Blockchain-based Loan Risk Prediction Method and Device

Abstract

The present invention discloses a loan risk prediction method and device based on blockchain, relating to the field of blockchain. The method includes: receiving a loan application request from a loan customer; obtaining key loan element information submitted by the loan customer according to the loan application request; comparing the key loan element information with the customer information of known loan risk customers, and determining a key loan element evaluation result corresponding to the key loan element information according to the comparison result; determining a set of loan risk types of the loan customer according to the key loan element evaluation result; searching for a risk prediction model corresponding to the loan risk type from a set of risk prediction models according to the set of loan risk types of the loan customer; and performing risk prediction on the loan customer based on one or more found risk prediction models to obtain a risk prediction result of the loan customer. The present invention can not only shorten the time-consuming of the model training process, but also improve the accuracy of the overall model prediction.

Description

Technical Field

[0001] The present invention relates to the field of blockchain, and in particular, to a loan risk prediction method and device based on blockchain. Background Art

[0002] This section aims to provide background or context for the embodiments of the present invention described in the claims. The description herein is not admitted to be prior art merely by including it in this section.

[0003] When a customer applies for a loan from a banking institution, the banking institution will predict the loan risk of the customer in order to determine whether to grant the corresponding loan to the customer according to the risk prediction result. With the help of artificial intelligence technology, a risk prediction model can be trained through machine learning to predict the risk of loan customers.

[0004] Due to the diversification of customer information and loan operations, there are technical problems of long training time and low model accuracy in training a model that can accurately predict the loan risk of loan customers. Summary of the Invention

[0005] An embodiment of the present invention provides a loan risk prediction method based on blockchain to solve the technical problems of long training time and low model accuracy in the existing loan risk prediction method with the help of artificial intelligence technology. The method includes: receiving a loan application request from a loan customer; obtaining key loan element information submitted by the loan customer according to the loan application request; comparing the key loan element information with the customer information of known loan risk customers, and determining the key loan element evaluation result corresponding to the key loan element information according to the comparison result, wherein the customer information of known loan risk customers includes: the key loan element information and key loan element evaluation result of known loan risk customers; determining a loan risk type set of the loan customer according to the key loan element evaluation result, wherein the loan risk type set includes one or more loan risk types of the loan customer; searching for a risk prediction model corresponding to the loan risk type from a risk prediction model set according to the loan risk type set of the loan customer, wherein the risk prediction model set includes: multiple risk prediction models uploaded by each lending institution through a blockchain network; and performing a risk prediction on the loan customer based on the found one or more risk prediction models to obtain a risk prediction result of the loan customer.

[0006] In an embodiment of the present invention, there is also provided a loan risk prediction device based on blockchain, which is used to solve the technical problems that the existing loan risk prediction method using artificial intelligence technology has a long training process and a low model accuracy rate. The device includes: a loan application module for receiving a loan application request from a loan customer; a key loan element information determination module for obtaining the key loan element information submitted by the loan customer according to the loan application request; a key loan element evaluation result acquisition module for comparing the key loan element information with the customer information of known loan risk customers and determining the key loan element evaluation result corresponding to the key loan element information according to the comparison result, wherein the customer information of known loan risk customers includes: the key loan element information and key loan element evaluation result of known loan risk customers; a customer loan risk type determination module for determining a set of loan risk types of the loan customer according to the key loan element evaluation result, wherein the set of loan risk types includes one or more loan risk types of the loan customer; a risk prediction model selection module for searching for a risk prediction model corresponding to the loan risk type from a set of risk prediction models according to the set of loan risk types of the loan customer, wherein the set of risk prediction models includes: multiple risk prediction models uploaded by each lending institution through a blockchain network; a risk prediction module for performing risk prediction on the loan customer based on the found one or more risk prediction models to obtain a risk prediction result of the loan customer.

[0007] In an embodiment of the present invention, there is also provided a computer device, which is used to solve the technical problems that the existing loan risk prediction method using artificial intelligence technology has a long training process and a low model accuracy rate. The computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned loan risk prediction method based on blockchain is implemented.

[0008] In an embodiment of the present invention, there is also provided a computer-readable storage medium, which is used to solve the technical problems that the existing loan risk prediction method using artificial intelligence technology has a long training process and a low model accuracy rate. The computer-readable storage medium stores a computer program for executing the above-mentioned loan risk prediction method based on blockchain.

