CN116596657A - Loan risk assessment method and device, storage medium and electronic equipment - Google Patents
Loan risk assessment method and device, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN116596657A CN116596657A CN202310539417.4A CN202310539417A CN116596657A CN 116596657 A CN116596657 A CN 116596657A CN 202310539417 A CN202310539417 A CN 202310539417A CN 116596657 A CN116596657 A CN 116596657A
- Authority
- CN
- China
- Prior art keywords
- account
- behaviors
- loan
- determining
- weight values
- 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.)
- Pending
Links
- 238000012502 risk assessment Methods 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 67
- 230000006399 behavior Effects 0.000 claims abstract description 296
- 230000015654 memory Effects 0.000 claims description 31
- 238000011156 evaluation Methods 0.000 claims description 20
- 238000013528 artificial neural network Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 13
- 230000006403 short-term memory Effects 0.000 claims description 12
- 230000007787 long-term memory Effects 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 abstract description 9
- 238000004590 computer program Methods 0.000 description 15
- 238000012549 training Methods 0.000 description 11
- 238000004422 calculation algorithm Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 9
- 230000008569 process Effects 0.000 description 8
- 206010000117 Abnormal behaviour Diseases 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 108010015780 Viral Core Proteins Proteins 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
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
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/25—Integrating or interfacing systems involving database management systems
- G06F16/258—Data format conversion from or to a database
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Biomedical Technology (AREA)
- Technology Law (AREA)
- Marketing (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The application discloses a loan risk assessment method, a loan risk assessment device, a storage medium and electronic equipment. Relates to the field of financial science and technology, and the method comprises the following steps: acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1; determining weight values corresponding to K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K; and determining a risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to one loan. The method solves the problems of low accuracy and incomplete risk prediction existing in the method for performing risk prediction based on single model output of the expert model in the related technology.
Description
Technical Field
The application relates to the field of financial science and technology, in particular to a loan risk assessment method, a loan risk assessment device, a storage medium and electronic equipment.
Background
Loans are a major core business for financial institutions, with the size of stock and new payouts increasing as economies develop. In order to develop economically well, financial risks, particularly systematic financial risks, are prevented, bad account generation is reduced, property quality is improved, and financial institutions evaluate and predict the risks before, during and after loans. The current risk monitoring mode mainly adopts an expert model to carry out corresponding risk assessment based on the business behaviors of the user, for example, when the user handles a loan for 10 years, if the user pays back normally in the previous 3 years, the monitoring model can be triggered to monitor the user or the transaction after the expiration occurs for the first time, and in fact, the user may be in a state of fund chain tension. The expert model has poor learning ability, can not perfect the accuracy of risk prediction by continuous learning, has poor risk prediction ability, and can not manage and control the risk behaviors as early as possible.
Aiming at the problems of low accuracy and incomplete risk prediction existing in the method for predicting the risk based on the single model output of the expert model in the related technology, no effective solution is proposed at present.
Disclosure of Invention
The application mainly aims to provide a loan risk assessment method, a loan risk assessment device, a storage medium and electronic equipment, so as to solve the problems of low and incomplete risk prediction accuracy in a method for performing risk prediction based on single model output of an expert model in the related technology.
To achieve the above object, according to one aspect of the present application, there is provided a loan risk assessment method. The method comprises the following steps: acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1; determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K; and determining a risk assessment result corresponding to the target account based on the weight values respectively corresponding to the K account behaviors and the loan amount corresponding to the loan.
To achieve the above object, according to another aspect of the present application, there is provided a loan risk assessment apparatus. The device comprises: the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after a loan is applied, and K is an integer greater than or equal to 1; the first determining module is used for determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K; and the second determining module is used for determining a risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the loan.
To achieve the above object, according to another aspect of the present application, there is also provided a nonvolatile storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform any one of the loan risk assessment methods described above.
In order to achieve the above object, according to another aspect of the present application, there is also 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 any one of the loan risk assessment methods described above.
According to the application, the following steps are adopted: acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1; determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K; and determining a risk assessment result corresponding to the target account based on the weight values respectively corresponding to the K account behaviors and the loan amount corresponding to the loan, so that the purpose of comprehensively determining the risk assessment result through the post-loan behaviors and the loan amount of the target account is achieved, and the problem that the risk prediction accuracy is low and incomplete in a method for performing risk prediction based on single model output of an expert model in the related technology is solved. Thereby achieving the effect of improving the accuracy and comprehensiveness of loan risk prediction.
