CN113011961A - Method, device and equipment for monitoring risk of company associated information and storage medium - Google Patents
Method, device and equipment for monitoring risk of company associated information and storage medium Download PDFInfo
- Publication number
- CN113011961A CN113011961A CN202110133551.5A CN202110133551A CN113011961A CN 113011961 A CN113011961 A CN 113011961A CN 202110133551 A CN202110133551 A CN 202110133551A CN 113011961 A CN113011961 A CN 113011961A
- Authority
- CN
- China
- Prior art keywords
- information
- company
- characteristic information
- feature
- risk
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000012544 monitoring process Methods 0.000 title claims abstract description 54
- 239000013598 vector Substances 0.000 claims abstract description 95
- 238000013135 deep learning Methods 0.000 claims abstract description 77
- 238000012549 training Methods 0.000 claims abstract description 61
- 238000012806 monitoring device Methods 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 15
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 238000013507 mapping Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 abstract description 15
- 238000012502 risk assessment Methods 0.000 description 11
- 238000004891 communication Methods 0.000 description 6
- 238000011835 investigation Methods 0.000 description 5
- 238000007670 refining Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000011144 upstream manufacturing Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Human Resources & Organizations (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Technology Law (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Animal Behavior & Ethology (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method, a device, equipment and a storage medium for monitoring risk of company associated information, which are used for acquiring customer characteristic information of a company and associated characteristic information of a company associated party; extracting feature embedding vectors based on the customer feature information and the associated feature information; training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model; and acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result. When the trained target model is used for risk prediction, the method predicts the characteristic information to be evaluated through the multi-head attention mechanism of the target model, and effectively improves the accuracy of the evaluation result when company risk evaluation is carried out.
Description
Technical Field
The invention relates to the technical field of risk monitoring, in particular to a method, a device, equipment and a storage medium for risk monitoring of company associated information.
Background
At present, when company risk assessment is carried out, risk signals can be manually refined according to expert rules or business experiences to carry out risk detection, and the refining mode is simple index statistics. Or a knowledge graph is constructed according to credit data of only the company, characteristics of learning of the knowledge graph of the company are embedded by utilizing deep learning of the graph, and the characteristics are embedded to carry out risk assessment through a risk assessment model.
However, whether the risk assessment is performed by manually refining the risk signal or the risk assessment is performed by the credit data of the company, the accuracy of the assessment result is low.
Disclosure of Invention
The invention mainly aims to provide a method, a device, equipment and a storage medium for monitoring company associated information risk, and aims to solve the technical problem that the accuracy of an evaluation result is low when company risk evaluation is carried out at present.
In order to achieve the above object, an embodiment of the present invention provides a method for monitoring risk of company-associated information, where the method for monitoring risk of company-associated information includes:
acquiring customer characteristic information of a company and associated characteristic information of a company associated party;
extracting feature embedding vectors based on the customer feature information and the associated feature information;
training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model;
and acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the company to be evaluated through the target model and the characteristic information to be evaluated to obtain a prediction result.
Preferably, the step of training the multi-head attention module and the feature embedding network of the preset deep learning network according to the feature embedding vector to obtain the target model includes:
processing the feature embedded vector through a multi-head attention module of a preset deep learning network to obtain a hidden vector;
mapping the hidden vector through a linear two-classifier of the preset deep learning network to obtain a classification result;
and training the feature embedded network and the multi-head attention module of the preset deep learning network according to the classification result to obtain a target model.
Preferably, the step of processing the feature embedding vector by a multi-head attention module of a preset deep learning network to obtain a hidden vector includes:
acquiring an attention weight parameter of a multi-head attention module in a preset deep learning network;
and performing preset operation on the feature embedding vector according to the attention weight parameter to obtain a hidden layer vector.
Preferably, the step of extracting a feature embedding vector based on the customer feature information and the associated feature information includes:
inputting the customer characteristic information and the associated characteristic information into a characteristic embedding network of a preset deep learning network;
and performing data conversion on the client characteristic information and the associated characteristic information through the characteristic embedding network to obtain a characteristic embedding vector.
Preferably, the step of acquiring the customer characteristic information of the company and the associated characteristic information of the company-associated party includes:
acquiring historical credit default information of a company, creating a training label according to the historical credit default information, acquiring label characteristic information of the company, and forming customer characteristic information by the training label and the label characteristic information;
and acquiring the associated map information of the company, and acquiring the associated feature information of the company associated party according to the associated map information.
Preferably, the step of obtaining the associated feature information of the company associated party according to the associated map information includes:
identifying a related party of the company according to the related map information;
and acquiring the associated characteristic information of the associated party.
Preferably, after the step of obtaining the feature information to be evaluated of the company to be evaluated, the method further includes:
and calculating the contribution degree of the characteristic information to be evaluated through a model explanatory module to generate associated risk monitoring information.
