A method and system for enterprise risk prediction based on federated learning

Through a method based on federated learning, combined with enterprise operating data and third-party data, a joint model is built to conduct risk prediction, which solves the problems of low analysis accuracy and efficiency in the existing technology, and achieves the accuracy and efficiency improvement of enterprise risk prediction.

CN114037359BActive Publication Date: 2024-12-06XIAMEN MEIYA YIAN INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111547710.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-16
Publication Date
2024-12-06
Estimated Expiration
2041-12-16

AI Technical Summary

Technical Problem

Existing enterprise risk prediction relies on subjective experience and manual integration of data, resulting in insufficient analysis accuracy and low efficiency, and inability to effectively predict enterprise risks.

Method used

Using a federated learning-based method, enterprise business data and third-party data are aligned through homomorphic encryption and business data, and a joint model is used to build a joint model for risk prediction and generate risk scores.

Benefits of technology

It improves computing and interaction efficiency, ensures data security, ensures the accuracy of prediction results and the effectiveness of enterprise risk prediction, and clarifies the company's business vulnerabilities and future risks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114037359B_ABST
    Figure CN114037359B_ABST
Patent Text Reader

Abstract

Disclosed is a method and system for enterprise risk prediction based on federated learning, including submitting enterprise business data and third-party data to a federated repository, wherein the third-party data includes consulting and auditing data; processing the enterprise business data and third-party data in the federated repository based on preset data specifications, and the processing includes homomorphic encryption and business data alignment; using machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, sampling from each model to obtain sample data, and inputting the sample data into a federated learning technology platform for joint modeling to generate a joint model; using the joint model for joint prediction, and feeding back the prediction results to the enterprise and third-party algorithm models. This application can be used to clarify the business vulnerabilities of current enterprises and predict the risks in future enterprises.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the application field of enterprise risk prediction, and in particular to an enterprise risk prediction method and system based on federated learning. Background Art

[0002] At present, risk prediction in the process of enterprise operation is mainly based on the experience of experts or subjective judgment of enterprise management based on previous work experience and market environment and other factors. The accuracy of analysis mainly depends on the personal experience and ability of analysts, with large subjective factors, inconsistent standards, and heavy manual screening tasks, which leads to insufficient market insight and enterprise risk prediction capabilities.

[0003] Moreover, due to the lack of unified standards for data involving third parties (such as audits), existing risk forecasts are huge in data volume, and require a large workload for manual data integration, analysis and summary, and conclusion drawing. Manual screening tasks are arduous and inefficient, and it is impossible to comprehensively screen information, which greatly reduces the accuracy of the prediction results and makes it difficult to help enterprises. Summary of the invention

[0004] In order to solve various technical problems existing in enterprise risk prediction in the prior art, the present invention proposes an enterprise risk prediction method and system based on federated learning to solve the above technical problems.

[0005] According to one aspect of the present invention, a method for enterprise risk prediction based on federated learning is proposed, comprising:

[0006] S1: Submit enterprise operating data and third-party data to the federal repository, where third-party data includes consulting and audit data;

[0007] S2: Processing of free samples of enterprise operating data and third-party data in the federated repository based on preset data specifications, including homomorphic encryption and business data alignment;

[0008] S3: Use machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, sample data from each model, and input the sample data into the federated learning technology platform for joint modeling to generate a joint model; and

[0009] S4: Use the joint model to make joint predictions and feed the prediction results back to the enterprise and third-party algorithm models.

[0010] In some specific embodiments, in step S2, the business operation data and third-party data are homomorphically encrypted and then stored in a federal repository, and all data in the entire risk prediction process are homomorphically encrypted ciphertexts.

[0011] In some specific embodiments, the business data alignment in step S2 includes quantifying and segmenting the enterprise and third-party data.

[0012] In some specific embodiments, the enterprise and third-party algorithm models obtain corresponding classification data, feature data and core business data, and the core business data includes sample ID, business feature variables and performance data.

[0013] In some specific embodiments, the sample data includes training parameters and gradients of enterprise and third-party algorithm models.

