CN117036022A

CN117036022A - Financing risk prediction method and device, storage medium and electronic equipment

Info

Publication number: CN117036022A
Application number: CN202310996993.1A
Authority: CN
Inventors: 祝福松
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2023-08-08
Filing date: 2023-08-08
Publication date: 2023-11-10

Abstract

The application discloses a financing risk prediction method and device, a storage medium and electronic equipment, and relates to the technical field of artificial intelligence. The method comprises the following steps: responding to a financing request sent by a first object and acquiring attribute information of the first object; inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; acquiring a knowledge graph; and determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks. According to the application, the problem that in the related technology, when the enterprise applies for financing to a financial institution, only whether a single enterprise is an inauguration enterprise can be identified, other inauguration enterprises associated with the inauguration enterprise are difficult to predict, and the risk of financing is difficult to accurately predict is solved.

Description

Financing risk prediction method and device, storage medium and electronic equipment

Technical Field

The application relates to the technical field of artificial intelligence, in particular to a financing risk prediction method and device, a storage medium and electronic equipment.

Background

In the process of operating an enterprise, as the enterprise scale is larger, the more funds are needed, any enterprise may face insufficient funds, and financing at this time becomes a solution to the problem. Moreover, general financing means are mainly by loaning to financial institutions. In addition, for financial institutions, various risks, such as default risks, etc., may be faced with the financing request of the enterprise, resulting in loss of the financial institution. The risk assessment model in the related art generally starts from the entity itself and analyzes the change of the entity according to the attribute characteristics of the entity itself, but the trend change of a single entity is not limited to the entity itself, but also propagates to other entities associated with the entity, so that the predicted risk is increased. I.e. the risk that it is difficult to accurately predict financing.

Aiming at the problems that in the related art, when enterprises apply for financing to financial institutions, whether a single enterprise is an inauguration enterprise can only be identified, other inauguration enterprises related to the inauguration enterprise are difficult to predict, and the inauguration risk of financing is difficult to accurately predict, no effective solution is proposed at present.

Disclosure of Invention

The application mainly aims to provide a financing risk prediction method and device, a storage medium and electronic equipment, and aims to solve the problems that in the related technology, when enterprises apply for financing to financial institutions, only whether a single enterprise is an inauguration enterprise can be identified, other inauguration enterprises associated with the inauguration enterprise are difficult to predict, and the inauguration risk of financing is difficult to accurately predict.

In order to achieve the above object, according to one aspect of the present application, there is provided a risk prediction method of financing. The method comprises the following steps: responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution; inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; acquiring a knowledge graph, wherein the knowledge graph at least comprises information of the first object and object information associated with the first object; and determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks, wherein N is a positive integer.

Further, determining the N objects at risk according to the classification result and the knowledge-graph includes: if the classification result indicates that the first object is an object with risk, determining M objects associated with the first object based on the knowledge graph, wherein M is a positive integer and is smaller than N; and taking the first object and the M objects as the N objects with risks.

Further, obtaining the knowledge graph includes: acquiring information of S second objects, wherein the information of the S second objects at least comprises information of the first object and object information associated with the first object, the information of each second object at least comprises attribute information of each second object, S is a positive integer, and S is larger than M; determining the association relation between every two second objects according to the attribute information of each second object; and acquiring the knowledge graph based on the information of the S second objects and the association relation between every two second objects.

Further, the object classification model is obtained by: acquiring a training sample set for model training, wherein the training sample set at least comprises T first attribute information samples and R second attribute information samples, the first attribute information samples are attribute information of objects with risks, the second attribute information samples are attribute information of objects without risks, and T and R are positive integers larger than 1; and learning and training the original classification model by adopting the training sample set to obtain the target classification model.

Further, after obtaining the knowledge-graph, the method further includes: judging whether the association relation between every two second objects in the knowledge graph needs to be updated or not; if the association relation between every two second objects in the knowledge graph does not need to be updated, determining the object with risk according to the knowledge graph; if the association relation between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph; and determining the object with risk according to the updated knowledge graph.

Further, if the association relationship between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph includes: if the association relation between every two second objects in the knowledge graph needs to be updated, acquiring the second objects needing to update the association relation, and acquiring the association relation needing to be updated; inputting the second object needing to update the association relationship and the association relationship needing to be updated into a link prediction model, and outputting Q second objects, wherein Q is a positive integer, and Q is smaller than S; and updating the knowledge graph according to the second object needing to update the association relationship, the association relationship needing to update and the Q second objects to obtain the updated knowledge graph.

Further, predicting the risk of the financing request based on the N risky objects comprises: acquiring the number of the N objects with risks; judging whether the number of the N objects with risks is larger than a preset number or not; if the number of the N objects with risks is larger than the preset number, predicting that the risk level corresponding to the financing request is a first risk level; and if the number of the N objects with risks is not greater than the preset number, predicting that the risk level corresponding to the financing request is a second risk level, wherein the risk corresponding to the second risk level is lower than the risk corresponding to the first risk level.

In order to achieve the above object, according to another aspect of the present application, there is provided a risk prediction apparatus for financing. The device comprises: the first processing unit is used for responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution; the first input unit is used for inputting attribute information of the first object into a target classification model to carry out classification processing and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; a first obtaining unit, configured to obtain a knowledge graph, where the knowledge graph at least includes information of the first object and object information associated with the first object; and the second processing unit is used for determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks, wherein N is a positive integer.