In order to achieve the above object, the present invention further provides a risk monitoring device for company-related information, including:
the acquisition module is used for acquiring the client characteristic information of a company and the associated characteristic information of the company associated party;
an extraction module for extracting feature embedding vectors based on the customer feature information and the associated feature information;
the training module is used for training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model;
and the prediction module is used for acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result.
Further, in order to achieve the above object, the present invention further provides a risk monitoring device for company-related information, which includes a memory, a processor, and a risk monitoring program for company-related information stored in the memory and operable on the processor, wherein the risk monitoring program for company-related information realizes the steps of the risk monitoring method for company-related information when executed by the processor.
Further, in order to achieve the above object, the present invention further provides a storage medium, wherein the storage medium stores a risk monitoring program for company-related information, and the risk monitoring program for company-related information realizes the steps of the risk monitoring method for company-related information when being executed by a processor.
The embodiment of the invention provides a method, a device, equipment and a storage medium for monitoring risks of company associated information, which are used for acquiring client characteristic information of a company and associated characteristic information of a company associated party; extracting feature embedding vectors based on the customer feature information and the associated feature information; training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model; and acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result. According to the method, the characteristic embedded vector is extracted from the client characteristic information of a company and the associated characteristic information of a company associated party to be used as a training basis, and the multi-head attention module and the characteristic embedded network of the preset deep learning network are trained, so that when a target model obtained by training is applied to risk prediction, the characteristic information to be evaluated is predicted through the multi-head attention mechanism of the target model, and the accuracy of an evaluation result in company risk evaluation is effectively improved.
Drawings
FIG. 1 is a schematic structural diagram of a hardware operating environment according to an embodiment of a risk monitoring method for company-associated information of the present invention;
FIG. 2 is a schematic flow chart illustrating a risk monitoring method for company-associated information according to a first embodiment of the present invention;
FIG. 3 is a schematic flow chart illustrating a risk monitoring method for company-associated information according to a second embodiment of the present invention;
FIG. 4 is a schematic flow chart illustrating a risk monitoring method for company-associated information according to a third embodiment of the present invention;
FIG. 5 is a schematic flow chart of a risk monitoring method for company-associated information according to a fourth embodiment of the present invention;
FIG. 6 is a functional block diagram of a risk monitoring device according to a preferred embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The embodiment of the invention provides a method, a device, equipment and a storage medium for monitoring risks of company associated information, which are used for acquiring client characteristic information of a company and associated characteristic information of a company associated party; extracting feature embedding vectors based on the customer feature information and the associated feature information; training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model; and acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result. According to the method, the characteristic embedded vector is extracted from the client characteristic information of a company and the associated characteristic information of a company associated party to be used as a training basis, and the multi-head attention module and the characteristic embedded network of the preset deep learning network are trained, so that when a target model obtained by training is applied to risk prediction, the characteristic information to be evaluated is predicted through the multi-head attention mechanism of the target model, and the accuracy of an evaluation result in company risk evaluation is effectively improved.
As shown in fig. 1, fig. 1 is a schematic structural diagram of a risk monitoring device for company-associated information of a hardware operating environment according to an embodiment of the present invention.
In the following description, suffixes such as "module", "component", or "unit" used to denote elements are used only for facilitating the explanation of the present invention, and have no specific meaning in itself. Thus, "module", "component" or "unit" may be used mixedly.
The risk monitoring device for the company associated information in the embodiment of the invention can be a PC, and can also be a mobile terminal device such as a tablet computer and a portable computer.
As shown in fig. 1, the risk monitoring device for company-associated information may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a storage device separate from the processor 1001.
Those skilled in the art will appreciate that the corporate-associated information risk monitoring device configuration shown in fig. 1 does not constitute a limitation of corporate-associated information risk monitoring devices, and may include more or fewer components than shown, or some components in combination, or a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a kind of storage medium, may include therein an operating system, a network communication module, a user interface module, and a company-related information risk monitoring program.
In the device shown in fig. 1, the network interface 1004 is mainly used for connecting to a backend server and performing data communication with the backend server; the user interface 1003 is mainly used for connecting a client (user side) and performing data communication with the client; and the processor 1001 may be configured to invoke the corporate linkage information risk monitoring program stored in the memory 1005 and perform the following operations:
acquiring customer characteristic information of a company and associated characteristic information of a company associated party;
extracting feature embedding vectors based on the customer feature information and the associated feature information;
training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model;
and acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the company to be evaluated through the target model and the characteristic information to be evaluated to obtain a prediction result.
Further, the step of training the multi-head attention module and the feature embedding network of the preset deep learning network according to the feature embedding vector to obtain the target model comprises:
processing the feature embedded vector through a multi-head attention module of a preset deep learning network to obtain a hidden vector;
mapping the hidden vector through a linear two-classifier of the preset deep learning network to obtain a classification result;
and training the feature embedded network and the multi-head attention module of the preset deep learning network according to the classification result to obtain a target model.