[0009] In the embodiments of the present invention, a blockchain-based loan risk prediction method, device, computer device, and computer-readable storage medium receive risk prediction models uploaded by each lending institution through a blockchain network to form a risk prediction model set. After receiving a loan application request from a loan customer, according to the loan application request of the loan customer, a loan risk type set of the loan customer is determined. Then, according to each loan risk type of the loan customer in the loan risk type set, a risk prediction model corresponding to the loan risk type is searched from the risk prediction model set. Finally, based on the one or more risk prediction models found, risk prediction is performed on the loan customer to obtain a risk prediction result of the loan customer.

[0010] Compared with the prior art technical solution of directly training a single model to predict the risk of loan customers, in the embodiments of the present invention, each lending institution trains multiple risk prediction models respectively, selects a risk prediction model corresponding to the loan risk type of the current loan application request of the loan customer from the multiple risk prediction models, and performs loan risk prediction on the loan customer, which can not only shorten the time-consuming of the model training process, but also improve the accuracy of the overall model prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. In the drawings:

[0012] Figure 1 It is a flowchart of a blockchain-based loan risk prediction method provided in the embodiments of the present invention;

[0013] Figure 2 It is a flowchart of a loan risk prediction provided in the embodiments of the present invention;

[0014] Figure 3 It is an alternative flowchart of a loan risk prediction provided in the embodiments of the present invention;

[0015] Figure 4 It is a flowchart of uploading a risk prediction result to the blockchain provided in the embodiments of the present invention;

[0016] Figure 5 It is a flowchart of uploading loan business data to the blockchain provided in the embodiments of the present invention;

[0017] Figure 6 It is a flowchart of a loan application based on blockchain and 5G messaging provided in the embodiments of the present invention;

[0018] Figure 7 A flowchart for generating a set of customer loan risk types provided in an embodiment of the present invention;

[0019] Figure 8 A flowchart for machine learning provided in an embodiment of the present invention;

[0020] Figure 9 A schematic diagram of a loan risk prediction device based on blockchain provided in an embodiment of the present invention;

[0021] Figure 10 A schematic diagram of an optional loan risk prediction device provided in an embodiment of the present invention;

[0022] Figure 11 A schematic diagram of a computer device provided in an embodiment of the present invention. Detailed implementation manners

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer and more understandable, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Herein, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but not to limit the present invention.

[0024] An embodiment of the present invention provides a loan risk prediction method based on blockchain. Figure 1 A flowchart of a loan risk prediction method based on blockchain provided in an embodiment of the present invention, as Figure 1 shown, the method includes the following steps:

[0025] S101, receiving a loan application request from a loan customer.

[0026] It should be noted that the loan application request in the embodiment of the present invention may be a loan application request initiated by a loan customer to any bank institution. Optionally, the loan application request includes at least: customer information of the loan customer and product information of the currently applied loan product.

[0027] S102, obtaining key loan element information submitted by the loan customer according to the loan application request;

[0028] It should be noted that the key loan element information obtained in S102 above includes but is not limited to: loan type, asset information of the customer, business conditions of corporate customers, industry information of the industry where the customer is located, loan purpose, credit information, historical repayment situation, and other information.

[0029] S103. Compare the key loan element information with the customer information of known loan risk customers, and determine the key loan element evaluation result corresponding to the key loan element information according to the comparison result. Among them, the customer information of known loan risk customers includes: the key loan element information and key loan element evaluation result of known loan risk customers;

[0030] It should be noted that the above customer information of known loan risk customers can be stored in the blockchain network to ensure the authenticity and immutability of the data. The above known loan risk customers can be customers with pre-stored known loan risk prediction results in the banking system. The key loan element information and key loan element evaluation results of these customers are stored in an associated manner. So that after receiving a loan application request from a new loan customer, according to the loan application request, obtain the key loan element information submitted by the loan customer, and then match the key loan element information submitted by the loan customer with the key loan element information of known loan risk customers, and determine the key loan element evaluation result of the customer with a successful match as the key loan element evaluation result of the loan customer.

[0031] S104. Determine the loan risk type set of the loan customer according to the key loan element evaluation result, where the loan risk type set includes one or more loan risk types of the loan customer.

[0032] Since for different loan operations and different customer information, the corresponding loan risk types are also different. In the embodiments of the present invention, after receiving a loan application request from a loan customer, according to the customer information of the loan customer included in the loan application request and the product information of the currently applied loan product, one or more loan risk types for evaluating the loan customer can be determined, and according to the determined loan risk types, a loan risk type set of the loan customer is formed. The loan risk types included in the loan risk type set in the embodiments of the present invention include but are not limited to the following risks: fraud risk, material fraud risk, bankruptcy risk, illegal fund transfer risk, default risk, etc.