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 loan risk assessment method provided in accordance with an embodiment of the application; and
FIG. 2 is a schematic illustration of an alternative loan risk assessment method, in accordance with an embodiment of the application;
FIG. 3 is a schematic diagram of a loan risk assessment apparatus provided in accordance with an embodiment of the application;
fig. 4 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.
For convenience of description, the following will describe some terms or terminology involved in the embodiments of the present application:
the Long Short-term memory artificial neural network model (Long Short-term memory) is a variant algorithm of the cyclic neural network algorithm. The information is allowed to pass down the sequence chain so that the information of an earlier time step can also be carried to a later time step. Some information may be added or removed by a "gate" structure that will learn what information this saves or forgets during the training process.
It should be noted that, the related information (including, but not limited to, user equipment information, user personal information, etc.) and data (including, but not limited to, data for presentation, analyzed data, etc.) related to the present application are information and data authorized by the user or sufficiently authorized by each party. For example, an interface is provided between the system and the relevant user or institution, before acquiring the relevant information, the system needs to send an acquisition request to the user or institution through the interface, and acquire the relevant information after receiving the consent information fed back by the user or institution.
The application will be described with reference to preferred implementation steps, and FIG. 1 is a flowchart of a loan risk assessment method provided according to an embodiment of the application, as shown in FIG. 1, the method comprising the steps of:
step S102, first behavior data of a target account is obtained, wherein the first behavior data are K account behaviors generated by the target account after a loan is applied, and K is an integer greater than or equal to 1.
It can be understood that the target account is an account corresponding to the target user, and the target account generates a series of behaviors after applying a loan, and the user default risk is predicted in advance by tracking the behaviors of the target account, so that the loan repayment risk is reduced.
Alternatively, the above-mentioned K account behaviors may include, but are not limited to: user basic information, user consumption information, user social relationship information, and the like.
Step S104, determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K.
Optionally, the above-mentioned types of symbolized units are obtained from a database based on historical behavior data of accounts with overdue loan repayment behaviors, and weight values corresponding to the K account behaviors are used to indicate probabilities that the K account behaviors will cause loan default. Through the method, the weight values corresponding to the K account behaviors are determined through the database comparison mode, and the probability that the K account behaviors of the target account respectively cause loan default is further obtained.
In an optional embodiment, the determining, based on the account behavior database, weight values corresponding to the K account behaviors respectively includes: based on the account behavior database, determining a weight value corresponding to any one of the K account behaviors in the following manner: obtaining similarity between the M account behaviors included in the account behavior database and any one of the account behaviors; determining a first account behavior corresponding to any one of the M account behaviors; and taking the weight value corresponding to the first account behavior as the weight value corresponding to any one of the account behaviors.
Optionally, the account behavior database includes M account behaviors, and weight values (M is greater than or equal to K) corresponding to the M account behaviors, where the weight values corresponding to the K account behaviors are performed in a database comparison manner. It should be noted that, because the text description corresponding to the same account behavior may be different due to different text expression modes, similarity calculation is performed respectively by calculating any one account behavior of the target account and the account behaviors included in the account behavior database, and when the account behavior database has the account behavior with the largest similarity with any one account behavior and greater than the preset similarity threshold, the account behavior with the largest similarity with any one account behavior and greater than the preset similarity threshold is taken as the weight value corresponding to any one account behavior.
In an optional embodiment, before determining the weight values corresponding to the K account behaviors respectively based on the account behavior database, the method further includes: acquiring historical behavior data corresponding to N accounts, wherein the N accounts are accounts with overdue loan repayment behaviors; based on the historical behavior data, a long-term and short-term memory artificial neural network model is adopted to obtain the M account behaviors and weight values corresponding to the M account behaviors respectively.
Optionally, the account behavior database is obtained based on historical behavior data of a plurality of accounts with excessive loan repayment behaviors (i.e. default behaviors), relevant information of the accounts is classified based on the historical behavior data of the accounts with excessive loan repayment behaviors, weights of various account behaviors for loan default are continuously trained through the long and short term memory neural network, and according to the weights, the judgment of which behaviors of the accounts are with high probability can lead to loan default.