Further, the second processing unit includes: the first determining module is used for determining M objects associated with the first object based on the knowledge graph if the classification result indicates that the first object is an object with risk, wherein M is a positive integer and M is smaller than N; and the second determining module is used for taking the first object and the M objects as N objects with risks.

Further, the first acquisition unit includes: the first acquisition module is used for acquiring information of S second objects, wherein the information of the S second objects at least comprises information of the first object and object information associated with the first object, the information of each second object at least comprises attribute information of each second object, S is a positive integer, and S is larger than M; the third determining module is used for determining the association relation between every two second objects according to the attribute information of each second object; and the second acquisition module is used for acquiring the knowledge graph based on the information of the S second objects and the association relation between every two second objects.

Further, the object classification model is obtained by: the second acquisition unit is used for acquiring a training sample set for model training, wherein the training sample set at least comprises T first attribute information samples and R second attribute information samples, the first attribute information samples are attribute information of objects with risks, the second attribute information samples are attribute information of objects without risks, and T and R are positive integers larger than 1; and the first training unit is used for learning and training the original classification model by adopting the training sample set to obtain the target classification model.

Further, the apparatus further comprises: the first judging unit is used for judging whether the association relation between every two second objects in the knowledge graph needs to be updated after the knowledge graph is acquired; the first determining unit is used for determining an object with risk according to the knowledge graph if the association relation between every two second objects in the knowledge graph does not need to be updated; the first updating unit is used for updating the knowledge graph to obtain an updated knowledge graph if the association relationship between every two second objects in the knowledge graph needs to be updated; and the second determining unit is used for determining the object with risk according to the updated knowledge graph.

Further, the first updating unit includes: the third acquisition module is used for acquiring the second objects needing to update the association relationship and acquiring the association relationship needing to be updated if the association relationship between every two second objects in the knowledge graph needs to be updated; the first input module is used for inputting the second object needing to update the association relation and the association relation needing to be updated into a link prediction model and outputting Q second objects, wherein Q is a positive integer and is smaller than S; and the first updating module is used for updating the knowledge graph according to the second object needing to update the association relationship, the association relationship needing to be updated and the Q second objects to obtain the updated knowledge graph.

Further, the second processing unit includes: a fourth obtaining module, configured to obtain the number of the N objects that are at risk; the first judging module is used for judging whether the number of the N objects with risks is larger than a preset number or not; the first prediction module is used for predicting that the risk level corresponding to the financing request is a first risk level if the number of the N objects with risks is larger than the preset number; and the second prediction module is used for predicting the risk level corresponding to the financing request to be a second risk level if the number of the N objects with risks is not greater than the preset number, wherein the risk corresponding to the second risk level is lower than the risk corresponding to the first risk level.

In order to achieve the above object, according to another aspect of the present application, there is provided a computer-readable storage medium storing a program, wherein the program performs the risk prediction method of financing according to any one of the above.

To achieve the above object, according to another aspect of the present application, there is provided an electronic device including one or more processors and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the risk prediction method of financing according to any one of the above.

According to the application, the following steps are adopted: responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution; inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; acquiring a knowledge graph, wherein the knowledge graph at least comprises information of a first object and object information associated with the first object; n objects with risks are determined according to the classification result and the knowledge graph, and the risks of financing requests are predicted based on the N objects with risks, wherein N is a positive integer, so that the problem that in the related art, when enterprises apply for financing to financial institutions, only whether a single enterprise is a risk enterprise can be identified, other risk enterprises associated with the risk enterprise are difficult to predict, and the risks of financing are difficult to accurately predict is solved. Responding to a financing request sent by a first object, acquiring attribute information of the first object, inputting the attribute information of the first object into a target classification model for classification processing, outputting a classification result of whether the first object is an object with risk or not, determining N objects with risk according to the classification result and a knowledge graph, and predicting the risk of the financing request based on the N objects with risk, so that when an enterprise applies for financing to a financial institution, other risk enterprises associated with the risk enterprises can be predicted, and further the effect of accurately predicting the risk of financing is achieved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

FIG. 1 is a flow chart of a risk prediction method for financing provided according to an embodiment of the present application;

FIG. 2 is a schematic diagram of an entity classification model in an embodiment of the application;

FIG. 3 is a schematic diagram of a link prediction model in an embodiment of the application;

FIG. 4 is a schematic diagram of a risk prediction apparatus for financing provided according to an embodiment of the present application;

fig. 5 is a schematic diagram of an electronic device according to an embodiment of the present application.

Detailed Description

It should be noted that, without conflict, the embodiments of the present application and features of the embodiments may be combined with each other. The application will be described in detail below with reference to the drawings in connection with embodiments.

In order that those skilled in the art will better understand the present application, a technical solution in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present application without making any inventive effort, shall fall within the scope of the present application.

It should be noted that the terms "first," "second," and the like in the description and the claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate in order to describe the embodiments of the application herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or fully authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region, and provide corresponding operation entries for the user to select authorization or rejection.

The application will be described with reference to preferred implementation steps, and fig. 1 is a flowchart of a method for predicting risk of financing according to an embodiment of the present application, as shown in fig. 1, and the method includes the following steps:

step S101, responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution.

For example, the first object may be an enterprise, a person, an organization, a group, etc. that applies for financing to a financial institution. When an enterprise, a person, an organization, a group, or the like who applies for financing to a financial institution applies for financing to the financial institution, attribute information of the enterprise, the person, the organization, the group, or the like who applies for financing to the financial institution can be acquired. In addition, for example, when the first object for applying financing to a financial institution is an enterprise, the attribute information of the first object may include an enterprise legal person, an enterprise organization code, an enterprise asset, or the like; for example, when the first object applying for financing to a financial institution is an individual, the attribute information of the first object may include personal information, a job position of the business, and the like.