[0033] In one embodiment, after finding the risk prediction model corresponding to the loan risk type from the risk prediction model set according to the loan risk type set of the loan customer, the loan risk prediction method based on blockchain provided in the embodiments of the present invention may further include the following steps: Determine the similar customers of the loan customer based on the pre-constructed customer relationship knowledge graph; Add the loan risk types of the similar customers to the loan risk type set of the loan customer.

[0034] In the embodiments of the present invention, adding the loan risk types of similar customers to the loan risk type set of the loan customer can perform risk prediction on more loan risk types of the loan customer, making the finally generated risk prediction result more accurate.

[0035] S105. Based on the loan risk type set of the loan customer, search for the risk prediction model corresponding to the loan risk type from the risk prediction model set, where the risk prediction model set includes: multiple risk prediction models uploaded by each lending institution through the blockchain network.

[0036] Due to different training sample data, the time-consuming of the model training process is also different. In the embodiments of the present invention, by making full use of the historical loan risk data of each lending institution, multiple machine learning models such as support vector machine models, Bayesian models, and neural network models are trained, and multiple risk prediction models of each lending institution can be obtained.

[0037] Based on the loan risk type set of the loan customer, search for the risk prediction model corresponding to the loan risk type from the risk prediction model set, where each loan risk prediction model can be uploaded to the blockchain network by each lending institution for use by the institutions corresponding to each node of the common blockchain. For example, some lending institutions use Bayesian models to warn of default risks and use deep learning models to warn of default risks. In this way, the lending institutions can upload the models with relatively high prediction performance to the blockchain.

[0038] S106. Based on the one or more risk prediction models found, perform risk prediction on the loan customer to obtain the risk prediction result of the loan customer.

[0039] It should be noted that if the loan risk type set of the loan customer includes one loan risk type, search for one or more risk prediction models corresponding to the loan risk type from the risk prediction model set; if the loan risk type set of the loan customer includes multiple loan risk types, search for one or more risk prediction models corresponding to each loan risk type from the risk prediction model set; using each risk prediction model, risk prediction can be performed on the loan customer to obtain multiple risk prediction sub-results, and then by integrating each risk prediction sub-result, the risk prediction result of the loan customer can be obtained.

[0040] Common integrated algorithm models include: Bagging algorithm, Boosting algorithm, Stacking algorithm, etc. In the embodiments of the present invention, any one of the above three integrated algorithms can be used to integrate multiple risk prediction sub-results in specific implementation. Generally, in order to build an integrated model with good generalization performance, individual learning algorithms should be good and different. For example, when predicting a certain type of risk (such as default risk), one or more risk prediction models found above can be classified according to the type of individual risk prediction models (for example, support vector machine, Bayesian network, deep learning algorithms with different network structures), and then the model with the best performance is selected from each type of model, and then the selected models are integrated to obtain the prediction result of this type of risk.

[0041] In the embodiments of the present invention, the loan requests of customers can be predicted based on each model respectively, and then the prediction results of each model are integrated based on various integrated algorithms of machine learning configured by the smart contract, and finally a final conclusion is obtained: that is, whether there is a corresponding risk (such as bankruptcy risk, default risk) in the customer's transaction and the probability of the risk occurrence.

[0042] In one embodiment, as Figure 2 shown, the blockchain-based loan risk prediction method provided in the embodiments of the present invention can be implemented through the following steps when predicting the risk of loan customers based on one or more found risk prediction models:

[0043] S201, obtain the customer information of the loan customer;

[0044] S202, input the customer information of the loan customer into each found risk prediction model to obtain the risk prediction results of each risk prediction model;

[0045] S203, generate the risk prediction result of the loan customer according to the risk prediction results of each risk prediction model.

[0046] It should be noted that the customer information of the loan customer obtained in the above S201 includes but is not limited to: the customer's asset information, the business data of corporate customers, the industry data of corporate customers, the historical predicted customer behavior data, and the historical risk prediction results.

[0047] In the specific implementation of the above S203, it can be implemented through the following steps: use the smart contract pre-stored on the blockchain network to select the target integrated algorithm model to be adopted by each lending institution from multiple integrated algorithm models pre-stored on the blockchain network; use the selected target integrated algorithm model to integrate the risk prediction results of each risk prediction model to obtain the risk prediction result of the loan customer; push the risk prediction result of the loan customer to each lending institution.

[0048] Optionally, the target integrated algorithm model to be adopted selected by each lending institution can be a model with relatively good performance. Further, each lending institution can upload the integrated algorithm model it selects, and the blockchain network can determine the performance of the integrated algorithm models uploaded by each lending institution based on historical loan business data.