[0014] In some specific embodiments, the joint modeling in step S3 specifically includes: matching sample IDs of input enterprise and third-party data samples, training the model after outputting the intersection of sample IDs, and generating a joint model.

[0015] In some specific embodiments, step S4 also includes: the enterprise and third-party algorithm models are updated and trained according to the parameters fed back by the joint model.

[0016] In some specific embodiments, the risk model formula in the joint model is Ris=min(150,F1)+min(150,F2)+min(150,F3)+min(150,F4)+min(150,F5), where Ris is the risk score, F1 represents the legal person shareholder situation, F2 represents the operating situation, F3 represents the credit situation, F4 represents the competitiveness, and F5 represents the industry trend.

[0017] According to a second aspect of the present invention, a computer-readable storage medium stores one or more computer programs, which implement the above method when executed by a computer processor.

[0018] According to a second aspect of the present invention, a corporate risk prediction system based on federated learning is proposed, the system comprising:

[0019] Data storage unit: configured to submit enterprise operation data and third-party data to the federated repository, wherein the third-party data includes consulting and auditing data;

[0020] Data processing unit: configured to process free samples of enterprise operating data and third-party data in the federated repository based on preset data specifications. The processing includes homomorphic encryption and business data alignment.

[0021] Joint model building unit: configured to use machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, sample data from each model, and input the sample data into the federated learning technology platform for joint modeling to generate a joint model;

[0022] Prediction unit: configured to use the joint model for joint prediction and feed the prediction results back to the enterprise and third-party algorithm models.

[0023] The enterprise risk prediction method and system based on federated learning of the present invention adopts a balanced "data security-computational efficiency-interaction efficiency" technical framework that meets application requirements. Under the premise of ensuring enterprise data privacy and enterprise data security, the computing and interaction efficiency are improved, and the performance of the model trained by federated learning is guaranteed to be the same as that of the model obtained by centralized training. By adjusting the structure of the federated learning model, the enterprise operating data and third-party data are sampled and trained separately, the two training models are periodically matched and a joint model is generated, and the trained joint model is input into the joint risk prediction model. Utilize the current business activity data of the enterprise, combined with third-party (such as audit) data for in-depth learning and prediction, clarify the current business loopholes of the enterprise, and predict the risks of the enterprise in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated into and constitute a part of this specification. The accompanying drawings illustrate the embodiments and are used together with the description to explain the principles of the present invention. It will be easy to recognize other embodiments and many expected advantages of the embodiments because they become better understood by reference to the following detailed description. The elements of the drawings are not necessarily to scale with each other. The same reference numerals refer to corresponding similar parts.

[0025] Figure 1 is a flowchart of a method for enterprise risk prediction based on federated learning according to an embodiment of the present application;

[0026] Figure 2 It is a flowchart of a method for enterprise risk prediction based on federated learning in a specific embodiment of the present application;

[0027] Figure 3 This is a framework diagram of an enterprise risk prediction system based on federated learning in one embodiment of the present application;

[0028] Figure 4 It is a structural diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application. DETAILED DESCRIPTION

[0029] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0030] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0031] Figure 1 A flowchart of a method for predicting enterprise risk based on federated learning according to an embodiment of the present application is shown as follows: Figure 1 As shown, the method comprises the following steps:

[0032] S101: Submit enterprise operating data and third-party data to the federal repository, where the third-party data includes consulting and auditing data.

[0033] S102: Process the enterprise business data and third-party data in the federated repository based on the preset data specifications. The processing includes homomorphic encryption and business data alignment. The enterprise and the third-party data provider process the free samples in the federated repository according to the specified data specifications. Each sample corresponds to a unique sample id, including homomorphic encryption in federated learning and alignment of the corresponding business data.

[0034] In a specific embodiment, in addition to realizing basic encryption operations, homomorphic encryption can also realize a variety of calculation functions between ciphertexts, that is, calculation first and then decryption is equivalent to decryption first and then calculation. In order to ensure the data security of enterprise and third-party data, both enterprise operating data and third-party data are homomorphically encrypted and saved in a federal repository after encryption. During the entire joint risk prediction process, all data are encrypted ciphertexts. After the results are returned, the recipients of the model result data will calculate and decrypt the data obtained by each party. Business data alignment quantifies and divides the data of enterprises and third parties in accordance with the optimal efficiency of model execution, such as: enterprise sales data is divided by month, and enterprise operation data is divided by quarter.