Step S102, inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk.

For example, attribute information of an enterprise, a person, an organization, a group, or the like, which applies for financing to a financial institution, may be input to an entity classification model (the above-described target classification model), and classification results of whether the enterprise, the person, the organization, the group, or the like, which applies for financing to the financial institution, is an entity at risk may be output. Moreover, the entity in this embodiment may be an entity such as an enterprise, a person, an organization, or a group.

Step S103, a knowledge graph is obtained, wherein the knowledge graph at least comprises information of a first object and object information associated with the first object.

For example, a constructed knowledge graph may be obtained, and attribute information of enterprises, individuals, institutions, groups, and the like that apply for financing to a financial institution, and attribute information of entities of enterprises, individuals, institutions, groups, and the like that apply for financing to a financial institution may be included in the previously constructed knowledge graph. Moreover, the entity in this embodiment may be an entity such as an enterprise, a person, an organization, or a group.

And step S104, determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks, wherein N is a positive integer.

For example, according to the classification result of whether the enterprises, individuals, institutions, groups and the like applying for financing to the financial institutions are the entities with risks and the association relationship among the entities in the pre-constructed knowledge graph, which are output by the entity classification model, a plurality of entities with risks can be determined, and according to the determined plurality of entities with risks, the risks when the enterprises, individuals, institutions, groups and the like applying for financing to the financial institutions apply for financing can be predicted.

It should be noted that, the risk prediction method for financing provided by the embodiment of the application can be applied to financial scenes.

Through the steps S101 to S104, the financing request sent by the first object is responded, the attribute information of the first object is obtained, the attribute information of the first object is input into the target classification model to perform classification processing, the classification result of whether the first object is an object with risk is output, the N objects with risk are determined according to the classification result and the knowledge graph, and the risk of the financing request is predicted based on the N objects with risk, so that when an enterprise applies for financing to a financial institution, other risk enterprises associated with the risk enterprises can be predicted, and the effect of accurately predicting the risk of financing is achieved.

Optionally, in the risk prediction method for financing provided by the embodiment of the present application, acquiring the knowledge graph includes: acquiring information of S second objects, wherein the information of the S second objects at least comprises information of a first object and object information related to the first object, the information of each second object at least comprises attribute information of each second object, S is a positive integer, and S is larger than M; determining the association relation between every two second objects according to the attribute information of each second object; and acquiring a knowledge graph based on the information of the S second objects and the association relation between every two second objects.

For example, constructing a knowledge graph can be divided into the following steps:

(1) An entity range is determined. And taking the collected entities (the S second objects) as the entities of the nodes in the knowledge graph. The entity mainly comprises: enterprises, individuals, institutions, communities, etc. have well-defined or socially-understood physical concepts.

(2) The entity attributes are noted. Specific attributes of the entity characteristics (the above-mentioned information of S second objects) are described. Such as businesses, including corporate legal, corporate organization codes, corporate assets, and the like; such as individuals, including personal information, the job position of the business, etc.

(3) A range of relationships is determined. A relationship definition is determined between the connection entity and the entity. Can be generally defined according to social learned relationships.

(4) The relationship attributes are annotated. The association relationship attribute is a specific description about the description relationship feature.

(5) An "entity-relationship-entity" triplet is established. The minimum unit body in the knowledge graph can be established according to the entity and the relation.

(6) And forming a knowledge graph based on the entity-relation-entity triples.

(7) And saving the formed knowledge graph in a graph database. The RDF storage system (a system for storing the knowledge graph) is generally used, operations such as adding, deleting, changing, checking and the like on the graph structure can be supported, and compared with the traditional database, the method has the advantages of high query speed, simplicity in operation and more abundant relation display modes.

By the scheme, the knowledge graph for predicting the financing risk can be quickly and accurately constructed.

Optionally, in the risk prediction method for financing provided by the embodiment of the present application, the target classification model is obtained by: acquiring a training sample set for model training, wherein the training sample set at least comprises T first attribute information samples and R second attribute information samples, the first attribute information samples are attribute information of objects with risks, the second attribute information samples are attribute information of objects without risks, and T and R are positive integers larger than 1; and carrying out learning training on the original classification model by adopting a training sample set to obtain a target classification model.

For example, attribute information samples of a plurality of objects with risks and attribute information samples of a plurality of objects without risks may be obtained, and the obtained attribute information samples of the plurality of objects with risks and the obtained attribute information samples of the plurality of objects without risks may be aggregated together to form the training sample set, and then an original entity classification model (the original classification model) may be learned and trained by using the constructed training sample set to obtain a trained entity classification model (the target classification model).

In summary, the training sample data is utilized to conveniently learn and train the original classification model, so that a trained classification model can be obtained.

Optionally, in the method for predicting risk of financing provided by the embodiment of the present application, after obtaining the knowledge graph, the method further includes: judging whether the association relation between every two second objects in the knowledge graph needs to be updated or not; if the association relation between every two second objects in the knowledge graph does not need to be updated, determining the object with risk according to the knowledge graph; if the association relation between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph; and determining the object with risk according to the updated knowledge graph.

For example, after the knowledge graph is built, if the relationship between the entities in the knowledge graph needs to be updated, the knowledge graph which is built originally needs to be updated, and the entities with risks are determined according to the updated knowledge graph. Otherwise, if the relation between the entities in the knowledge graph does not need to be updated, the entities with risks can be directly determined according to the updated knowledge graph.