[0049] In one embodiment, as Figure 3 shown, when generating the risk prediction result of a loan customer according to the risk prediction results of each risk prediction model in the loan risk prediction method based on blockchain provided in the embodiment of the present invention, the following steps may be specifically included:

[0050] S301, obtain the performance evaluation results of each risk prediction model;

[0051] S302, determine the weight coefficients of each risk prediction model according to the performance evaluation results of each risk prediction model;

[0052] S303, generate the risk prediction result of the loan customer according to the weight coefficients and risk prediction results of each risk prediction model.

[0053] In one embodiment, as Figure 4 shown, after predicting the risk of a loan customer based on one or more found risk prediction models to obtain the risk prediction result of the loan customer, the loan risk prediction method based on blockchain provided in the embodiment of the present invention may further include the following steps:

[0054] S401, upload the risk prediction result of the loan customer to the blockchain network;

[0055] S402, execute the loan application service of the loan customer based on the risk prediction result of the loan customer stored on the blockchain network, wherein when the risk prediction result corresponding to any risk type of the loan customer exceeds the preset risk condition, the loan application service of the loan customer is rejected.

[0056] Through the above embodiments, uploading the risk prediction result to the blockchain network can facilitate subsequent execution of other services based on the risk prediction result stored on the blockchain network, which can be but is not limited to loan application services.

[0057] When applying the loan risk prediction method based on blockchain provided in the embodiment of the present invention to the loan risk prediction of corporate or personal customers of a bank, it may specifically include:

[0058] 1) Receive a loan application request from a loan customer and upload the loan application request data to the blockchain. Based on this loan application request, obtain the key loan elements submitted by the customer, such as the loan type submitted by the customer, the customer's asset information, the customer's company operation status, the industry information where the customer is located, the loan purpose, the credit information of the loan customer, the customer's other loans, and the repayment status of previous loans.

[0059] 2) Obtain the customer information stored in the blockchain by each lending institution, including the customer data collected by each branch of the bank, the behavioral prediction result data of the customer by the bank's customer behavior prediction model, the analysis data of the industry where the customer is located, and the previous risk prediction results of the customer at the bank, and compare and analyze this information with the key loan elements to obtain the evaluation result of the key loan elements.

[0060] 3) According to the evaluation results of the above key loan elements, determine the set of loan risk types of the loan customer. For example, the set of loan risk types includes fraud risk, material fraud risk, bankruptcy risk, money laundering risk, default risk, etc.

[0061] 4) According to the set of loan risk types of the loan customer, search for the risk prediction models corresponding to the loan risk types from the set of risk prediction models. Each loan risk prediction model can be uploaded to the blockchain network by each lending institution for use by the institutions corresponding to each node of the blockchain. For example, some lending institutions use the Bayesian model to warn of default risk and the deep learning model to warn of default risk. In this way, the lending institution can upload the model with relatively high prediction performance to the blockchain.

[0062] 5) Based on one or more loan risk prediction models found from the blockchain, conduct a loan risk prediction on the loan customer to obtain the risk prediction result of the loan customer. For example, the loan request of the customer can be predicted based on each model respectively, and then based on various integration algorithms of machine learning configured by the smart contract, the prediction results of each model are integrated, and finally a final conclusion is obtained: that is, whether there is a corresponding risk for the customer's transaction, such as bankruptcy risk, default risk, and the probability of the risk occurring.

[0063] 6) Based on the above-obtained risk prediction results, determine whether to grant a loan to the customer and the recommended loan amount, and upload this data to the blockchain. For example, upload information data such as a 0.1 probability of default risk and a maximum loan amount of 100,000 yuan to the blockchain.

[0064] 7) The loan application information, risk prediction results, and subsequent loan execution risk results of each loan application are uploaded to the blockchain. Then, based on this data, the performance of the risk prediction model is evaluated and its parameters are adjusted. Previous model evaluations may not be accurate due to insufficient sample data, and the newly emerged sample data can exactly make up for the previous deficiencies.

[0065] Further, as Figure 5 shown, after executing the loan application business of the loan customer based on the risk prediction results of the loan customer stored on the blockchain network, the blockchain-based loan risk prediction method provided in the embodiments of the present invention may further include the following steps:

[0066] S501, obtain the business data of the loan application business;

[0067] S502, upload the business data of the loan application business to the blockchain network.

[0068] Storing the business data on the blockchain network can facilitate the traceability of the business data.

[0069] In one embodiment, as Figure 6 shown, before obtaining the key loan element information submitted by the loan customer according to the loan application request, the blockchain-based loan risk prediction method provided in the embodiments of the present invention further includes the following steps:

[0070] S601, receive the loan application request sent by the loan customer through 5G messaging.