Further, the step of processing the feature embedding vector through a multi-head attention module of a preset deep learning network to obtain a hidden vector includes:
acquiring an attention weight parameter of a multi-head attention module in a preset deep learning network;
and performing preset operation on the feature embedding vector according to the attention weight parameter to obtain a hidden layer vector.
Further, the step of extracting feature embedding vectors based on the customer feature information and the associated feature information includes:
inputting the customer characteristic information and the associated characteristic information into a characteristic embedding network of a preset deep learning network;
and performing data conversion on the client characteristic information and the associated characteristic information through the characteristic embedding network to obtain a characteristic embedding vector.
Further, the step of obtaining the customer characteristic information of the company and the associated characteristic information of the company-associated party includes:
acquiring historical credit default information of a company, creating a training label according to the historical credit default information, acquiring label characteristic information of the company, and forming customer characteristic information by the training label and the label characteristic information;
and acquiring the associated map information of the company, and acquiring the associated feature information of the company associated party according to the associated map information.
Further, the step of obtaining the associated feature information of the company associated party according to the associated map information includes:
identifying a related party of the company according to the related map information;
and acquiring the associated characteristic information of the associated party.
Further, after the step of obtaining the characteristic information to be evaluated of the company to be evaluated, the processor 1001 may be configured to call the company-related information risk monitoring program stored in the memory 1005, and perform the following operations:
and calculating the contribution degree of the characteristic information to be evaluated through a model explanatory module to generate associated risk monitoring information.
For a better understanding of the above technical solutions, exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
In order to better understand the technical solution, the technical solution will be described in detail with reference to the drawings and the specific embodiments.
Referring to fig. 2, a first embodiment of the present invention provides a schematic flow chart of a risk monitoring method for company-related information. In this embodiment, the risk monitoring method for company-associated information includes the following steps:
step S10, obtaining the client characteristic information of the company and the related characteristic information of the company related party;
the risk monitoring method for the company associated information in the embodiment is applied to a risk monitoring system for the company associated information, and it can be understood that for convenience of description, the risk monitoring system for the company associated information is simply referred to as a system below; the system comprises a preset deep learning network and a model explanatory module, wherein the preset deep learning network is preferably a deep learning network, the deep learning network consists of a feature embedded network, a linear two-classifier and a multi-head attention module, and the feature embedded network is used for converting data into features with fixed sizes for representation; a linear two classifier in machine learning for using a quadric surface to classify an object or event into two or more classifications; the multi-head attention module is used for utilizing a plurality of queries to calculate and select a plurality of information from the input information in parallel; the model interpretation module is used to calculate how a plurality of input variables in the model work and how the prediction of the model changes according to the values of the input variables.
It can be understood that, at present, when company risk assessment is performed, risk detection can be performed by manually refining the risk signal according to expert rules or business experiences, and the refining mode is simple index statistics. Or a knowledge graph is constructed according to credit data of only the company, characteristics of learning of the knowledge graph of the company are embedded by utilizing deep learning of the graph, and the characteristics are embedded to carry out risk assessment through a risk assessment model. However, whether the risk assessment is performed by manually refining the risk signal or the risk assessment is performed by the credit data of the company, the accuracy of the assessment result is low. On the basis, the method for monitoring the risk of the company associated information extracts the characteristic embedded vector as a training basis from the client characteristic information of a company and the associated characteristic information of a company associated party, trains the multi-head attention module and the characteristic embedded network of the preset deep learning network, predicts the characteristic information to be evaluated through the multi-head attention mechanism of the target model when the risk prediction is carried out on the target model obtained by applying the training, and effectively improves the accuracy of the evaluation result when the risk evaluation of the company is carried out.
Further, the system acquires historical credit default information of at least one company, creates a training label through the historical credit default information, and simultaneously acquires label characteristic information of the company, wherein the label characteristic information refers to various kinds of characteristic information corresponding to the company with the training label. In order to more accurately perform risk assessment, the system needs to acquire the characteristic information of other companies or enterprises related to the current company, and if the historical credit default information of a plurality of companies is acquired, the characteristic information of the related company or enterprise of each company needs to be acquired respectively to obtain the related characteristic information of the company related party.
Step S20, extracting a feature embedding vector based on the customer feature information and the associated feature information;
further, in order to more conveniently process the client characteristic information and the associated characteristic information, the system needs to extract the characteristic information from the client characteristic information and the associated characteristic information, specifically, the system inputs the client characteristic information and the associated characteristic information into a characteristic embedding network of a preset deep learning network, performs data conversion on the input client characteristic information and the associated characteristic information through the characteristic embedding network, and obtains a characteristic embedding vector after the data conversion is completed.