By the scheme, whether the knowledge graph needs to be updated can be rapidly and accurately judged.

Optionally, in the risk prediction method for financing provided by the embodiment of the present application, if the association relationship between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain the updated knowledge graph includes: if the association relation between every two second objects in the knowledge graph needs to be updated, acquiring the second objects needing to update the association relation, and acquiring the association relation needing to be updated; inputting a second object needing to update the association relationship and the association relationship needing to be updated into a link prediction model, and outputting Q second objects, wherein Q is a positive integer, and Q is smaller than S; and updating the knowledge graph according to the second object needing to update the association relationship, the association relationship needing to be updated and the Q second objects to obtain an updated knowledge graph.

For example, if the relationship between the entities in the knowledge graph needs to be updated, the entity needing to be updated and the relationship needing to be updated may be input into the link prediction model, the entity needing to be updated in the knowledge graph is output, and then the entity needing to be updated in the knowledge graph and the relationship needing to be updated are output according to the model.

By the scheme, the knowledge graph can be updated rapidly and accurately.

Optionally, in the method for predicting risk of financing provided by the embodiment of the present application, determining, according to the classification result and the knowledge graph, N objects having risks includes: if the classification result indicates that the first object is an object with risk, determining M objects associated with the first object based on a knowledge graph, wherein M is a positive integer and is smaller than N; the first object and the M objects are taken as N objects with risks.

For example, if the enterprises, individuals, institutions, groups, and the like that apply for financing to the financial institutions output by the entity classification model are risk-bearing entities, the entities associated with the enterprises, individuals, institutions, groups, and the like that apply for financing to the financial institutions are determined from the pre-constructed knowledge graph, and the enterprises, individuals, institutions, groups, and the like that apply for financing to the financial institutions, and the entities associated with the enterprises, individuals, institutions, groups, and the like that apply for financing to the financial institutions are risk-bearing entities.

By the scheme, the object with financing risk can be rapidly and accurately determined.

Optionally, in the method for predicting risk of financing provided by the embodiment of the present application, predicting risk of a financing request based on N objects with risk includes: acquiring the number of N objects with risks; judging whether the number of N objects with risks is larger than a preset number or not; if the number of the N objects with risks is larger than the preset number, predicting that the risk level corresponding to the financing request is a first risk level; if the number of the N objects with the risks is not greater than the preset number, predicting that the risk level corresponding to the financing request is a second risk level, wherein the risk corresponding to the second risk level is lower than the risk corresponding to the first risk level.

For example, the preset number may be 5, and if it is determined that there are 8 entities with risk, and the number of the entities with risk is 8 or greater than the preset number of 5, it may be predicted that the risk of financing is greater (corresponding to the first risk level); if it is determined that there are 3 risky entities and the number of risky entities 3 is less than the preset number of entities 5, then it can be predicted that the risk of financing is less (corresponding to the second risk level described above).

By the scheme, when enterprises, individuals, institutions, communities and the like apply for financing to financial institutions, the risk of financing can be rapidly and accurately predicted.

For example, a financial institution may detect events and situations that may pose a liquidity risk based on business size, nature, complexity, and risk status of the enterprise, and focusing on associations between various enterprises. Moreover, with the wide application of artificial intelligence technology, by means of a machine learning algorithm, entity trend change propagation targeting a network is researched, so that decision-making personnel can make more comprehensive decision judgment.

In addition, knowledge-graph is a branch of artificial intelligence. The essence of the knowledge graph is a semantic knowledge base based on a graph structure, concepts, entities and relations among the concepts and the entities in the objective world are described in a structured form, the basic constituent units are triples, and the triples are formed by 'entities-relations-entities', and the entities are mutually connected through the relations to form a netlike knowledge base. In the embodiment, the entity trend change is combined with the knowledge graph, and the entity change propagation path is predicted by combining a machine learning technology from the perspective of communities.

The embodiment can solve the problem that when the financial institution examines the supplier financing application, the risk transmission is predicted so as to assist personnel in making decisions. Through knowledge graph technology, the following functions can be realized:

(1) For a large amount of data, knowledge maps of supplier cooperation relations, group relations, investment relations, asset and liability situation relations and the like can be established.

(2) And identifying the risk group and the risk client according to the knowledge graph.

(3) And automatically updating the knowledge graph.

In addition, the embodiment mainly comprises a knowledge graph construction module, a risk prediction module and an automatic updating module.

And the knowledge graph construction module constructs the bottom data of the knowledge graph, which is also the whole device and the basic module. The risk prediction module predicts the risk of other entities by blacklisting the entities (which may include entities that are at risk of financing), and is the core module of the overall apparatus. The automatic updating module is used for updating the data of the strong knowledge graph, and is an important module of the whole device.

1. Constructing a knowledge graph

(6) And forming a knowledge graph based on the entity-relation-entity triples.

2. Application of R-GCN in knowledge graph

The R-GCN (Relational Graph Convolutional Networks, relation graph rolling network) is based on the aggregation neighbor operation of the GCN (Graph Convolutional Network, graph rolling network), and the dimension of an aggregation relation is increased, so that the aggregation operation of the nodes is changed into a double aggregation process, each layer of node characteristics of the R-GCN are obtained from the node characteristics and the node relation of the last layer, the node characteristics of the node neighbors and the node characteristics are weighted and summed to obtain new characteristics, and the R-GCN considers self-loop in order to keep the information of the node. The R-GCN constructs an encoder and completes different modeling problems by accessing different layers, such as accessing a Softmax (mathematical function) layer for entity classification. The joining decoder performs link prediction.