[0035] S103: Use machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, sample sample data from each model, and input the sample data into the federated learning technology platform for joint modeling to generate a joint model. Enterprises and third-party data providers use machine learning and deep learning model training to build relevant algorithm models to obtain classification, features, and core business data (including sample id, business feature variables, performance data, etc.). Sample data is obtained from the model, and the sample data includes the training parameters and gradients of the enterprise and third-party models. The sample data is input into the federated learning technology platform for joint modeling.

[0036] In a specific embodiment, during joint modeling, sample ID matching is performed on the input enterprise and third-party data samples, the ID intersection is output, and then the model is trained to generate a joint model. The joint model is a joint model after the enterprise and third-party models are integrated on the federated learning platform.

[0037] S104: Use the joint model to make joint predictions, and feed back the prediction results to the enterprise and third-party algorithm models. The enterprise and third-party's own local models are updated and trained based on the parameters fed back by the joint model.

[0038] Continue to refer Figure 2 , Figure 2 A flowchart of a method for predicting enterprise risk based on federated learning according to a specific embodiment of the present application is shown as follows: Figure 2 As shown, the business data of an enterprise includes the business situation, main business income, sales situation, employee data, etc., and the third-party data includes capital structure, total tax payment, competitiveness, number of patents, etc. The above two types of data are submitted to the federal repository respectively, and the classification / feature / business data are extracted respectively, and the sample data is obtained by sampling. The sample data is input into the federal risk model, sample matching is performed, the sample intersection is output, and the model is trained. Finally, a joint risk prediction model including the enterprise model and the third-party model is generated to perform full risk prediction, and the prediction results are fed back to the enterprise or third-party federal data repository respectively, so that the free local models of the enterprise and the third party can be updated and trained according to the feedback parameters of the joint model. Using the federated learning model, combined with the enterprise customized risk calculation algorithm model, using the enterprise's current business activity data, combined with third-party (such as audit) data for in-depth learning and prediction, clarify the current business loopholes of the enterprise, and predict the risks in the enterprise in the future.

[0039] In a specific embodiment, the risk prediction model uses the following model formula to predict the risk score Ris: Ris = min(150, F1) + min(150, F2) + min(150, F3) + min(150, F4) + min(150, F5), where Ris is the risk score, F1 represents the legal person shareholder situation, F2 represents the operating situation, F3 represents the credit situation, F4 represents the competitiveness, and F5 represents the industry trend.

[0040] In a specific embodiment, the evaluation index of the legal person shareholder situation is specifically the participating enterprises, and the participating enterprises include core enterprises and non-core enterprises. The specific definition rules are as follows in Table 1:

[0041] Table 1. Legal person shareholders

[0042]

[0043] Definition: root node R1, child node R2, industry ranking T1, blacklist B1. For enterprises in non-enterprise forests: F1 = 120; for enterprises in enterprise forests:

[0044] F1 = 120 - (R1 - R2) / 5 * 10;

[0045] F1=F1-10,T1≤10%;

[0046] F1=F1-5,10% <T1≤50%;

[0047] F1=F1-5,B1≥10;

[0048] F1=F1-10,B1≥100;

[0049] Among them, deductions are made based on the distance from the root, and ten points will be deducted for more than 5 layers; for the average registered capital ranking, ten points will be deducted if the tree is in the bottom 10% of the industry ranking, and 5 points will be deducted if 10% < ranking < 50%; if the tree is involved in the blacklist, 5 points will be deducted for more than 10 times in three years, and ten points will be deducted for more than 100 times.