In an optional embodiment, the obtaining the M account behaviors and weight values corresponding to the M account behaviors respectively based on the historical behavior data and using a long-short-term memory artificial neural network model includes: processing the historical behavior data by adopting a word embedding model to obtain first behavior data corresponding to the K account behaviors, wherein the historical behavior data are unstructured data, and the first behavior data are structured data; based on the first behavior data, the long-term and short-term memory artificial neural network model is adopted to obtain the M account behaviors and weight values corresponding to the M account behaviors respectively.
Optionally, based on the historical behavior data of the account with the overdue loan repayment behavior, classifying the historical behavior data according to M account behaviors to obtain classified historical behavior data; and converting the unstructured classified historical behavior data into structured first behavior data through a word embedding model (namely a word2Vec pre-training model), learning and training different types of data (namely different account behaviors) in the first behavior data through a Long Short-term memory neural network (namely a Long Short-term memory, LSTM model), and finally training out weight outputs corresponding to the M account behaviors respectively. These weight outputs are used for later post-credit risk assessment for the target account.
Optionally, processing the historical behavior data by using a word embedding model to obtain first behavior data corresponding to the K account behaviors, where the processing includes: the historical behavior data of accounts for which loan overdue repayment behaviors exist is classified, for example:
user personal information (age, sex, wedding, work, education level, personal asset, information collection time)
Post-loan consumption information (consumption frequency, average consumption per pen, average consumption per day, consumption category, information acquisition time)
User social relationship information (social credit level, whether there is abnormal behavior experience, information acquisition time)
Establishing a data conversion dictionary, and converting the classified historical behavior data in an unstructured form into structured first behavior data, wherein the classified historical behavior data in the unstructured form is as follows:
user personal information (25, man, marriage, software engineer, master, high assets, 2022-01-01)
Post-loan consumption information (20/day, 50, 1000, electronic goods, 2022-01-01)
User social relationship information (no belief record, no abnormal behavior record, 2022-01-01)
The first row of data in the converted structured form is as follows:
user personal information (25, 1, 0, 4, 3, 2022-01-01)
Post-loan consumption information (20, 50, 1000, 7, 2022-01-01)
User social relationship information (4, 5, 2022-01-01)
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.
Alternatively, the long-term and short-term memory artificial neural network algorithm is a variant of the recurrent neural network algorithm. The information is allowed to pass down the sequence chain so that the information of an earlier time step can also be carried to a later time step. Some information can be added or removed through a door structure, the door structure is divided into a forgetting door, an input door and an output door, and the information which is saved or forgotten can be learned in the training process, so that important information and unimportant information can be screened. Wherein the input gate also acts on the cell state, and new information can be selectively recorded in the cell state. By doing so, meaningless or insignificant information in the newly added information at each time step can be removed, so that the cell state at the moment can be calculated so that the cell state at the moment cannot be influenced by redundant useless information. The cell state at the current time t is formed by the influence of the forgetting gate control history information and the influence of the input gate control input at the current time. It is determined by the output gate how much of the current hidden node state is externally visible. By means of the three gate structures, the influence of historical information and current input on subsequent results can be controlled, and when training reaches a certain degree, possible situations in subsequent time steps can be predicted more scientifically. The method is used as the weight value of different account behaviors, and the influence of different types of information on risk assessment can be effectively predicted, so that the risk assessment result is more scientific and accurate.
Through the training of the LSTM model, different weight values can be given to different categories of information of the user, and the weight values represent the influence of the information on loan default. For example, consider a young engineer of high school to be less likely to violate, while a young, older, unmarketed man of low school is more likely to violate, i.e., to manifest itself as:
user personal information (25, male, not married, software engineer, master, high asset) weight value: 2
User personal information (35, male, unmarketed, no business, high, low asset) weight value: 9
Others such as:
post-loan consumption information (5/day, 20, 100, daily overhead) weight value: 2
Post-loan consumption information (20/day, 50, 1000, entertainment, 2022-01-01) weight value: 7
User social relationship information (no trust record, no abnormal behavior record) weight value: 2
User social relationship information (no trust record, no abnormal behavior record) weight value: 5
After the training is finished, a series of some default core account behaviors (namely the K account behaviors) are obtained, and the account behaviors and the weight values trained corresponding to the account behaviors are stored in an account behavior database together.
