile internet, network security, chip, virtual reality, augmented reality, holographic technology, quantum computing, quantum communication, quantum measurement, digital twin or finance. The above is only an example, and does not limit the application field of a credit risk evaluation method and device, storage medium and electronic device provided by the present invention.

It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. For the same and similar parts among the various embodiments, refer to each other Can. As for the apparatus type embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant part, please refer to the partial description of the method embodiment.

Finally, it should also be noted that in this document, relational terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply these entities or there is any such actual relationship or sequence between operations. Moreover, the terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion such that a process, method, article or device comprising a list of elements includes not only those elements, but also includes not explicitly listed or other elements inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

For the convenience of description, when describing the above device, the functions are divided into various units and described respectively. Of course, when implementing the present invention, the functions of each unit may be implemented in one or more software and/or hardware.

From the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solutions of the present invention can be embodied in the form of software products in essence or the parts that make contributions to the prior art. The computer software products can be stored in storage media, such as ROM/RAM, magnetic disks, etc. , CD, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

A credit risk evaluation method provided by the present invention has been introduced in detail above. Specific examples are used in this paper to illustrate the principles and implementations of the present invention. Its core idea; at the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation and application scope. limit.

Claims (10)

1. A credit risk assessment method, characterized in that, comprising: In response to the credit risk evaluation instruction, determining the user corresponding to the credit risk evaluation instruction; Obtain the credit behavior data of the user; Perform feature extraction on the credit behavior data of the user by using the first feature extraction module in the pre-built credit risk evaluation model to obtain the first credit behavior feature of the credit behavior data; Perform feature extraction on the user's credit behavior data by the second feature extraction module in the credit risk evaluation model, to obtain the second credit behavior feature of the credit behavior data; The credit risk evaluation result of the user is obtained based on the first credit behavior characteristic and the second credit behavior characteristic through the output module in the credit risk evaluation model. 2 . The method according to claim 1 , wherein, obtaining the user’s profile based on the first credit behavior feature and the second credit behavior feature through an output module in the credit risk evaluation model. 3 . Credit risk assessment results, including: Perform feature fusion on the first credit behavior feature and the second credit behavior feature through the feature fusion layer of the output module in the credit risk evaluation model to obtain a fusion feature; Through the output layer of the output module in the credit risk evaluation model, the probability value between the user and each preset candidate tag is determined based on the fusion feature; the target probability with the largest value in each of the probability values is determined The candidate label to which the value belongs is used as the credit risk evaluation result of the user. 3. The method according to claim 1, wherein the process of constructing the credit risk evaluation model comprises: Obtain a training data set and an initial credit risk evaluation model; the training data set includes a plurality of training samples, and each of the training samples includes historical credit behavior data of a historical user with a sample label; the sample label represents the Credit risk evaluation results of historical users; Use each training sample in the training data set to train the initial credit risk evaluation model until the initial credit risk evaluation model satisfies preset training conditions; An initial credit risk evaluation model that satisfies the training conditions is determined as a credit risk evaluation model. 4 . The method according to claim 1 , wherein the first feature extraction module in the pre-built credit risk evaluation model performs feature extraction on the credit behavior data of the user, and obtains the first feature of the credit behavior data. 5 . -Characteristics of credit behavior, including: Through the convolutional neural network in the first feature extraction module of the credit risk evaluation model, feature extraction is performed on the credit behavior data to obtain the initial characteristics of the first credit behavior; Through the multi-head attention mechanism layer in the first feature extraction module of the credit risk evaluation model, the initial feature of the first credit behavior is processed to obtain the first credit behavior feature of the credit behavior data. 5. The method according to claim 1, wherein after obtaining the credit risk evaluation result of the user, the method further comprises: In the case that the credit risk evaluation result of the user satisfies the preset credit risk warning condition, output the credit risk warning information for the user. 6. A credit risk assessment device, comprising: a determining unit, configured to, in response to the credit risk evaluation instruction, determine the user corresponding to the credit risk evaluation instruction; an obtaining unit, used to obtain the credit behavior data of the user; a first feature extraction unit, configured to perform feature extraction on the credit behavior data of the user through the first feature extraction module in the pre-built credit risk evaluation model, to obtain the first credit behavior feature of the credit behavior data; a second feature extraction unit, which performs feature extraction on the credit behavior data of the user through the second feature extraction module in the credit risk evaluation model to obtain the second credit behavior feature of the credit behavior data; An output unit, configured to obtain a credit risk evaluation result of the user based on the first credit behavior feature and the second credit behavior feature through an output module in the credit risk evaluation model. 7. The device according to claim 6, wherein the output unit comprises: a feature fusion subunit, configured to perform feature fusion on the first credit behavior feature and the second credit behavior feature through the feature fusion layer of the output module in the credit risk evaluation model to obtain a fusion feature; The output subunit is used to determine the probability value between the user and each preset candidate tag based on the fusion feature through the output layer of the output module in the credit risk evaluation model; each of the probability values The candidate tag to which the target probability value with the largest median value belongs is used as the credit risk evaluation result of the user. 8. The apparatus according to claim 6, further comprising: a model training unit; the model training unit is used for: Obtain a training data set and an initial credit risk evaluation model; the training data set includes a plurality of training samples, and each of the training samples includes historical credit behavior data of a historical user with a sample label; the sample label represents the Credit risk evaluation results of historical users; Use each training sample in the training data set to sequentially train the initial credit risk evaluation model until the initial credit risk evaluation model satisfies preset training conditions; An initial credit risk evaluation model that satisfies the training conditions is determined as a credit risk evaluation model. 9 . A storage medium, characterized in that the storage medium comprises a storage instruction, wherein when the instruction is executed, a device where the storage medium is located is controlled to execute the credit risk according to any one of claims 1 to 5 Evaluation method. 10. An electronic device, comprising a memory, and one or more instructions, wherein the one or more instructions are stored in the memory and configured to be executed by one or more processors as claimed in claims 1- 5 any one of the credit risk assessment methods described.

Images