[0050] In a specific embodiment, the evaluation index of the operating condition F2 is as follows: in, Z4=(W10-W11) / W11, W10, W11≠0; the names corresponding to the above variable symbols are as shown in Table 2:

[0051] Table 2. Variable definition table

[0052]

[0053]

[0054] In a specific embodiment, the credit status F3 has different calculation methods depending on whether the enterprise belongs to the enterprise forest: For an enterprise not in the enterprise forest:

[0055] in, Oap: represents the original administrative penalty score, Ode represents the original number of times the enterprise was sued, Es represents the scale index, and Rdh represents the percentile ranking of the dishonest execution.

[0056] For enterprises in the enterprise forest: F3 = 150 + 10a + 10b + 10c, where

[0057]

[0058]

[0059]

[0060] In a specific embodiment, the evaluation index calculation formula of the competitiveness F4 is:

[0061] Among them: Rt represents the percentile ranking of trademarks; Rp represents the percentile ranking of patents; Rw represents the percentile ranking of work copyrights; Rs represents the percentile ranking of software copyrights; Re represents the percentile ranking of registered capital; Rc represents the percentile ranking of qualification certificate scores; Rm represents the percentile ranking of employees' basic medical insurance.

[0062] In a specific embodiment, the evaluation index calculation formula of the industry trend F5 is: F5=120+6a+4b+6c; where:

[0063] Riu represents the operating income growth rate of large and medium-sized enterprises representative of the industry;

[0064] Reu represents the annual number of new enterprises in the industry;

[0065] Rpu represents the growth rate of the number of employees in the industry.

[0066] Finally, the formula Ris=min(150,F1)+min(150,F2)+min(150,F3)+min(150,F4)+min(150,F5) is used, that is, each evaluation indicator takes the sum of 150 and the minimum calculated value to obtain the final risk score of the risk model.

[0067] Continue to refer Figure 3 , Figure 3 A framework diagram of an enterprise risk prediction system based on federated learning according to an embodiment of the present application is shown. Figure 3As shown, the system includes a data storage unit 301, a data processing unit 302, a joint model building unit 303 and a prediction unit 304. The data storage unit 301 is configured to submit enterprise operation data and third-party data to the federated repository, wherein the third-party data includes consulting and auditing data; the data processing unit 302 is configured to process the enterprise operation data and third-party data in the federated repository based on the preset data specification, and the processing includes homomorphic encryption and business data alignment; the joint model building unit 303 is configured to use machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, sample sample data from each model, and input the sample data into the federated learning technology platform for joint modeling to generate a joint model; the prediction unit 304 is configured to use the joint model for joint prediction, and feed the prediction results back to the enterprise and third-party algorithm models.

[0068] Reference below Figure 4 , which shows a schematic diagram of the structure of a computer system 400 suitable for implementing an electronic device of an embodiment of the present application. Figure 4 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0069] like Figure 4 As shown, the computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage part 408 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the system 400 are also stored. The CPU 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0070] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, etc.; an output section 407 including a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed, so that a computer program read therefrom is installed into the storage section 408 as needed.

[0071] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 409, and / or installed from the removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, the above functions defined in the method of the present application are executed. It should be noted that the computer-readable storage medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, - but not limited to - a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wireless, wireline, optical cable, RF, etc., or any suitable combination of the foregoing.

[0072] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0073] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0074] The modules involved in the embodiments described in the present application may be implemented by software or hardware. The units described may also be set in a processor, for example, it may be described as: a processor includes a deployment unit, an instruction processing unit, and a file access unit. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0075] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or it may exist independently without being assembled into the electronic device. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed by the electronic device, the electronic device: submits the business data and third-party data to the federal repository, wherein the third-party data includes consulting and auditing data; performs free sample processing on the business data and third-party data in the federal repository based on the preset data specification, and the processing includes homomorphic encryption and business data alignment; uses machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, samples from each model to obtain sample data, and inputs the sample data into the federated learning technology platform for joint modeling to generate a joint model; uses the joint model for joint prediction, and feeds the prediction results back to the enterprise and third-party algorithm models.