The R-GCN defines a propagation model of a relational graph in which node vi (representing an entity) is updated as follows:

wherein,the updated characteristics of the own node; σ () is an activation function; />A neighbor node set with the relation r representing the node i; c (C) _i R is a regularized vector, wherein C _i The value of r is ∈>W _r ^(l) Is a linear transformation function, and uses the same parameter matrix W for neighboring nodes of the similar type edge _r ^(l) Performing transformation; w (W) _r ^(l) The weight matrix of the corresponding relation characteristic is considered, and because different relations are considered, how many kinds of relations have what Wr; />The input characteristics of the neighbor nodes are referred to; />Is the own characteristic weight matrix; />Refers to the characteristics of the input own node.

Moreover, as can be seen from the above formula (1), the R-GCN uses a double-layer traversal loop to traverse each relationship, superimpose the features of the neighboring points of each point for fusion, and finally add the central node feature of the previous layer, and output the feature as the output feature of the central node through an activation function. The R-GCN network has a set of parameters for each relationship, which is also the existence that the R-GCN can model multiple relationships.

(1) Realizing entity classification based on R-GCN+softmax

For example, FIG. 2 is a schematic diagram of an entity classification model in an embodiment of the present application, and Input in FIG. 2 represents Input, encoder represents encoder, and Node loss represents a loss function. And predicting the type label of the specific node, and using a softmax classifier for each node on the graph, wherein the node obtained by the R-GCN is used as the classifier input. The model and the R-GCN parameters are obtained by optimizing cross entropy loss learning.

Moreover, for semi-supervised classification of entities, the following cross entropy losses of the marker nodes are minimized by:

wherein Y is the set with the index of the tag node and +.>The network that is the ith marker node outputs the kth entry. t is t _ik Representing the respective authentic tag.

(2) Knowledge Graph Completion (KGC) is realized based on R-GCN+ConvE (2D convolution) link prediction

The goal of KGC is to predict valid but unobserved triples by basing them on the triples in a known knowledge graph. The KGC model is applied to each potential triplet (h _i ,r _j ,t _k ) Epsilon X R X epsilon define a scoreFunction s:the correlation score s (h _i ,r _j ,t _k ). The score measures the rationality of the triplet. For queries (h _i ,r _j What is? ) Or (? R, r _j ,t _k ) The model first fills in the blank with each entity in the knowledge graph and then scores the resulting triples. Valid triples are expected to score higher than invalid triples.

For example, fig. 3 is a schematic diagram of a link prediction model in an embodiment of the present application, where Input in fig. 3 represents an Input, an encoder represents an encoder, a decoder represents a decoder, edge loss represents an Edge loss function, and as shown in fig. 3, the link prediction model may be regarded as a self-encoder consisting of an encoder and a decoder, where the encoder generates a feature representation of a hidden layer for an entity by R-GCN, and the decoder is a decomposition model of tensor, and the label generated by the encoder is used to predict a predicted Edge. And the present embodiment uses ConvE (Convolutional Embeddings for Knowledge Graphs, convolution embedded model for knowledge-graph) as scoring model.

Moreover, convE uses convolutional neural networks to define a scoring function, which is:where σ represents the activation function, ω represents the convolution kernel in the convolution layer, +.>Representing the 2D real vector, W is a trainable weight matrix. The weight calculation formula is->Wherein d _ij Representing the Euclidean distance between two nodes, σ is a scale parameter variable in an n-dimensional space.

3. Risk prediction module

The present embodiment may label other entities that may be blacklisted by known blacklisted entities (which may be risky entities). And the main steps of the prediction are as follows:

the first step: the target variables of the model, namely the blacklist entities, are defined.

And a second step of: characteristic variables for modeling are designed. Feature variables are typically determined by entity attributes, and derivative variables are typically generated based on entity attribute processing.

And a third step of: modeling training, verification and testing are carried out through the entity classification model.

Fourth step: and obtaining the high-risk entity in the graph, and updating the blacklist entity list.

4. Automatic knowledge graph updating module

When the knowledge graph is constructed, along with the enrichment of future data, the constructed knowledge graph is more and more huge, and the relationship between entities is more and more sound. The knowledge graph is updated manually, which is time-consuming and labor-consuming, and the comprehensiveness of the relationship cannot be guaranteed. The main steps of automatic updating are as follows:

The first step: entity data to be updated is prepared, and the entity of the data to be updated is extracted through relation extraction.

And a third step of: modeling training, verification and testing are carried out through the link prediction model.

Fourth step: and updating the knowledge graph.

In summary, according to the risk prediction method for financing provided by the embodiment of the application, by responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution; inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; acquiring a knowledge graph, wherein the knowledge graph at least comprises information of a first object and object information associated with the first object; n objects with risks are determined according to the classification result and the knowledge graph, and the risks of financing requests are predicted based on the N objects with risks, wherein N is a positive integer, so that the problem that in the related art, when enterprises apply for financing to financial institutions, only whether a single enterprise is a risk enterprise can be identified, other risk enterprises associated with the risk enterprise are difficult to predict, and the risks of financing are difficult to accurately predict is solved. Responding to a financing request sent by a first object, acquiring attribute information of the first object, inputting the attribute information of the first object into a target classification model for classification processing, outputting a classification result of whether the first object is an object with risk or not, determining N objects with risk according to the classification result and a knowledge graph, and predicting the risk of the financing request based on the N objects with risk, so that when an enterprise applies for financing to a financial institution, other risk enterprises associated with the risk enterprises can be predicted, and further the effect of accurately predicting the risk of financing is achieved.