Further, the step of extracting feature embedding vectors based on the customer feature information and the associated feature information includes:
step S21, inputting the customer characteristic information and the associated characteristic information into a characteristic embedding network of a preset deep learning network;
and step S22, performing data conversion on the client characteristic information and the associated characteristic information through the characteristic embedding network to obtain a characteristic embedding vector.
Further, in order to facilitate training of the preset deep learning network through the client characteristic information and the associated characteristic information, data format conversion needs to be performed on the client characteristic information and the associated characteristic information, specifically, the system inputs the client characteristic information and the associated characteristic information into a characteristic embedded network of the preset deep learning network, and performs data conversion on the client characteristic information and the associated characteristic information through the characteristic embedded network, specifically, since data dimensions of the client characteristic information and the associated characteristic information may be different from data dimensions that can be identified by the preset deep learning network, dimension information capable of identifying data is set in the characteristic embedded network of the preset deep learning network, after the client characteristic information and the associated characteristic information are received by the preset deep learning network, the data dimensions of the client characteristic information and the associated characteristic information are identified through the characteristic embedded network, performing dimension increasing or dimension reducing processing on the client characteristic information and the associated characteristic information to enable the data dimension of the client characteristic information and the associated characteristic information after the dimension reducing processing to be the same as the dimension information of identifiable data in the characteristic embedding network, and converting to obtain a characteristic embedding vector; the system calculates the characteristic embedding vectors of multiple types of associators of the company and finds out the node set of each type of associator of the company by circularly executing the process of acquiring the characteristic embedding vectors. It can be understood that when the customer feature information and the associated feature information are input into the feature embedded network of the preset deep learning network, the information such as the abnormal graph composed of the node convolution kernel and the edge set, the attention head number, the attention parameter, the linear classifier parameter, the maximum iteration parameter, and the like is also input. The characteristic embedding vectors of the client characteristic information and the associated characteristic information are extracted through the preset deep learning network, the characteristic embedding vectors are prevented from being extracted manually, the data processing efficiency is improved, errors generated during manual extraction can be reduced, and the accuracy of data extraction is improved.
Step S30, training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model;
further, the system firstly carries out operation on the feature embedded vector through a multi-head attention module of a preset deep learning network to calculate a hidden layer vector; mapping the hidden vector through a linear two-classifier in a preset deep learning network to obtain a classification result consisting of a plurality of classifications; and finally, training the multi-head attention module and the feature embedded network of the preset deep learning network according to the generated classification result to obtain a target model.
And step S40, obtaining the characteristic information to be evaluated of the company to be evaluated, and carrying out risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result.
Further, when a company needs to perform risk business handling under the system, the system can send a business handling request to the system, and when the system receives the business handling request sent by the company, the company needing to perform risk business handling is firstly determined as a company to be evaluated, then characteristic information to be evaluated of the company to be evaluated is obtained, then the risk information to be evaluated of the company to be evaluated is input into a trained target model, the characteristic information to be evaluated is subjected to risk prediction through the target model, and finally a prediction result of the risk prediction of the company is obtained, wherein the prediction result is a default probability between 0 and 1, the system can output the prediction result through a display device, so that a worker can make a decision on the business handling request of the company according to the default probability in the prediction result, specifically, whether the business handling (such as loan) request of the company to be evaluated is made according to the default probability, if the default probability is higher than the preset default probability set by the financial institution (such as a bank), the loan is made to the company to be assessed, the company to be assessed has a high probability of not paying, and if the loan transaction request of the company to be assessed is passed, the bank benefit is damaged, which is not beneficial to the development of the bank.
The embodiment provides a method, a device, equipment and a storage medium for monitoring risk of company associated information, which are used for acquiring customer characteristic information of a company and associated characteristic information of a company associated party; extracting feature embedding vectors based on the customer feature information and the associated feature information; training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model; and acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result. According to the method, the characteristic embedded vector is extracted from the client characteristic information of a company and the associated characteristic information of a company associated party to be used as a training basis, and the multi-head attention module and the characteristic embedded network of the preset deep learning network are trained, so that when a target model obtained by training is applied to risk prediction, the characteristic information to be evaluated is predicted through the multi-head attention mechanism of the target model, and the accuracy of an evaluation result in company risk evaluation is effectively improved.
Further, referring to fig. 3, a second embodiment of the method for monitoring risk of company-associated information according to the present invention is provided based on the first embodiment of the method for monitoring risk of company-associated information according to the present invention, in the second embodiment, the step of training the multi-head attention module and the feature-embedded network of the preset deep learning network according to the feature-embedded vector to obtain the target model includes:
step S31, processing the feature embedding vector through a multi-head attention module of a preset deep learning network to obtain a hidden vector;
step S32, mapping the hidden vector through a linear two-classifier of the preset deep learning network to obtain a classification result;
and step S33, training the feature embedded network of the preset deep learning network and the multi-head attention module according to the classification result to obtain a target model.