[0076] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above invention concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. A method for enterprise risk prediction based on federated learning, characterized in that: include: S1: Submitting enterprise operating data and third-party data to the federal repository, wherein the third-party data includes consulting and audit data; S2: performing free sample processing on the enterprise operation data and the third-party data in the federated repository based on a preset data specification, wherein the processing includes homomorphic encryption and business data alignment; S3: Use machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, sample data from each model, and input the sample data into the federated learning technology platform for joint modeling to generate a joint model; and S4: Perform joint prediction using the joint model, and feed the prediction results back to the enterprise and the third-party algorithm model; The enterprise and third-party algorithm models obtain corresponding classification data, feature data and core business data, and the core business data includes sample ID, business feature variables and performance data; The joint modeling in step S3 specifically includes: performing sample ID matching on the input data samples of the enterprise and the third party, outputting the intersection of the sample IDs, training the model, and generating the joint model; The risk model formula in the joint model is Ris=min(150,F1)+min(150,F2)+min(150,F3)+min(150,F4)+min(150,F5), where Ris is the risk score, F1 represents the legal person shareholder situation, F2 represents the operating situation, F3 represents the credit situation, F4 represents the competitiveness, and F5 represents the industry trend; Define the root node R1, child node R2, industry ranking T1, blacklist B1, for enterprises in non-enterprise forest: F1 = 120; for enterprises in enterprise forest: F1 = 120 - (R1 - R2) / 5 * 10; F1=F1-10,T1≤10%; F1=F1-5,10% <T1≤50%; F1=F1-5,B1≥10; F1=F1-10,B1≥100; Among them, deductions are made for the distance from the root, and ten points will be deducted for more than 5 layers; for the average registered capital ranking, if the tree is ranked in the bottom 10% of the industry, ten points will be deducted, and if it is 10% < ranking < 50%, five points will be deducted; if the tree is involved in the blacklist for more than 10 times in three years, five points will be deducted, and if it is more than 100 times, ten points will be deducted; The evaluation indicators of operating conditions F2 are as follows: Among them, Z1 represents the average growth rate of tax payment, Z2 represents the growth rate of the number of employees covered by basic medical insurance, Z3 represents the return on net assets in the past two years, Z4 represents the growth rate of sales revenue, W22 represents the percentile ranking of the company's existence time, W23 represents the percentile ranking of the debt ratio, and W24 represents the percentile ranking of the number of valid administrative licenses; Credit status F3 has different calculation methods depending on whether the enterprise belongs to the enterprise forest: For enterprises not in the enterprise forest: in, Oap: represents the original administrative penalty score, Ode represents the original number of times the enterprise was sued, Es represents the scale index, and Rdh represents the percentile ranking of dishonesty enforcement; For enterprises in the enterprise forest: F3 = 150 + 10a + 10b + 10c, where Ap5 represents the administrative penalty score within five years; De3 indicates the number of times the enterprise was sued in three years; Bl3 indicates the number of blacklistings within three years; The evaluation index calculation formula of competitiveness F4 is: in: Rt represents the percentile ranking of trademarks; Rp represents the percentile ranking of patents; Rw represents the percentile ranking of copyrights of works; Rs represents the percentile ranking of software copyrights; Re represents the percentile ranking of registered capital; Rc represents the percentile ranking of qualification certificate scores; Rm represents the percentile ranking of employees’ basic medical insurance; The calculation formula for the evaluation index of industry trend F5 is: F5 = 120 + 6A + 4B + 6C; where: Riu represents the operating income growth rate of large and medium-sized enterprises that are representative of the industry; Reu represents the annual number of new enterprises in the industry; Rpu represents the growth rate of the number of employees in the industry.

2. The enterprise risk prediction method based on federated learning according to claim 1 is characterized in that: In the step S2, the business operation data and the third-party data are homomorphically encrypted and then stored in the federal repository, and all data in the entire risk prediction process are homomorphically encrypted ciphertexts.

3. The enterprise risk prediction method based on federated learning according to claim 1 is characterized in that: The business data alignment in step S2 includes quantifying and segmenting the data of the enterprise and the third party.

4. The enterprise risk prediction method based on federated learning according to claim 1 is characterized in that: The sample data includes the training parameters and gradients of the enterprise and third-party algorithm models.