Optionally, after paying for a specific loan of the target account, determining weight values corresponding to the K account behaviors respectively through a database comparison mode, so as to obtain probabilities that the K account behaviors of the target account respectively cause loan default, and evaluating the loan risk of the target account by a user.
Step S106, determining a risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the loan.
Optionally, the risk assessment result is used for indicating that the target account cannot pay the risk assessment result of the loan according to the period after performing the loan.
In an optional embodiment, the determining the risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the one loan includes: determining a risk assessment value corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the loan; judging whether the risk evaluation value is larger than a preset risk threshold value or not; and determining that the risk assessment result is that the target account has risk behaviors under the condition that the risk assessment value is larger than the preset risk threshold value.
Optionally, the determining the risk assessment value corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the one loan includes: based on the weight values respectively corresponding to the K account behaviors and the loan amount corresponding to the loan, determining a risk evaluation value corresponding to the target account in the following manner:
where Xi (i=1, 2, …, K) is a weight value corresponding to any one of the K account behaviors, Y is a preset limit value, and Z is a loan amount corresponding to the one loan.
Optionally, based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the one loan, the weight values corresponding to the K account behaviors are calculated by using the formula, for example, when the K account behaviors include user personal information, post-loan consumption information and user social relationship information, the K account behaviors matched to the target account and the weight values corresponding to the K account behaviors are as follows:
user personal information (25, male, not married, software engineer, master, high asset) weight value: 2
Post-loan consumption information (20/day, 50, 1000, entertainment, 2022-01-01) weight value: 7
User social relationship information (no trust record, no abnormal behavior record) weight value: 5
Taking risk weight values of the three matched default core behaviors (namely weight values corresponding to 3 account behaviors): x1=2, x2=7, x3=5. It should be noted that, the larger the loan amount is, the lower the tolerance to the risk is, so the loan amount is introduced as another factor for evaluating the risk, the loan amount is taken as an independent influence factor Z, meanwhile, a maximum upper limit value y=100 is preset, the result of risk evaluation is calculated by adopting the above-mentioned W formula, the risk evaluation result is compared with the preset risk threshold, and then a decision is made.
In an alternative embodiment, the method further comprises: and determining a risk threshold interval to which the risk evaluation value belongs when the risk evaluation value is greater than the preset risk threshold value: and determining a decision result aiming at the target account according to the risk threshold value interval to which the risk evaluation value belongs.
Optionally, the number of the preset risk thresholds may be one or more, table 1 shows a relationship between the risk evaluation value and the corresponding decision result, and as shown in table 1, the closer W is to 0, the more controllable the risk is, so that the normal use of the loan can be maintained; if W is closer to 1, the risk is higher, and a decision to collect or freeze loan funds can be made to avoid that after the loan is paid out, the risk is not found until the user is overdue because of insufficient post-evaluation of the loan risk. By the method, some risk thresholds are set, the calculated risk evaluation value is compared with the preset risk threshold, and when the risk evaluation value accords with a certain preset risk threshold, corresponding measures are taken to control the loan funds of the user. Therefore, the risk management and control of the loan is carried out after the loan is released, and the economic loss caused by the fact that a user cannot repayment is reduced as much as possible.
TABLE 1
Through the steps S102 to S106, the purpose of comprehensively determining the risk assessment result through the post-loan behavior and the loan amount of the target account can be achieved, and the problems of low and incomplete risk prediction accuracy existing in the method for performing risk prediction based on single model output of the expert model in the related technology are solved. Thereby achieving the effect of improving the accuracy and comprehensiveness of loan risk prediction.
Based on the above embodiment and the optional embodiment, the present application proposes an optional implementation, and fig. 2 is a flowchart of an optional loan risk assessment method according to an embodiment of the present application, as shown in fig. 2, the method includes: account behavior database construction, risk assessment after lending and decision output, wherein:
in the construction stage of an account behavior database, based on the existing historical behavior data with the violations, classifying the historical behavior data according to the violations (namely, expected loan repayment behaviors) to obtain classified historical behavior data (comprising three types of account behaviors including user basic information, user consumption information and user social relationship information), converting unstructured historical behavior data classified by a word2Vec pre-training model into structured first behavior data, inputting the first behavior data into an LSTM model for training to obtain a plurality of account behaviors (namely, a plurality of core violations) and weight values corresponding to the account behaviors respectively; an account behavior database is formed based on the plurality of account behaviors and the corresponding weight values.