It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is illustrated in the flowcharts, in some cases the steps illustrated or described may be performed in an order other than that illustrated herein.

The embodiment of the application also provides a financing risk prediction device, and the financing risk prediction device can be used for executing the financing risk prediction method provided by the embodiment of the application. The risk prediction device for financing provided by the embodiment of the application is described below.

Fig. 4 is a schematic diagram of a risk prediction apparatus for financing according to an embodiment of the present application. As shown in fig. 4, the apparatus includes: a first processing unit 401, a first input unit 402, a first acquisition unit 403, and a second processing unit 404.

Specifically, the first processing unit 401 is configured to respond to a financing request sent by a first object, and obtain attribute information of the first object, where the financing request is a request that the first object applies for financing to a financial institution;

a first input unit 402, configured to input attribute information of a first object into a target classification model for classification processing, and output a classification result, where the classification result is used to indicate whether the first object is an object with risk;

A first obtaining unit 403, configured to obtain a knowledge graph, where the knowledge graph includes at least information of a first object and object information associated with the first object;

the second processing unit 404 is configured to determine N objects with risks according to the classification result and the knowledge graph, and predict the risk of the financing request based on the N objects with risks, where N is a positive integer.

In summary, in the risk prediction device for financing provided by the embodiment of the present application, a first processing unit 401 responds to a financing request sent by a first object, and acquires attribute information of the first object, where the financing request is a request that the first object applies for financing to a financial institution; the first input unit 402 inputs attribute information of the first object into the target classification model to perform classification processing, and outputs a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; the first obtaining unit 403 obtains a knowledge graph, where the knowledge graph includes at least information of a first object and object information associated with the first object; the second processing unit 404 is configured to determine N objects with risks according to the classification result and the knowledge graph, and predict the risk of the financing request based on the N objects with risks, where N is a positive integer, so as to solve the problem in the related art that when an enterprise applies for financing to a financial institution, only whether a single enterprise is a inauguration enterprise can be identified, and it is difficult to predict other inauguration enterprises associated with the inauguration enterprise, resulting in difficulty in accurately predicting the risk of financing. Responding to a financing request sent by a first object, acquiring attribute information of the first object, inputting the attribute information of the first object into a target classification model for classification processing, outputting a classification result of whether the first object is an object with risk or not, determining N objects with risk according to the classification result and a knowledge graph, and predicting the risk of the financing request based on the N objects with risk, so that when an enterprise applies for financing to a financial institution, other risk enterprises associated with the risk enterprises can be predicted, and further the effect of accurately predicting the risk of financing is achieved.

Optionally, in the risk prediction apparatus for financing provided by an embodiment of the present application, the second processing unit 404 includes: the first determining module is used for determining M objects associated with the first object based on the knowledge graph if the classification result indicates that the first object is an object with risk, wherein M is a positive integer and M is smaller than N; and the second determining module is used for taking the first object and the M objects as N objects with risks.

Optionally, in the risk prediction apparatus for financing provided in the embodiment of the present application, the first obtaining unit 403 includes: the first acquisition module is used for acquiring information of S second objects, wherein the information of the S second objects at least comprises information of a first object and object information related to the first object, the information of each second object at least comprises attribute information of each second object, S is a positive integer, and S is larger than M; the third determining module is used for determining the association relation between every two second objects according to the attribute information of each second object; and the second acquisition module is used for acquiring the knowledge graph based on the information of the S second objects and the association relation between every two second objects.

Optionally, in the risk prediction device for financing provided by the embodiment of the present application, the target classification model is obtained by: the second acquisition unit is used for acquiring a training sample set for model training, wherein the training sample set at least comprises T first attribute information samples and R second attribute information samples, the first attribute information samples are attribute information of objects with risks, the second attribute information samples are attribute information of objects without risks, and T and R are positive integers larger than 1; the first training unit is used for learning and training the original classification model by adopting a training sample set to obtain a target classification model.

Optionally, in the risk prediction device for financing provided by the embodiment of the present application, the device further includes: the first judging unit is used for judging whether the association relation between every two second objects in the knowledge graph needs to be updated after the knowledge graph is acquired; the first determining unit is used for determining an object with risk according to the knowledge graph if the association relation between every two second objects in the knowledge graph does not need to be updated; the first updating unit is used for updating the knowledge graph to obtain an updated knowledge graph if the association relationship between every two second objects in the knowledge graph needs to be updated; and the second determining unit is used for determining the object with risk according to the updated knowledge graph.

Optionally, in the risk prediction apparatus for financing provided in the embodiment of the present application, the first updating unit includes: the third acquisition module is used for acquiring the second objects needing to update the association relationship and acquiring the association relationship needing to be updated if the association relationship between every two second objects in the knowledge graph needs to be updated; the first input module is used for inputting the second object needing to update the association relation and the association relation needing to be updated into the link prediction model and outputting Q second objects, wherein Q is a positive integer and Q is smaller than S; the first updating module is used for updating the knowledge graph according to the second object needing to update the association relationship, the association relationship needing to be updated and the Q second objects to obtain an updated knowledge graph.

Optionally, in the risk prediction apparatus for financing provided by an embodiment of the present application, the second processing unit 404 includes: a fourth obtaining module, configured to obtain the number of N objects that have a risk; the first judging module is used for judging whether the number of N objects with risks is larger than a preset number or not; the first prediction module is used for predicting that the risk level corresponding to the financing request is a first risk level if the number of N objects with risks is greater than the preset number; and the second prediction module is used for predicting the risk level corresponding to the financing request as a second risk level if the number of the N objects with the risk is not greater than the preset number, wherein the risk corresponding to the second risk level is lower than the risk corresponding to the first risk level.