5. The enterprise risk prediction method based on federated learning according to claim 1 is characterized in that: After step S4, the following further includes: the enterprise and third-party algorithm models are updated and trained according to the parameters fed back by the joint model.

6. A computer-readable storage medium having one or more computer programs stored thereon, characterized in that: When the one or more computer programs are executed by a computer processor, the method according to any one of claims 1 to 5 is implemented.

7. An enterprise risk prediction system based on federated learning, characterized in that: The system comprises: A data storage unit: configured to submit enterprise operation data and third-party data to a federal repository, wherein the third-party data includes consulting and auditing data; Data processing unit: configured to process the enterprise operation data and the third-party data in the federated repository based on a preset data specification, wherein the processing includes homomorphic encryption and business data alignment; Joint model building unit: configured to use machine learning and deep learning model training to build enterprise and third-party algorithm models respectively, sample data from each model, and input the sample data into the federated learning technology platform for joint modeling to generate a joint model; Prediction unit: configured to use the joint model to perform joint prediction and feed back the prediction results to the enterprise and the third-party algorithm model; The enterprise and third-party algorithm models obtain corresponding classification data, feature data and core business data, and the core business data includes sample ID, business feature variables and performance data; The joint modeling in the joint model building unit specifically includes: matching sample IDs of the input data samples of the enterprise and the third party, outputting the intersection of the sample IDs, training the model, and generating the joint model; The risk model formula in the joint model is Ris=min(150,F1)+min(150,F2)+min(150,F3)+min(150,F4)+min(150,F5), where Ris is the risk score, F1 represents the legal person shareholder situation, F2 represents the operating situation, F3 represents the credit situation, F4 represents the competitiveness, and F5 represents the industry trend; Define the root node R1, child node R2, industry ranking T1, blacklist B1, for enterprises in non-enterprise forest: F1 = 120; for enterprises in enterprise forest: F1 = 120 - (R1 - R2) / 5 * 10; F1=F1-10,T1≤10%; F1=F1-5,10% <T1≤50%; F1=F1-5,B1≥10; F1=F1-10,B1≥100; Among them, deductions are made for the distance from the root, and ten points will be deducted for more than 5 layers; for the average registered capital ranking, if the tree is ranked in the bottom 10% of the industry, ten points will be deducted, and if it is 10% < ranking < 50%, five points will be deducted; if the tree is involved in the blacklist for more than 10 times in three years, five points will be deducted, and if it is more than 100 times, ten points will be deducted; The evaluation indicators of operating conditions F2 are as follows: Among them, Z1 represents the average growth rate of tax payment, Z2 represents the growth rate of the number of employees covered by basic medical insurance, Z3 represents the return on net assets in the past two years, Z4 represents the growth rate of sales revenue, W22 represents the percentile ranking of the company's existence time, W23 represents the percentile ranking of the debt ratio, and W24 represents the percentile ranking of the number of valid administrative licenses; Credit status F3 has different calculation methods depending on whether the enterprise belongs to the enterprise forest: For enterprises not in the enterprise forest: in, Oap: represents the original administrative penalty score, Ode represents the original number of times the enterprise was sued, Es represents the scale index, and Rdh represents the percentile ranking of dishonesty enforcement; For enterprises in the enterprise forest: F3 = 150 + 10a + 10b + 10c, where Ap5 represents the administrative penalty score within five years; De3 indicates the number of times the enterprise was sued in three years; Bl3 indicates the number of blacklistings within three years; The evaluation index calculation formula of competitiveness F4 is: in: Rt represents the percentile ranking of trademarks; Rp represents the percentile ranking of patents; Rw represents the percentile ranking of copyrights of works; Rs represents the percentile ranking of software copyrights; Re represents the percentile ranking of registered capital; Rc represents the percentile ranking of qualification certificate scores; Rm represents the percentile ranking of employees’ basic medical insurance; The calculation formula for the evaluation index of industry trend F5 is: F5 = 120 + 6A + 4B + 6C; where: Riu represents the operating income growth rate of large and medium-sized enterprises that are representative of the industry; Reu represents the annual number of new enterprises in the industry; Rpu represents the growth rate of the number of employees in the industry.