[0071] In the embodiments of the present invention, sending the loan application request through 5G messaging enables the customer to handle the loan application business through 5G messaging, avoiding the trouble of downloading the client, and taking advantage of the high data transmission rate of 5G messaging to greatly improve the real-time performance of business handling.

[0072] In order to enable the loan customer to quickly understand their risk prediction results, in one embodiment, as Figure 6 shown, after performing risk prediction on the loan customer based on one or more found risk prediction models to obtain the risk prediction results of the loan customer, the blockchain-based loan risk prediction method provided in the embodiments of the present invention further includes the following steps:

[0073] S602, determine the loan amount information of the loan customer according to the risk prediction results of the loan customer;

[0074] S603, send the loan amount information of the loan customer to the loan customer through 5G messaging.

[0075] In one embodiment, as Figure 7As shown in the figure, the blockchain-based loan risk prediction method provided in the embodiments of the present invention may further include the following steps:

[0076] S701, obtaining the loan risk data corresponding to each type of loan stored on the blockchain network;

[0077] S702, using the smart contracts pre-stored on the blockchain network to statistically calculate the probability of each type of loan risk corresponding to each type of loan;

[0078] S703, associatively storing the loan risk types with a probability higher than the preset threshold and the corresponding loan types into the blockchain network;

[0079] S704, querying the corresponding loan risk types from the blockchain network according to the loan type of the loan customer and adding them to the loan risk type set of the loan customer.

[0080] It should be noted that the loan types in the embodiments of the present invention include but are not limited to: personal micro-loans, student loans, housing loans, auto loans, and corporate business loans; through the above embodiments, the loan risk types with higher probabilities can be added to the loan risk type set of the loan customer.

[0081] In one embodiment, as Figure 8 shown in the figure, the blockchain-based loan risk prediction method provided in the embodiments of the present invention may further include the following steps:

[0082] S801, obtaining the historical loan risk data of each loan institution;

[0083] S802, training different machine learning models according to the historical loan risk data of each loan institution to obtain multiple risk prediction models corresponding to each loan institution.

[0084] In the embodiments of the present invention, using the historical loan risk data of each loan institution to train the risk prediction model can shorten the time-consuming of the model training process, and training different machine learning models can obtain more risk prediction models.

[0085] Since the risk prediction models of each loan institution are shared, in order to prevent the model parameters from being tampered with, further, in one embodiment, as Figure 8 shown in the figure, the blockchain-based loan risk prediction method provided in the embodiments of the present invention may further include the following steps:

[0086] S803, uploading the multiple risk prediction models corresponding to each loan institution to the blockchain network.

[0087] Based on the same inventive concept, an embodiment of the present invention also provides a blockchain-based loan risk prediction device, as described in the following embodiments. Since the principle of the device for solving problems is similar to that of the blockchain-based loan risk prediction method, the implementation of the device can refer to the implementation of the blockchain-based loan risk prediction method, and the repeated parts will not be elaborated.

[0088] Figure 9 FIG. is a schematic diagram of a blockchain-based loan risk prediction device provided in an embodiment of the present invention, as Figure 9 shown, the device includes: a loan application module 901, a key loan element information determination module 902, a key loan element evaluation result acquisition module 903, a customer loan risk type determination module 904, a risk prediction model selection module 905, and a risk prediction module 906.

[0089] Among them, the loan application module 901 is used to receive a loan application request from a loan customer;

[0090] The key loan element information determination module 902 is used to obtain the key loan element information submitted by the loan customer according to the loan application request;

[0091] The key loan element evaluation result acquisition module 903 is used to compare the key loan element information with the customer information of known loan risk customers, and determine the key loan element evaluation result corresponding to the key loan element information according to the comparison result, wherein the customer information of the known loan risk customers includes: the key loan element information and key loan element evaluation results of the known loan risk customers;

[0092] The customer loan risk type determination module 904 is used to determine the loan risk type set of the loan customer according to the key loan element evaluation result, wherein the loan risk type set includes one or more loan risk types of the loan customer;

[0093] The risk prediction model selection module 905 is used to find the risk prediction model corresponding to the loan risk type from the risk prediction model set according to the loan risk type set of the loan customer, wherein the risk prediction model set includes: multiple risk prediction models uploaded by each lending institution through the blockchain network;

[0094] The risk prediction module 906 is used to perform risk prediction on the loan customer based on the found one or more risk prediction models to obtain the risk prediction result of the loan customer.

[0095] In one embodiment, as Figure 10As shown in the figure, in the loan risk prediction device based on blockchain provided in the embodiment of the present invention, the risk prediction module 906 includes: a customer information acquisition unit 9061, a risk prediction unit 9062, and a risk prediction result integration unit 9063.