Further, the system inputs the feature embedding vector into a preset deep learning network, the input feature embedding vector is processed through a multi-head attention module in the preset deep learning network, specifically, the multi-head attention module performs weighted average operation on the input feature embedding vector, and a hidden vector is obtained after the operation is completed. Furthermore, the system maps the calculated hidden layer vector by a linear two-classifier of a preset deep learning network, maps the hidden layer vector into a plurality of classes by the linear two-classifier, and forms a mapping classification result by the plurality of classes. Further, the system trains the feature embedded network and the multi-head attention module of the preset deep learning network according to the classification result obtained by mapping, specifically trains parameters in the feature embedded network and the multi-head attention module, judges that the training is finished when the parameters are trained to be convergent, ends the cycle process, and obtains a final target model. After a final target model is obtained through training, only characteristic information to be evaluated of a company to be evaluated needs to be input into the target model, risk prediction can be conducted on the company to be evaluated through the target model according to the characteristic information to be evaluated, a prediction result is generated, and accuracy of the prediction result obtained through the training of the target model through training is improved through training of the characteristic embedding network of the preset deep learning network and the multi-head attention module.
Further, the step of processing the feature embedding vector through a multi-head attention module of a preset deep learning network to obtain a hidden vector includes:
step S311, acquiring attention weight parameters of a multi-head attention module in a preset deep learning network;
step S312, performing a preset operation on the feature embedding vector according to the attention weight parameter to obtain a hidden vector.
It can be understood that, in the multi-head attention module of the preset deep learning network, the attention weight parameter is assigned by the system to represent the weight of each head in the multi-head attention module in the whole module, and the weight of each head is not fixed, but is assigned by the system after being updated according to a certain update rule. Therefore, the system obtains an attention weight parameter composed of the weight of each head from a multi-head attention module of a preset deep learning network, performs preset weighted average operation on the feature embedded vector according to the attention weight parameter, specifically performs product operation on the feature embedded vector and the weight of each head respectively to obtain a plurality of product results, and concatenates the product results through addition operation to form a hidden layer vector of the feature embedded vector, so that the feature embedded network of the preset deep learning network and the multi-head attention module are trained after a classification result is mapped through the hidden layer vector, and the accuracy of a predicted result of a target model obtained through training is improved.
In the embodiment, the multi-head attention module and the linear two-classifier are used for processing the feature embedded vector in sequence, and then the multi-head attention module and the feature embedded network of the preset deep learning network are trained according to the classification result obtained by processing, so that the accuracy of the prediction result obtained by predicting the target model through the training is higher.
Further, referring to fig. 4, a third embodiment of the risk monitoring method for company-related information according to the present invention is proposed based on the first embodiment of the risk monitoring method for company-related information according to the present invention, and in the third embodiment, the step of obtaining the customer characteristic information of a company and the related characteristic information of the company-related party includes:
step S11, obtaining historical credit default information of a company, creating a training label according to the historical credit default information, obtaining label characteristic information of the company, and forming customer characteristic information by the training label and the label characteristic information;
and step S12, acquiring the association map information of the company, and acquiring the association feature information of the company association party according to the association map information.
Furthermore, when a company transacts business such as loan to a financial institution such as a bank, the company needs to provide characteristic information such as credit investigation, industry and commerce, law, transaction, credit, finance, enterprise owner, early warning, risk and the like to the bank, or an authorized financial institution can directly obtain the characteristic information such as credit investigation, industry and commerce, law, transaction, credit, finance, enterprise owner, early warning, risk and the like of the company through the system, the system can obtain the characteristic information of whether the company has an influence on the default such as credit investigation, industry and commerce, law, transaction, credit, finance, enterprise owner, early warning, risk and the like of the company from a risk information database of the financial institution or a risk information storage unit of the company as label characteristic information, and can also obtain historical credit default information of the company from a risk information database of the financial institution or a risk information storage unit of the company, after obtaining the historical credit default information, the system creates a training label according to the historical credit default condition, for example, if the default condition exists, an identifier 1 is set, and if the default condition does not exist, an identifier 0 is set. And obtaining label characteristic information and a training label, and forming customer characteristic information by the training label and the label characteristic information together, so as to input the customer characteristic information into a preset deep learning network in the subsequent process, and training to obtain an association relation between the training label and the label characteristic information.
It can be understood that, when the evaluation is performed on the financial-related business transaction, the evaluation is not accurate only based on the characteristic information of the current company, and if the company has an associated company such as a subsidiary company or a parent company, the characteristic information of the associated company may also affect the credit capability of the current company. Therefore, the system needs to acquire the association map information of the current company, where the association map information is used to display the relationship between the current company and other companies having an association relationship with the current company, and after acquiring the association map information, the system acquires the association party of the current company through the association map information and acquires the association feature information of the association party.