In the post-loan risk assessment stage, acquiring 3 types of account behaviors (namely user basic information, user consumption information and user social relationship information) of a target user after a loan is initiated, comparing the 3 types of account behaviors of the target user with the account behaviors stored in an account behavior database, and determining weight values corresponding to the 3 types of account behaviors of the target account respectively; and determining the risk assessment value of the target account through the W calculation formula based on the weight value corresponding to the 3-class account behaviors of the target account and the loan amount of the target account.
And in the decision output stage, comparing the calculated risk evaluation value with a preset risk threshold value, and obtaining a final decision result according to the comparison result.
It should be noted that, according to the application, by introducing the long-short-term memory artificial neural network algorithm, the high-risk behavior is extracted by a deep learning method based on the existing massive user financial behavior data of the financial institution, so as to deduce the core behavior which causes that the user cannot repay the loan, that is, the weight value which may cause the default of the loan due to different behaviors of the user is trained by the LSTM algorithm, and the behavior with the weight value greater than a certain threshold is marked as the core behavior. After a loan is paid out, the user is tracked to obtain various behavior information after the loan, and the behavior information is compared with trained core behaviors which can cause default. By calculating a risk assessment value for the post-credit behavior of the user hitting the offending core behavior, corresponding risk management measures can be implemented for different value ranges of the risk value. Therefore, the risk prediction and corresponding management and control can be carried out by technical means before the true default of the loan, and the economic loss is reduced.
The application can at least realize the following technical effects: and after the loan is paid, the risk management and control of the loan is carried out, so that the economic loss caused by the fact that the user cannot pay back is reduced as much as possible. Through continuous training, the neural network algorithm can obtain more reasonable weight values, the influence of different behaviors on loan violations is distinguished, and the risk assessment value is calculated by utilizing the violating core behaviors more accurately, so that the risk management and control is more reasonable.
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 loan risk assessment device, and the loan risk assessment device can be used for executing the loan risk assessment method provided by the embodiment of the application. The loan risk assessment device provided by the embodiment of the application is described below.
Fig. 3 is a schematic diagram of a loan risk assessment apparatus according to an embodiment of the application. As shown in fig. 3, the apparatus includes: an acquisition module 300, a first determination module 302, a second determination module 304, wherein,
The obtaining module 300 is configured to obtain first behavior data of a target account, where the first behavior data is K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1;
the first determining module 302, coupled to the obtaining module 300, is configured to determine weight values corresponding to the K account behaviors based on an account behavior database, where the account behavior database includes M account behaviors, and the weight values corresponding to the M account behaviors are integers greater than or equal to K;
the second determining module 304 is connected to the first determining module 302, and is configured to determine a risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the one loan.
In the present application, the acquiring module 300 is configured to acquire first behavior data of a target account, where the first behavior data is K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1; the first determining module 302, coupled to the obtaining module 300, is configured to determine weight values corresponding to the K account behaviors based on an account behavior database, where the account behavior database includes M account behaviors, and the weight values corresponding to the M account behaviors are integers greater than or equal to K; the second determining module 304 is connected to the first determining module 302, and is configured to determine a risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the one loan, so as to achieve the purpose of comprehensively determining the risk assessment result through the post-loan behavior of the target account and the loan amount, and solve the problem of low and incomplete risk prediction accuracy in the method for performing risk prediction based on the single model output of the expert model in the related art. Thereby achieving the effect of improving the accuracy and comprehensiveness of loan risk prediction.
It should be noted that each of the above modules may be implemented by software or hardware, for example, in the latter case, it may be implemented by: the above modules may be located in the same processor; alternatively, the various modules described above may be located in different processors in any combination.
It should be noted that, the acquiring module 300, the first determining module 302, and the second determining module 304 correspond to steps S102 to S106 in the embodiment, and the modules are the same as the examples and application scenarios implemented by the corresponding steps, but are not limited to the disclosure of the embodiment. It should be noted that the above modules may be run in a computer terminal as part of the apparatus.
It should be noted that, the optional or preferred implementation manner of this embodiment may be referred to the related description in the embodiment, and will not be repeated herein.
The loan risk assessment device comprises a processor and a memory, wherein the units and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.
The processor includes a kernel, and the kernel fetches the corresponding program unit from the memory. The kernel may be provided with one or more by adjusting the kernel parameters (object of the present application).