[0096] Among them, the customer information acquisition unit 9061 is used to acquire the customer information of the loan customer; the risk prediction unit 9062 is used to input the customer information of the loan customer into each found risk prediction model to obtain the risk prediction results of each risk prediction model; the risk prediction result integration unit 9063 is used to generate the risk prediction result of the loan customer according to the risk prediction results of each risk prediction model.

[0097] In this embodiment, the risk prediction result integration unit 9063 is further used to: acquire the performance evaluation results of each risk prediction model; determine the weight coefficients of each risk prediction model according to the performance evaluation results of each risk prediction model; and generate the risk prediction result of the loan customer according to the weight coefficients and risk prediction results of each risk prediction model.

[0098] Optionally, the risk prediction result integration unit 9063 is further used to: select the target integration algorithm model to be adopted by each lending institution from multiple pre-stored integration algorithm models on the blockchain network by using the smart contract pre-stored on the blockchain network; use the selected target integration algorithm model to integrate the risk prediction results of each risk prediction model to obtain the risk prediction result of the loan customer; and push the risk prediction result of the loan customer to each lending institution.

[0099] In one embodiment, as Figure 10 shown, the loan risk prediction device based on blockchain provided in the embodiment of the present invention further includes: a risk prediction result on-chain module 907, which is used to upload the risk prediction result of the loan customer to the blockchain network; a loan application service handling module 908, which is used to execute the loan application service of the loan customer based on the risk prediction result of the loan customer stored on the blockchain network, where when the risk prediction result corresponding to any risk type of the loan customer exceeds the preset risk condition, the loan application service of the loan customer is refused to be executed.

[0100] In one embodiment, as Figure 10 shown, the loan risk prediction device based on blockchain provided in the embodiment of the present invention further includes: a service data on-chain module 909, which is used to acquire the service data of the loan application service; and upload the service data of the loan application service to the blockchain network.

[0101] In one embodiment, as Figure 10As shown in the figure, the blockchain-based loan risk prediction device provided in the embodiment of the present invention further includes: a 5G message communication module 910, configured to receive a loan application request sent by a loan customer through a 5G message; and determine the loan amount information of the loan customer according to the risk prediction result of the loan customer, and send the loan amount information of the loan customer to the loan customer through a 5G message.

[0102] In one embodiment, as Figure 10 As shown in the figure, the blockchain-based loan risk prediction device provided in the embodiment of the present invention further includes: a similar customer determination module 911, configured to determine similar customers of a loan customer based on a pre-constructed customer relationship knowledge graph; in this embodiment, the customer loan risk type determination module 904 is further configured to add the loan risk types of the similar customers to the loan risk type set of the loan customer.

[0103] In one embodiment, the customer loan risk type determination module 904 is further configured to: obtain the loan risk data corresponding to each loan type stored on the blockchain network; use the smart contract pre-stored on the blockchain network to count the probability of the loan risk type corresponding to each loan type; associate and store the loan risk type with a probability higher than a preset threshold and the corresponding loan type on the blockchain network; query the corresponding loan risk type from the blockchain network according to the loan type of the loan customer, and add it to the loan risk type set of the loan customer.

[0104] In one embodiment, as Figure 10 As shown in the figure, the blockchain-based loan risk prediction device provided in the embodiment of the present invention further includes: a machine learning module 912, configured to obtain the historical loan risk data of each loan institution; and train different machine learning models according to the historical loan risk data of each loan institution to obtain multiple risk prediction models corresponding to each loan institution.

[0105] In one embodiment, as Figure 10 As shown in the figure, the blockchain-based loan risk prediction device provided in the embodiment of the present invention further includes: a risk prediction model uploading module 913, configured to upload multiple risk prediction models corresponding to each loan institution to the blockchain network.

[0106] Based on the same inventive concept, the embodiment of the present invention further provides a computer device to solve the technical problems that the existing loan risk prediction method using artificial intelligence technology has a long training process and low model accuracy. Figure 11 It is a schematic diagram of a computer device provided in the embodiment of the present invention, as Figure 11As shown in the figure, the computer device 11 includes a memory 111, a processor 112, and a computer program stored on the memory 111 and executable on the processor 112. When the processor 112 executes the computer program, the above-mentioned blockchain-based loan risk prediction method is implemented.

[0107] Based on the same inventive concept, an embodiment of the present invention also provides a computer-readable storage medium to solve the technical problems that the existing loan risk prediction method using artificial intelligence technology has a long training process and a low model accuracy. The computer-readable storage medium stores a computer program for executing the above-mentioned blockchain-based loan risk prediction method.