Further, the step of obtaining the associated feature information of the company associated party according to the associated map information includes:
step S121, identifying the related party of the company according to the related map information;
and step S122, acquiring the associated characteristic information of the associated party.
Further, after obtaining the associated map information, the system may identify other companies associated with the current company from the associated map information, use the other companies associated with the current company as associated parties, obtain characteristic information of security, upstream and downstream, same entity controller, same legal person representative, litigation, transaction, external investment, share right, etc. of the associated parties, obtain characteristic information of credit investigation, industry quotient, law, transaction, credit, finance, business owner, early warning, risk, etc. of the associated parties, and combine the characteristic information of security, upstream and downstream, same entity controller, same legal person representative, litigation, transaction, external investment, share right, etc. and the characteristic information of credit investigation, industry quotient, law, transaction, credit, finance, business owner, early warning, risk, etc. to form associated characteristic information of the associated parties.
In the embodiment, the client characteristic information of the company and the associated characteristic information of the company associated party are obtained, and the characteristic embedding vector obtained by performing characteristic embedding extraction on the risk information of the company and the associated party through a preset deep learning network is more accurate through the client characteristic information generated by the training label and the associated characteristic information of the associated party. Furthermore, the model interpretability module gives risk influence factors with interpretability, and the persuasion of company associated risk monitoring conclusion can be improved.
Further, referring to fig. 5, a fourth embodiment of the risk monitoring method for company-related information according to the present invention is proposed based on the first embodiment of the risk monitoring method for company-related information according to the present invention, and in the fourth embodiment, after the step of obtaining the feature information to be evaluated of the company to be evaluated, the method further includes:
and S100, calculating the contribution degree of the characteristic information to be evaluated through a model explanatory module to generate associated risk monitoring information.
Furthermore, a model explanatory module is arranged in the system, and the model explanatory module can calculate how a plurality of input variables in the model work and how the prediction of the model changes according to the values of the input variables. Therefore, after the system trains the multi-head attention module and the feature embedded network of the preset deep learning network through the feature embedded vector to obtain the target model, and when risk prediction is carried out on the company to be evaluated through the target model and the acquired characteristic information to be evaluated, calculating the contribution degree of the characteristic information to be evaluated input into the target model through a model explanatory module, determining the contribution degree of each characteristic in the characteristic information to be evaluated in risk evaluation, therefore, model interpretability is made on the characteristic information to be evaluated, the associated risk monitoring information of the company to be evaluated is generated after calculation is completed, and the associated risk monitoring information can be output through a display screen, specifically, the associated risk monitoring information can be output together when the prediction result is output, so that the purpose of comprehensively monitoring the risk of the company to be evaluated is achieved.
In the embodiment, the contribution degree of the characteristic information to be evaluated is calculated through the model explanatory module, the associated risk monitoring information is generated, and the comprehensiveness of enterprise risk monitoring is effectively improved.
Further, the invention also provides a risk monitoring device for the company associated information.
Referring to fig. 6, fig. 6 is a functional module schematic diagram of a first embodiment of the risk monitoring device for company-related information according to the present invention.
The company associated information risk monitoring device comprises:
an obtaining module 10, configured to obtain customer feature information of a company and associated feature information of a company associated party;
an extraction module 20, configured to extract a feature embedding vector based on the customer feature information and the associated feature information;
the training module 30 is used for training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model;
and the prediction module 40 is used for acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result.
Further, the obtaining module 10 includes:
the system comprises a first acquisition unit, a second acquisition unit and a third acquisition unit, wherein the first acquisition unit is used for acquiring historical credit default information of a company, creating a training label according to the historical credit default information, acquiring label characteristic information of the company, and forming customer characteristic information by the training label and the label characteristic information;
and the second acquisition unit is used for acquiring the associated map information of the company and acquiring the associated characteristic information of the company associated party according to the associated map information.
Further, the obtaining module 10 further includes:
the identification unit is used for identifying the related party of the company according to the related map information;
and the third acquisition unit is used for acquiring the associated characteristic information of the associated party.
Further, the extraction module 20 includes:
the input unit is used for inputting the client characteristic information and the associated characteristic information into a characteristic embedded network of a preset deep learning network;
and the conversion unit is used for performing data conversion on the client characteristic information and the associated characteristic information through the characteristic embedded network to obtain a characteristic embedded vector.
Further, the training module 30 includes:
the data processing unit is used for processing the feature embedded vector through a multi-head attention module of a preset deep learning network to obtain a hidden vector;
the mapping unit is used for mapping the hidden layer vector through a linear two-classifier of the preset deep learning network to obtain a classification result;
and the training unit is used for training the feature embedded network of the preset deep learning network and the multi-head attention module according to the classification result to obtain a target model.