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 application provides a nonvolatile storage medium having a program stored thereon, which when executed by a processor, implements the loan risk assessment method described above.
The embodiment of the application provides a processor, which is used for running a program, wherein the loan risk assessment method is executed when the program runs.
As shown in fig. 4, an embodiment of the present application provides an electronic device, where the electronic device 10 includes a processor, a memory, and a program stored on the memory and executable on the processor, and the processor implements the following steps when executing the program: acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1; determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K; and determining a risk assessment result corresponding to the target account based on the weight values respectively corresponding to the K account behaviors and the loan amount corresponding to the loan. 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: acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1; determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K; and determining a risk assessment result corresponding to the target account based on the weight values respectively corresponding to the K account behaviors and the loan amount corresponding to the loan.
Optionally, the above computer program product is further adapted to execute a program initialized with the method steps of: based on the account behavior database, determining a weight value corresponding to any one of the K account behaviors in the following manner: obtaining similarity between the M account behaviors included in the account behavior database and any one of the account behaviors; determining a first account behavior corresponding to any one of the M account behaviors; and taking the weight value corresponding to the first account behavior as the weight value corresponding to any one of the account behaviors.
Optionally, the above computer program product is further adapted to execute a program initialized with the method steps of: determining a risk assessment value corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the loan; judging whether the risk evaluation value is larger than a preset risk threshold value or not; and determining that the risk assessment result is that the target account has risk behaviors under the condition that the risk assessment value is larger than the preset risk threshold value.
Optionally, the above computer program product is further adapted to execute a program initialized with the method steps of: and determining a risk threshold interval to which the risk evaluation value belongs when the risk evaluation value is greater than the preset risk threshold value: and determining a decision result aiming at the target account according to the risk threshold interval to which the risk evaluation value belongs.
Optionally, the above computer program product is further adapted to execute a program initialized with the method steps of: acquiring historical behavior data corresponding to N accounts, wherein the N accounts are accounts with overdue loan repayment behaviors; based on the historical behavior data, a long-term and short-term memory artificial neural network algorithm is adopted to obtain M account behaviors and weight values corresponding to the M account behaviors respectively.
Optionally, the above computer program product is further adapted to execute a program initialized with the method steps of: processing the historical behavior data by adopting a word embedding model to obtain first behavior data corresponding to the K account behaviors, wherein the historical behavior data are unstructured data, and the first behavior data are structured data; based on the first behavior data, the M account behaviors and weight values corresponding to the M account behaviors are obtained by adopting the long-term and short-term memory artificial neural network algorithm.
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 magnetic 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.
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 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 (10)
1. A loan risk assessment method, comprising:
acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after applying a loan, and K is an integer greater than or equal to 1;
determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K;
And determining a risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the loan.
2. The method of claim 1, wherein determining the weight values for the K account behaviors, respectively, based on the account behavior database, comprises:
based on the account behavior database, determining a weight value corresponding to any one of the K account behaviors in the following manner:
obtaining similarity between the M account behaviors included in the account behavior database and any one of the account behaviors;
determining a first account behavior corresponding to any one account behavior from the M account behaviors; and taking the weight value corresponding to the first account behavior as the weight value corresponding to any one account behavior.
3. The method of claim 1, wherein the determining the risk assessment result corresponding to the target account based on the weight values respectively corresponding to the K account behaviors and the loan amount corresponding to the one loan comprises:
determining a risk assessment value corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the loan;
Judging whether the risk assessment value is larger than a preset risk threshold value or not;
and determining that the risk assessment result is that the target account has risk behaviors under the condition that the risk assessment value is larger than the preset risk threshold value.
4. The method of claim 3, wherein the determining the risk assessment value corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the one loan, respectively, comprises:
based on the weight values respectively corresponding to the K account behaviors and the loan amount corresponding to the loan, determining a risk evaluation value corresponding to the target account by the following method:
where Xi (i=1, 2, …, K) is a weight value corresponding to any one of the K account behaviors, Y is a preset limit value, and Z is a loan amount corresponding to the one loan.
5. A method according to claim 3, characterized in that the method further comprises:
and determining a risk threshold interval to which the risk evaluation value belongs when the risk evaluation value is larger than the preset risk threshold value:
and determining a decision result aiming at the target account according to the risk threshold interval to which the risk evaluation value belongs.