[0108] In summary, in the embodiment of the present invention, the blockchain-based loan risk prediction method, device, computer device, and computer-readable storage medium receive risk prediction models uploaded by each lending institution through a blockchain network to form a risk prediction model set. After receiving a loan application request from a loan customer, according to the loan application request of the loan customer, determine the loan risk type set of the loan customer, and then, according to each loan risk type of the loan customer in the loan risk type set, search for the risk prediction model corresponding to the loan risk type from the risk prediction model set. Finally, based on the one or more risk prediction models found, perform a risk prediction on the loan customer to obtain the risk prediction result of the loan customer.

[0109] Compared with the prior art technical solution of directly training a single model to predict the risk of loan customers, in the embodiment of the present invention, each lending institution trains multiple risk prediction models respectively, and selects the risk prediction model corresponding to the loan risk type of the current loan application request of the loan customer from the multiple risk prediction models to predict the loan risk of the loan customer, which can not only shorten the time-consuming of the model training process, but also improve the accuracy of the overall model prediction.

[0110] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0111] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0112] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0113] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0114] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A loan risk prediction method based on blockchain, characterized in that, it includes: Receiving a loan application request from a loan customer; According to the loan application request, obtaining the key loan element information submitted by the loan customer; Comparing the key loan element information with the customer information of known loan risk customers, and determining the key loan element evaluation result corresponding to the key loan element information according to the comparison result. Among them, the customer information of the known loan risk customers includes: the key loan element information and the key loan element evaluation result of the known loan risk customers; According to the key loan element evaluation result, determining the loan risk type set of the loan customer, where the loan risk type set includes one or more loan risk types of the loan customer; According to the loan risk type set of the loan customer, searching for the risk prediction model corresponding to the loan risk type from the risk prediction model set. Among them, the risk prediction model set includes: multiple risk prediction models uploaded by each lending institution through the blockchain network; Based on the one or more risk prediction models found, performing risk prediction on the loan customer to obtain the risk prediction result of the loan customer, which includes: obtaining the customer information of the loan customer; inputting the customer information of the loan customer into each found risk prediction model to obtain the risk prediction results of each risk prediction model; generating the risk prediction result of the loan customer according to the risk prediction results of each risk prediction model; Generating the risk prediction result of the loan customer according to the risk prediction results of each risk prediction model, including: using the smart contract pre-stored on the blockchain network to select the target integration algorithm model to be adopted by each lending institution from the multiple integration algorithm models pre-stored on the blockchain network; using the selected target integration algorithm model to integrate the risk prediction results of each risk prediction model to obtain the risk prediction result of the loan customer; pushing the risk prediction result of the loan customer to each lending institution.

2. The method according to claim 1, characterized in that, after performing risk prediction on the loan customer based on the one or more risk prediction models found to obtain the risk prediction result of the loan customer, the method further includes: Uploading the risk prediction result of the loan customer to the blockchain network; Based on the risk prediction result of the loan customer stored on the blockchain network, executing the loan application business of the loan customer. Among them, when the risk prediction result corresponding to any risk type of the loan customer exceeds the preset risk condition, the loan application business of the loan customer is rejected.

3. The method according to claim 2, characterized in that, after executing the loan application business of the loan customer based on the risk prediction result of the loan customer stored on the blockchain network, the method further includes: Obtaining the business data of the loan application business; Uploading the business data of the loan application business to the blockchain network.

4. The method according to claim 1, wherein, receiving a loan application request from a loan customer, including: receiving a loan application request sent by the loan customer via 5G messaging; wherein, after performing a risk prediction on the loan customer based on one or more found risk prediction models to obtain a risk prediction result of the loan customer, the method further includes: determining loan amount information of the loan customer according to the risk prediction result of the loan customer; sending the loan amount information of the loan customer to the loan customer via 5G messaging.

5. The method according to claim 1, wherein, after finding a risk prediction model corresponding to the loan risk type from the risk prediction model set according to the loan risk type set of the loan customer, the method further includes: determining similar customers of the loan customer based on a pre-constructed customer relationship knowledge graph; adding the loan risk types of the similar customers to the loan risk type set of the loan customer.

6. The method according to claim 1, wherein, after finding a risk prediction model corresponding to the loan risk type from the risk prediction model set according to the loan risk type set of the loan customer, the method further includes: obtaining loan risk data corresponding to each loan type stored on the blockchain network; using a smart contract pre-stored on the blockchain network to statistically calculate the probability of the loan risk type corresponding to each loan type; associatively storing the loan risk types with a probability higher than a preset threshold and the corresponding loan types on the blockchain network; querying the corresponding loan risk types from the blockchain network according to the loan type of the loan customer and adding them to the loan risk type set of the loan customer.