The risk prediction device for financing includes a processor and a memory, where the first processing unit 401, the first input unit 402, the first obtaining unit 403, the second processing unit 404, and the like are stored as program units, and the processor executes the program units stored in the memory to implement corresponding functions.

The processor includes a kernel, and the kernel fetches the corresponding program unit from the memory. The kernel can be provided with one or more than one, and the risk of financing can be accurately predicted by adjusting kernel parameters.

The memory may include volatile memory, random Access Memory (RAM), and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM), among other forms in computer readable media, the memory including at least one memory chip.

An embodiment of the present invention provides a computer-readable storage medium having stored thereon a program that, when executed by a processor, implements a risk prediction method for financing.

The embodiment of the invention provides a processor which is used for running a program, wherein the risk prediction method of financing is executed when the program runs.

As shown in fig. 5, an embodiment of the present invention provides an electronic device, where the device includes a processor, a memory, and a program stored in the memory and executable on the processor, and when the processor executes the program, the following steps are implemented: responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution; inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; acquiring a knowledge graph, wherein the knowledge graph at least comprises information of the first object and object information associated with the first object; and determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks, wherein N is a positive integer.

The processor also realizes the following steps when executing the program: determining the N objects with risks according to the classification result and the knowledge graph comprises the following steps: if the classification result indicates that the first object is an object with risk, determining M objects associated with the first object based on the knowledge graph, wherein M is a positive integer and is smaller than N; and taking the first object and the M objects as the N objects with risks.

The processor also realizes the following steps when executing the program: the obtaining of the knowledge graph comprises the following steps: acquiring information of S second objects, wherein the information of the S second objects at least comprises information of the first object and object information associated with the first object, the information of each second object at least comprises attribute information of each second object, S is a positive integer, and S is larger than M; determining the association relation between every two second objects according to the attribute information of each second object; and acquiring the knowledge graph based on the information of the S second objects and the association relation between every two second objects.

The processor also realizes the following steps when executing the program: the target classification model is obtained by the following steps: acquiring a training sample set for model training, wherein the training sample set at least comprises T first attribute information samples and R second attribute information samples, the first attribute information samples are attribute information of objects with risks, the second attribute information samples are attribute information of objects without risks, and T and R are positive integers larger than 1; and learning and training the original classification model by adopting the training sample set to obtain the target classification model.

The processor also realizes the following steps when executing the program: after obtaining the knowledge-graph, the method further comprises: judging whether the association relation between every two second objects in the knowledge graph needs to be updated or not; if the association relation between every two second objects in the knowledge graph does not need to be updated, determining the object with risk according to the knowledge graph; if the association relation between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph; and determining the object with risk according to the updated knowledge graph.

The processor also realizes the following steps when executing the program: if the association relationship between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph, wherein the updating of the knowledge graph comprises the following steps: if the association relation between every two second objects in the knowledge graph needs to be updated, acquiring the second objects needing to update the association relation, and acquiring the association relation needing to be updated; inputting the second object needing to update the association relationship and the association relationship needing to be updated into a link prediction model, and outputting Q second objects, wherein Q is a positive integer, and Q is smaller than S; and updating the knowledge graph according to the second object needing to update the association relationship, the association relationship needing to update and the Q second objects to obtain the updated knowledge graph.

The processor also realizes the following steps when executing the program: predicting the risk of the financing request based on the N risky objects comprises: acquiring the number of the N objects with risks; judging whether the number of the N objects with risks is larger than a preset number or not; if the number of the N objects with risks is larger than the preset number, predicting that the risk level corresponding to the financing request is a first risk level; and if the number of the N objects with risks is not greater than the preset number, predicting that the risk level corresponding to the financing request is a second risk level, wherein the risk corresponding to the second risk level is lower than the risk corresponding to the first risk level.

The device herein may be a server, PC, PAD, cell phone, etc.

The application also provides a computer program product adapted to perform, when executed on a data processing device, a program initialized with the method steps of: responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution; inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk; acquiring a knowledge graph, wherein the knowledge graph at least comprises information of the first object and object information associated with the first object; and determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks, wherein N is a positive integer.

When executed on a data processing device, is further adapted to carry out a program initialized with the method steps of: determining the N objects with risks according to the classification result and the knowledge graph comprises the following steps: if the classification result indicates that the first object is an object with risk, determining M objects associated with the first object based on the knowledge graph, wherein M is a positive integer and is smaller than N; and taking the first object and the M objects as the N objects with risks.

When executed on a data processing device, is further adapted to carry out a program initialized with the method steps of: the obtaining of the knowledge graph comprises the following steps: acquiring information of S second objects, wherein the information of the S second objects at least comprises information of the first object and object information associated with the first object, the information of each second object at least comprises attribute information of each second object, S is a positive integer, and S is larger than M; determining the association relation between every two second objects according to the attribute information of each second object; and acquiring the knowledge graph based on the information of the S second objects and the association relation between every two second objects.

When executed on a data processing device, is further adapted to carry out a program initialized with the method steps of: the target classification model is obtained by the following steps: acquiring a training sample set for model training, wherein the training sample set at least comprises T first attribute information samples and R second attribute information samples, the first attribute information samples are attribute information of objects with risks, the second attribute information samples are attribute information of objects without risks, and T and R are positive integers larger than 1; and learning and training the original classification model by adopting the training sample set to obtain the target classification model.