Further, the training module 30 further includes:
the fourth acquisition unit is used for acquiring the attention weight parameter of the multi-head attention module in the preset deep learning network;
and the operation unit is used for carrying out preset operation on the feature embedding vector according to the attention weight parameter to obtain a hidden vector.
Further, the prediction module 40 includes:
and the calculation unit is used for calculating the contribution degree of the characteristic information to be evaluated through the model explanatory module to generate the associated risk monitoring information.
Furthermore, the present invention also provides a storage medium, preferably a computer-readable storage medium, on which a risk monitoring program of company-related information is stored, which when executed by a processor implements the steps of the embodiments of the risk monitoring method of company-related information described above.
In the embodiments of the risk monitoring device and the computer-readable medium for company-associated information according to the present invention, all technical features of the embodiments of the risk monitoring method for company-associated information are included, and the description and explanation contents are basically the same as those of the embodiments of the risk monitoring method for company-associated information, and are not repeated herein.
It should be noted that, in this document, 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 an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention or a part contributing to the prior art may be embodied in the form of a software product, where the computer software product is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk), and includes a plurality of instructions for enabling a terminal device (which may be a fixed terminal, such as an internet of things smart device including smart homes, such as a smart air conditioner, a smart lamp, a smart power supply, a smart router, etc., or a mobile terminal, including a smart phone, a wearable networked AR/VR device, a smart sound box, an autonomous driving automobile, etc.) to execute the method according to each embodiment of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A risk monitoring method for company associated information is characterized by comprising the following steps:
acquiring customer characteristic information of a company and associated characteristic information of a company associated party;
extracting feature embedding vectors based on the customer feature information and the associated feature information;
training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model;
and acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the company to be evaluated through the target model and the characteristic information to be evaluated to obtain a prediction result.
2. The method for monitoring risk of company-associated information according to claim 1, wherein the step of training a multi-head attention module and a feature-embedded network of a preset deep learning network according to the feature-embedded vector to obtain a target model comprises:
processing the feature embedded vector through a multi-head attention module of a preset deep learning network to obtain a hidden vector;
mapping the hidden vector through a linear two-classifier of the preset deep learning network to obtain a classification result;
and training the feature embedded network and the multi-head attention module of the preset deep learning network according to the classification result to obtain a target model.
3. The method for monitoring risk of company-associated information according to claim 2, wherein the step of processing the feature-embedded vector by a multi-head attention module of a preset deep learning network to obtain a hidden vector comprises:
acquiring an attention weight parameter of a multi-head attention module in a preset deep learning network;
and performing preset operation on the feature embedding vector according to the attention weight parameter to obtain a hidden layer vector.
4. The corporate linkage information risk monitoring method of claim 1, wherein the step of extracting feature embedding vectors based on the customer feature information and the linkage feature information comprises:
inputting the customer characteristic information and the associated characteristic information into a characteristic embedding network of a preset deep learning network;
and performing data conversion on the client characteristic information and the associated characteristic information through the characteristic embedding network to obtain a characteristic embedding vector.
5. The method for risk monitoring of corporate linkage information according to claim 1, wherein the step of obtaining customer characteristic information of a corporation and linkage characteristic information of the corporate linkage party comprises:
acquiring historical credit default information of a company, creating a training label according to the historical credit default information, acquiring label characteristic information of the company, and forming customer characteristic information by the training label and the label characteristic information;
and acquiring the associated map information of the company, and acquiring the associated feature information of the company associated party according to the associated map information.
6. The method for risk monitoring of corporate linkage information according to claim 1, wherein the step of obtaining linkage characteristic information of the corporate linkage party according to the linkage map information comprises:
identifying a related party of the company according to the related map information;
and acquiring the associated characteristic information of the associated party.
7. The method for monitoring risk of company-associated information according to claim 1, wherein after the step of obtaining the characteristic information to be evaluated of the company to be evaluated, the method further comprises:
and calculating the contribution degree of the characteristic information to be evaluated through a model explanatory module to generate associated risk monitoring information.
8. A risk monitoring device for company-associated information, the risk monitoring device comprising:
the acquisition module is used for acquiring the client characteristic information of a company and the associated characteristic information of the company associated party;
an extraction module for extracting feature embedding vectors based on the customer feature information and the associated feature information;
the training module is used for training a multi-head attention module and a feature embedding network of a preset deep learning network according to the feature embedding vector to obtain a target model;
and the prediction module is used for acquiring the characteristic information to be evaluated of the company to be evaluated, and performing risk prediction on the characteristic information to be evaluated through the target model to obtain a prediction result.
9. A corporate linkage information risk monitoring device comprising a memory, a processor and a corporate linkage information risk monitoring program stored on the memory and executable on the processor, the corporate linkage information risk monitoring program when executed by the processor implementing the steps of the corporate linkage information risk monitoring method of any one of claims 1 to 7.