7. The method according to any one of claims 1 to 6, wherein, the method further includes: obtaining historical loan risk data of each loan institution; training different machine learning models according to the historical loan risk data of each loan institution to obtain multiple risk prediction models corresponding to each loan institution; uploading the multiple risk prediction models corresponding to each loan institution to the blockchain network.

8. A blockchain-based loan risk prediction device, wherein, including: a loan application module for receiving a loan application request from a loan customer; a key loan element information determination module for obtaining key loan element information submitted by the loan customer according to the loan application request; a key loan element evaluation result acquisition module for comparing the key loan element information with the customer information of known loan risk customers and determining a key loan element evaluation result corresponding to the key loan element information according to the comparison result, wherein the customer information of the known loan risk customers includes: key loan element information and key loan element evaluation results of the known loan risk customers; A customer loan risk type determination module, configured to determine a set of loan risk types of the loan customer according to the evaluation result of the key loan elements, where the set of loan risk types includes one or more loan risk types of the loan customer; A risk prediction model selection module, configured to find a risk prediction model corresponding to the loan risk type from a set of risk prediction models according to the set of loan risk types of the loan customer, where the set of risk prediction models includes: multiple risk prediction models uploaded by each lending institution through the blockchain network; A risk prediction module, configured to perform risk prediction on the loan customer based on one or more found risk prediction models to obtain a risk prediction result of the loan customer; The risk prediction module includes: a customer information acquisition unit, configured to acquire customer information of the loan customer; a risk prediction unit, configured to input the customer information of the loan customer into each found risk prediction model to obtain a risk prediction result of each risk prediction model; a risk prediction result integration unit, configured to generate a risk prediction result of the loan customer according to the risk prediction results of each risk prediction model; Generating the risk prediction result of the loan customer according to the risk prediction results of each risk prediction model includes: using a smart contract pre-stored on the blockchain network to select a target integration algorithm model to be adopted by each lending institution from multiple integration algorithm models pre-stored on the blockchain network; using the selected target integration algorithm model to integrate the risk prediction results of each risk prediction model to obtain a risk prediction result of the loan customer; pushing the risk prediction result of the loan customer to each lending institution.

9. The apparatus according to claim 8, wherein, the apparatus further includes: A risk prediction result on-chain module, configured to upload the risk prediction result of the loan customer to the blockchain network; A loan application service handling module, configured to execute the loan application service of the loan customer based on the risk prediction result of the loan customer stored on the blockchain network, where when the risk prediction result corresponding to any risk type of the loan customer exceeds a preset risk condition, the loan application service of the loan customer is refused to be executed.

10. The apparatus according to claim 9, wherein, after executing the loan application service of the loan customer based on the risk prediction result of the loan customer stored on the blockchain network, the apparatus further includes: Acquiring service data of the loan application service; Uploading the service data of the loan application service to the blockchain network.

11. The apparatus according to claim 8, wherein, the apparatus further includes: A 5G message communication module, configured to receive a loan application request sent by a loan customer through 5G message; and determine loan amount information of the loan customer according to the risk prediction result of the loan customer, and send the loan amount information of the loan customer to the loan customer through 5G message.

12. The apparatus according to claim 8, It is characterized in that the device further comprises: a similar customer determination module, configured to determine similar customers of the loan customer based on a pre-constructed customer relationship knowledge graph; wherein, the customer loan risk type determination module is further configured to add the loan risk types of the similar customers to the loan risk type set of the loan customer.

13. The device according to claim 8, it is characterized in that the customer loan risk type determination module is further configured to: obtain the loan risk data corresponding to each loan type stored on the blockchain network; utilize the smart contract pre-stored on the blockchain network to statistically calculate the probability of the loan risk type corresponding to each loan type; associate and store the loan risk types with a probability higher than a preset threshold and the corresponding loan types on the blockchain network; query the corresponding loan risk type from the blockchain network according to the loan type of the loan customer, and add it to the loan risk type set of the loan customer.

14. The device according to any one of claims 8 to 13, it is characterized in that the device further comprises: a machine learning module, configured to obtain the historical loan risk data of each lending institution; and train different machine learning models according to the historical loan risk data of each lending institution to obtain multiple risk prediction models corresponding to each lending institution; a risk prediction model uploading module, configured to upload the multiple risk prediction models corresponding to each lending institution to the blockchain network.

15. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, it is characterized in that when the processor executes the computer program, the blockchain-based loan risk prediction method according to any one of claims 1 to 7 is implemented.

16. A computer-readable storage medium, it is characterized in that the computer-readable storage medium stores a computer program for executing the blockchain-based loan risk prediction method according to any one of claims 1 to 7.

Images