When executed on a data processing device, is further adapted to carry out a program initialized with the method steps of: after obtaining the knowledge-graph, the method further comprises: judging whether the association relation between every two second objects in the knowledge graph needs to be updated or not; if the association relation between every two second objects in the knowledge graph does not need to be updated, determining the object with risk according to the knowledge graph; if the association relation between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph; and determining the object with risk according to the updated knowledge graph.

When executed on a data processing device, is further adapted to carry out a program initialized with the method steps of: if the association relationship between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph, wherein the updating of the knowledge graph comprises the following steps: if the association relation between every two second objects in the knowledge graph needs to be updated, acquiring the second objects needing to update the association relation, and acquiring the association relation needing to be updated; inputting the second object needing to update the association relationship and the association relationship needing to be updated into a link prediction model, and outputting Q second objects, wherein Q is a positive integer, and Q is smaller than S; and updating the knowledge graph according to the second object needing to update the association relationship, the association relationship needing to update and the Q second objects to obtain the updated knowledge graph.

When executed on a data processing device, is further adapted to carry out a program initialized with the method steps of: predicting the risk of the financing request based on the N risky objects comprises: acquiring the number of the N objects with risks; judging whether the number of the N objects with risks is larger than a preset number or not; if the number of the N objects with risks is larger than the preset number, predicting that the risk level corresponding to the financing request is a first risk level; and if the number of the N objects with risks is not greater than the preset number, predicting that the risk level corresponding to the financing request is a second risk level, wherein the risk corresponding to the second risk level is lower than the risk corresponding to the first risk level.

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application 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 application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations 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.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises an element.

The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.

Claims

1. A method of risk prediction for financing, comprising:

responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution;

inputting attribute information of the first object into a target classification model for classification processing, and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk;

Acquiring a knowledge graph, wherein the knowledge graph at least comprises information of the first object and object information associated with the first object;

and determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks, wherein N is a positive integer.

2. The method of claim 1, wherein determining N objects at risk from the classification result and the knowledge-graph comprises:

if the classification result indicates that the first object is an object with risk, determining M objects associated with the first object based on the knowledge graph, wherein M is a positive integer and is smaller than N;

and taking the first object and the M objects as the N objects with risks.

3. The method of claim 1, wherein obtaining a knowledge-graph comprises:

acquiring information of S second objects, wherein the information of the S second objects at least comprises information of the first object and object information associated with the first object, the information of each second object at least comprises attribute information of each second object, S is a positive integer, and S is larger than M;

Determining the association relation between every two second objects according to the attribute information of each second object;

and acquiring the knowledge graph based on the information of the S second objects and the association relation between every two second objects.

4. The method of claim 1, wherein the object classification model is obtained by:

acquiring a training sample set for model training, wherein the training sample set at least comprises T first attribute information samples and R second attribute information samples, the first attribute information samples are attribute information of objects with risks, the second attribute information samples are attribute information of objects without risks, and T and R are positive integers larger than 1;

and learning and training the original classification model by adopting the training sample set to obtain the target classification model.

5. The method of claim 1, wherein after obtaining the knowledge-graph, the method further comprises:

judging whether the association relation between every two second objects in the knowledge graph needs to be updated or not;

if the association relation between every two second objects in the knowledge graph does not need to be updated, determining the object with risk according to the knowledge graph;

If the association relation between every two second objects in the knowledge graph needs to be updated, updating the knowledge graph to obtain an updated knowledge graph;

and determining the object with risk according to the updated knowledge graph.

6. The method of claim 5, wherein if the association relationship between every two second objects in the knowledge-graph needs to be updated, updating the knowledge-graph to obtain the updated knowledge-graph comprises:

if the association relation between every two second objects in the knowledge graph needs to be updated, acquiring the second objects needing to update the association relation, and acquiring the association relation needing to be updated;

inputting the second object needing to update the association relationship and the association relationship needing to be updated into a link prediction model, and outputting Q second objects, wherein Q is a positive integer, and Q is smaller than S;

and updating the knowledge graph according to the second object needing to update the association relationship, the association relationship needing to update and the Q second objects to obtain the updated knowledge graph.

7. The method of claim 1, wherein predicting the risk of the financing request based on the at-risk N subjects comprises:

Acquiring the number of the N objects with risks;

judging whether the number of the N objects with risks is larger than a preset number or not;

if the number of the N objects with risks is larger than the preset number, predicting that the risk level corresponding to the financing request is a first risk level;

and if the number of the N objects with risks is not greater than the preset number, predicting that the risk level corresponding to the financing request is a second risk level, wherein the risk corresponding to the second risk level is lower than the risk corresponding to the first risk level.

8. A risk prediction apparatus for financing, comprising:

the first processing unit is used for responding to a financing request sent by a first object and acquiring attribute information of the first object, wherein the financing request is a request for the first object to apply for financing to a financial institution;

the first input unit is used for inputting attribute information of the first object into a target classification model to carry out classification processing and outputting a classification result, wherein the classification result is used for indicating whether the first object is an object with risk;

a first obtaining unit, configured to obtain a knowledge graph, where the knowledge graph at least includes information of the first object and object information associated with the first object;

And the second processing unit is used for determining N objects with risks according to the classification result and the knowledge graph, and predicting the risk of the financing request based on the N objects with risks, wherein N is a positive integer.

9. A computer-readable storage medium storing a program, wherein the program performs the risk prediction method of financing according to any one of claims 1 to 7.

10. An electronic device comprising one or more processors and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the risk prediction method of financing of any of claims 1-7.