6. The method of any one of claims 1 to 5, wherein prior to determining the weight values for the K account behaviors, respectively, based on an account behavior database, the method further comprises:
acquiring historical behavior data corresponding to N accounts, wherein the N accounts are accounts with overdue loan repayment behaviors;
based on the historical behavior data, a long-term and short-term memory artificial neural network model is adopted to obtain the M account behaviors and weight values corresponding to the M account behaviors respectively.
7. The method of claim 6, wherein the obtaining the M account behaviors and the weight values corresponding to the M account behaviors respectively using a long-short-term memory artificial neural network model based on the historical behavior data comprises:
processing the historical behavior data by adopting a word embedding model to obtain first behavior data corresponding to the K account behaviors, wherein the historical behavior data are unstructured data, and the first behavior data are structured data;
based on the first behavior data, the long-term and short-term memory artificial neural network model is adopted to obtain the M account behaviors and weight values corresponding to the M account behaviors respectively.
8. A loan risk assessment device, comprising:
the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring first behavior data of a target account, wherein the first behavior data are K account behaviors generated by the target account after a loan is applied, and K is an integer greater than or equal to 1;
the first determining module is used for determining weight values corresponding to the K account behaviors respectively based on an account behavior database, wherein the account behavior database comprises M account behaviors and weight values corresponding to the M account behaviors respectively, and M is an integer greater than or equal to K;
and the second determining module is used for determining a risk assessment result corresponding to the target account based on the weight values corresponding to the K account behaviors and the loan amount corresponding to the loan.
9. A non-volatile storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the loan risk assessment method of 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 loan risk assessment method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310539417.4A CN116596657A (en) | 2023-05-12 | 2023-05-12 | Loan risk assessment method and device, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310539417.4A CN116596657A (en) | 2023-05-12 | 2023-05-12 | Loan risk assessment method and device, storage medium and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116596657A true CN116596657A (en) | 2023-08-15 |
Family
ID=87607467
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310539417.4A Pending CN116596657A (en) | 2023-05-12 | 2023-05-12 | Loan risk assessment method and device, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116596657A (en) |
-
2023
- 2023-05-12 CN CN202310539417.4A patent/CN116596657A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11132624B2 (en) | Model integration method and device | |
| CN110490304B (en) | Data processing method and device | |
| CN113554228B (en) | Training method of repayment rate prediction model and repayment rate prediction method | |
| CN112200402B (en) | Risk quantification method, device and equipment based on risk portrait | |
| CN116450951A (en) | Service recommendation method and device, storage medium and electronic equipment | |
| CN112766814A (en) | Training method, device and equipment for credit risk pressure test model | |
| CN110163470B (en) | Event evaluation method and device | |
| CN116630020A (en) | Risk assessment method and device, storage medium and electronic equipment | |
| KR20210063005A (en) | Prediction method of price changing in real estate market using temporal causal modeling | |
| CN115860924A (en) | Supply chain financial credit risk early warning method and related equipment | |
| CN116596657A (en) | Loan risk assessment method and device, storage medium and electronic equipment | |
| CN112308294B (en) | Method and device for predicting default probability | |
| Zakowska | A New Credit Scoring Model to Reduce Potential Predatory Lending: A Design Science Approach | |
| Locatelli et al. | Artificial Intelligence and Credit Risk | |
| CN114861680B (en) | Dialogue processing method and device | |
| KR102409019B1 (en) | System and method for risk assessment of financial transactions and computer program for the same | |
| Kaisar et al. | Explainable Machine Learning Models for Credit Risk Analysis: A Survey | |
| CN117557086B (en) | Secret-related carrier supervision method, device, equipment and readable storage medium | |
| CN118297736A (en) | Processing method and device of fund information, storage medium and electronic equipment | |
| CN116777624A (en) | Method, device, electronic equipment and storage medium for evaluating reliability of risk decision | |
| CN116737904A (en) | Data information processing method and device, storage medium and electronic equipment | |
| CN116934087A (en) | Financial risk assessment method and device and electronic equipment | |
| CN116823440A (en) | Method for obtaining scientific enterprise risk scoring model, risk scoring method and device | |
| CN116523648A (en) | Risk object identification method and device, storage medium and electronic equipment | |
| CN114757763A (en) | Object capability prediction method and device, storage medium and electronic device |
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 |