10. A storage medium having stored thereon a corporate-associated information risk monitoring program, the corporate-associated information risk monitoring program when executed by a processor implementing the steps of a corporate-associated information risk monitoring method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110133551.5A CN113011961B (en) | 2021-01-29 | 2021-01-29 | Method, device, equipment and storage medium for monitoring risk of company-related information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110133551.5A CN113011961B (en) | 2021-01-29 | 2021-01-29 | Method, device, equipment and storage medium for monitoring risk of company-related information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113011961A true CN113011961A (en) | 2021-06-22 |
| CN113011961B CN113011961B (en) | 2024-04-19 |
Family
ID=76384700
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110133551.5A Active CN113011961B (en) | 2021-01-29 | 2021-01-29 | Method, device, equipment and storage medium for monitoring risk of company-related information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113011961B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113989043A (en) * | 2021-10-28 | 2022-01-28 | 支付宝(杭州)信息技术有限公司 | Event risk identification method, device and equipment |
| CN116796909A (en) * | 2023-08-16 | 2023-09-22 | 浙江同信企业征信服务有限公司 | Judicial litigation risk prediction method, device, equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657918A (en) * | 2018-11-19 | 2019-04-19 | 平安科技（深圳）有限公司 | Method for prewarning risk, device and the computer equipment of association assessment object |
| CN111724083A (en) * | 2020-07-21 | 2020-09-29 | 腾讯科技（深圳）有限公司 | Training method and device for financial risk recognition model, computer equipment and medium |
| US20200357060A1 (en) * | 2019-05-10 | 2020-11-12 | Fair Ip, Llc | Rules/model-based data processing system for intelligent default risk prediction |
-
2021
- 2021-01-29 CN CN202110133551.5A patent/CN113011961B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657918A (en) * | 2018-11-19 | 2019-04-19 | 平安科技（深圳）有限公司 | Method for prewarning risk, device and the computer equipment of association assessment object |
| US20200357060A1 (en) * | 2019-05-10 | 2020-11-12 | Fair Ip, Llc | Rules/model-based data processing system for intelligent default risk prediction |
| CN111724083A (en) * | 2020-07-21 | 2020-09-29 | 腾讯科技（深圳）有限公司 | Training method and device for financial risk recognition model, computer equipment and medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113989043A (en) * | 2021-10-28 | 2022-01-28 | 支付宝(杭州)信息技术有限公司 | Event risk identification method, device and equipment |
| CN116796909A (en) * | 2023-08-16 | 2023-09-22 | 浙江同信企业征信服务有限公司 | Judicial litigation risk prediction method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113011961B (en) | 2024-04-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110009174B (en) | Risk recognition model training method and device and server | |
| US10373262B1 (en) | Image processing system for vehicle damage | |
| CN107563757B (en) | Data risk identification method and device | |
| CN113255815B (en) | User behavior abnormity analysis method, device, equipment and storage medium | |
| US10380696B1 (en) | Image processing system for vehicle damage | |
| CN112668859A (en) | Big data based customer risk rating method, device, equipment and storage medium | |
| US20100161526A1 (en) | Ranking With Learned Rules | |
| US20160335551A1 (en) | Optimization of fraud detection strategies | |
| CN108876545A (en) | Order recognition methods, device and readable storage medium storing program for executing | |
| CN113011961A (en) | Method, device and equipment for monitoring risk of company associated information and storage medium | |
| CN111754044A (en) | Employee behavior auditing method, device, equipment and readable storage medium | |
| CN116187718B (en) | Intelligent goods identification and sorting method and system based on computer vision | |
| CN111192133A (en) | Method and device for generating risk model after user loan and electronic equipment | |
| CN113205403A (en) | Method and device for calculating enterprise credit level, storage medium and terminal | |
| CN111445058A (en) | Data analysis method, device, equipment and computer readable storage medium | |
| CN111415336A (en) | Image tampering identification method and device, server and storage medium | |
| CN111353514A (en) | Model training method, image recognition method, device and terminal equipment | |
| CN111754195A (en) | Information processing method and device, electronic equipment and computer readable storage medium | |
| CN115760182A (en) | Method and device for acquiring large data of automobile industry, electronic equipment and storage medium | |
| CN114581249A (en) | Financial product recommendation method and system based on investment risk bearing capacity assessment | |
| CN114399367A (en) | Insurance product recommendation method, device, equipment and storage medium | |
| CN112634057A (en) | Fund similarity calculation method, platform, device and readable storage medium | |
| CN111245815B (en) | Data processing method and device, storage medium and electronic equipment | |
| CN113283901A (en) | Byte code-based fraud contract detection method for block chain platform | |
| CN109544348B (en) | Asset security screening method, device and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |