A data encryption method for realizing multi-party secure sharing of financial data

Abstract

The present invention relates to the technical field of electrical digital data processing, and particularly relates to a data encryption method for realizing multi-party secure sharing of financial data. By obtaining financial data of different customers from each data provider, respectively performing sensitivity analysis on the financial data to obtain the data sensitivity corresponding to each financial data, determining the importance degree of data features of each financial data according to the data sensitivity, and respectively performing stratification on the financial data according to the importance degree of data features to obtain the target stratification corresponding to each financial data, using the differential privacy algorithm to process the financial data according to the target stratification to obtain the differential privacy result corresponding to each financial data, and obtaining the encrypted data for sharing among each data provider according to the differential privacy result and the homomorphic encryption algorithm, the encryption is more targeted, which is beneficial to reducing the computational complexity, improving the computational efficiency and enhancing the security.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic digital data processing, and particularly to a data encryption method for realizing multi-party secure sharing of financial data. Background Art

[0002] With the rapid development of Internet finance, some data providers, such as traditional financial institutions, are facing new challenges and market competition pressures. Since the financial industry has a vast amount of financial business data, in order to enable financial institutions to reduce data collection costs, optimize product design, and better meet customer needs, data sharing is required within the industry. However, financial data sharing involves the division of interests and powers among multiple enterprises, and Internet finance faces difficulties such as large data asset management and shortage of high-quality available external data. Subsequently, problems such as poor data security and insecure privacy protection methods have emerged, resulting in weak willingness to share data within the industry, and thus difficulties and pain points such as "data islands", "trust gaps", and "data security" exist within and between enterprises. To solve the security problem, some solutions introduce encryption algorithms to directly encrypt and share data. However, during the process of multi-party data sharing, due to the large amount of data, when performing unified encryption algorithm processing on large-scale data, the computational complexity of the encryption operation is relatively high, which easily leads to low computational efficiency and low security due to errors. Summary of the Invention

[0003] In order to solve the above technical problems, an object of the present invention is to provide a data encryption method for realizing multi-party secure sharing of financial data, and the specific technical solution adopted is as follows:

[0004] In a first aspect, an embodiment of the present application provides a method for encrypting data for realizing multi-party secure sharing of financial data, including:

[0005] Obtain financial data of different customers from each data provider;

[0006] Respectively perform sensitivity analysis on the financial data to obtain the data sensitivity corresponding to each financial data, where the data sensitivity is used to characterize the stability of the credit dimension in the financial data;

[0007] According to the data sensitivity, determine the importance degree of the data features of each financial data, and respectively layer the financial data according to the importance degree of the data features to obtain the target layer corresponding to each financial data;

[0008] According to the target stratification, the differential privacy algorithm is used to process the financial data to obtain the differential privacy result corresponding to each piece of financial data, and the encrypted data is obtained according to the differential privacy result and the homomorphic encryption algorithm, and the encrypted data is used for sharing among the respective data providers.

[0009] In one implementation manner, the step of respectively performing sensitivity analysis on the financial data to obtain the data sensitivity corresponding to each piece of financial data includes:

[0010] Respectively perform local sensitivity analysis on the financial data to obtain the local sensitivity corresponding to each piece of financial data;

[0011] Respectively perform global sensitivity analysis on the financial data to obtain the global sensitivity corresponding to each piece of financial data;

[0012] Respectively obtain the data sensitivity corresponding to each piece of financial data according to the product of the local sensitivity and the global sensitivity.

[0013] In one implementation manner, the step of respectively performing local sensitivity analysis on the financial data to obtain the local sensitivity corresponding to each piece of financial data includes:

[0014] Based on a preset time interval, respectively segment the financial data to obtain the financial sub-data corresponding to several time periods of each piece of financial data, and determine the credit score of each piece of financial sub-data;

[0015] Determine the target data from the financial data, and respectively calculate the credit score differences between two adjacent time periods in the target data;

[0016] Perform a summation process according to each of the credit score differences and the number of the credit score differences, and obtain the local sensitivity corresponding to the target data according to the first ratio of the summation process result to the number of the credit score differences;

[0017] Return to the step of determining the target data from the financial data until the local sensitivity corresponding to each piece of financial data is obtained.

[0018] In one implementation manner, the step of respectively performing global sensitivity analysis on the financial data to obtain the global sensitivity corresponding to each piece of financial data includes:

[0019] Respectively determine the target credit score difference between the latest two time periods in each piece of financial data, and determine the minimum credit score difference and the maximum credit score difference among all the financial data;

[0020] Determine the first difference between the target credit score difference and the minimum credit score difference respectively, and determine the second difference between the maximum credit score difference and the minimum credit score difference, and obtain the global sensitivity corresponding to each piece of financial data according to the second ratio of the first difference and the second difference respectively.

[0021] In one implementation manner, the determining the importance degree of the data characteristics of each piece of financial data according to the data sensitivity includes:

[0022] Determine the life cycle of the customer corresponding to each piece of financial data, the number of data providers involved in each customer's financial data, and the total number of the data providers;

[0023] Determine the data sharing range value of each customer according to the number of data providers and the total number;

[0024] Determine the importance degree of the data characteristics of each piece of financial data according to the data sharing range value, the life cycle, and the data sensitivity respectively.

[0025] In one implementation manner, the determining the life cycle of the customer corresponding to each piece of financial data includes:

[0026] Respectively determine the basic information of each customer from each piece of financial data, where the basic information includes customer age, occupation, historical loan record, and transaction frequency;

[0027] Input the basic information of each customer into the life cycle model respectively, and obtain the life cycle of the customer corresponding to each piece of financial data.

[0028] In one implementation manner, the determining the data sharing range value of each customer according to the number of data providers and the total number includes:

[0029] Determine the third ratio of the number of data providers to the total number;

[0030] Determine the data sharing range value of each customer according to the third ratio and the normalization function.

[0031] In one implementation manner, the determining the importance degree of the data characteristics of each piece of financial data according to the data sharing range value, the life cycle, and the data sensitivity respectively includes:

[0032] Respectively determine the sum value of the life cycle and the data sharing range value, and determine the normalization result according to the sum value and the normalization function;

[0033] Determine the importance degree of the data characteristics of each piece of financial data according to the product of the normalization result and the data sensitivity respectively.

[0034] In one implementation, the step of stratifying the financial data according to the importance degree of the data characteristics respectively to obtain the target stratification corresponding to each piece of financial data includes:

[0035] When the importance degree of the data characteristics corresponding to the financial data is greater than or equal to a first threshold, determine that the target stratification of the financial data is the first stratification;

[0036] When the importance degree of the data characteristics corresponding to the financial data is greater than or equal to a second threshold and less than the first threshold, determine that the target stratification of the financial data is the second stratification;

[0037] When the importance degree of the data characteristics corresponding to the financial data is less than the second threshold, determine that the target stratification of the financial data is the third stratification;

[0038] Among them, the data importance degrees represented by the first stratification, the second stratification, and the third stratification decrease in sequence.

[0039] In one implementation, the step of processing the financial data by using the differential privacy algorithm according to the target stratification to obtain the differential privacy result corresponding to each piece of financial data includes:

[0040] When the target stratification is the first stratification, determine that the initial privacy budget is a first value; when the target stratification is the second stratification, determine that the initial privacy budget is a second value; when the target stratification is the third stratification, determine that the initial privacy budget is a third value, and the magnitudes of the first value, the second value, and the third value increase in sequence;

[0041] Determine the target privacy budget according to the importance degree of the data characteristics corresponding to each piece of financial data and a fourth ratio of the initial privacy budget;

[0042] Process the financial data by using the differential privacy algorithm according to the target privacy budget to obtain the differential privacy result corresponding to each piece of financial data.

[0043] The present invention has the following beneficial effects:

[0044] By obtaining the financial data of different customers from various data providers, performing sensitivity analysis on the financial data respectively to obtain the data sensitivity corresponding to each financial data, determining the importance degree of data features of each financial data according to the data sensitivity, and respectively stratifying the financial data according to the importance degree of data features to obtain the target stratification corresponding to each financial data, using the differential privacy algorithm to process the financial data according to the target stratification to obtain the differential privacy result corresponding to each financial data, and obtaining the encrypted data for sharing among various data providers according to the differential privacy result and the homomorphic encryption algorithm. Stratifying by using the importance degree of data features and performing hierarchical encryption of data based on the target stratification through the differential privacy algorithm makes the encryption more targeted, which is beneficial to reducing the computational complexity, improving the computational efficiency, and enhancing the security. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0046] Figure 1 It is a schematic flowchart of the steps of a data encryption method for realizing multi-party secure sharing of financial data provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features, and effects of a data encryption method for realizing multi-party secure sharing of financial data proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0049] It should be noted that the "exemplary" in the embodiments of the present application refers to examples listed for convenience of description, and in other embodiments, it is not limited to the listed examples.

[0050] The following specifically describes the specific solution of a data encryption method for realizing multi-party secure sharing of financial data provided by the present invention in combination with the accompanying drawings.

[0051] Please refer to Figure 1 which shows a flowchart of a data encryption method for realizing multi-party secure sharing of financial data provided by an embodiment of the present invention. The data encryption method for realizing multi-party secure sharing of financial data may at least include steps S100 - S400:

[0052] S100. Obtain financial data of different customers from each data provider.

[0053] S200. Respectively perform sensitivity analysis on the financial data to obtain the data sensitivity corresponding to each financial data.

[0054] In the embodiment of the present application, the data sensitivity is used to characterize the stability of the credit dimension in the financial data.

[0055] S300. Determine the importance degree of the data characteristics of each financial data according to the data sensitivity, and respectively layer the financial data according to the importance degree of the data characteristics to obtain the target layer corresponding to each financial data.

[0056] S400. Process the financial data by using the differential privacy algorithm according to the target layer to obtain the differential privacy result corresponding to each financial data, and obtain the encrypted data according to the differential privacy result and the homomorphic encryption algorithm. The encrypted data is used for sharing among each data provider.

[0057] The technical solution of the embodiment of the present application, by obtaining financial data of different customers from each data provider, respectively performing sensitivity analysis on the financial data to obtain the data sensitivity corresponding to each financial data, determining the importance degree of the data characteristics of each financial data according to the data sensitivity, and respectively layer the financial data according to the importance degree of the data characteristics to obtain the target layer corresponding to each financial data, process the financial data by using the differential privacy algorithm according to the target layer to obtain the differential privacy result corresponding to each financial data, and obtain the encrypted data for sharing among each data provider according to the differential privacy result and the homomorphic encryption algorithm. Laying layers by using the importance degree of the data characteristics, and performing hierarchical encryption of the data based on the target layer through the differential privacy algorithm, the encryption is more targeted, which is beneficial to reducing the computational complexity, improving the computational efficiency and improving the security.

[0058] The method of the embodiment of the present application can be executed by a computer, a mobile phone and other terminals or a server. Exemplarily, taking the execution by a background server data sharing platform as an example for illustration.

[0059] In one embodiment, the data sharing platform can obtain the financial data of different customers of different data providers. Subsequently, the data sharing platform can perform processing such as encrypted storage on the obtained financial data to provide a basis for subsequent use by each data provider. Optionally, the data providers include but are not limited to financial institutions (such as banks, insurance companies), payment platforms, etc.; the financial data can include multiple dimensions such as the basic information of customers, credit records, debt situations, asset changes, etc.; after obtaining the original financial data, preprocessing can be performed, including but not limited to missing value processing, outlier detection, and data normalization, etc., and when the data sharing platform stores the financial data, it can be classified and stored according to different dimensions based on the customer ID corresponding to different customers.

[0060] In one embodiment, step S200 includes steps S201 - S203:

[0061] S201. Perform local sensitivity analysis on the financial data respectively to obtain the local sensitivity corresponding to each financial data.

[0062] First, a preset time interval can be set in advance, such as months, years. Exemplarily, taking years as an example, the financial data is segmented based on years respectively to obtain the financial sub - data of several time periods corresponding to each financial data, that is, in each financial data, the financial sub - data of the customer in several time periods (years). Each financial sub - data represents the credit status of the corresponding time period, and each financial sub - data can contain a credit score, or the credit score of each financial sub - data can be calculated using corresponding methods based on actual needs, without specific limitation.

[0063] Second, determine the target data from the financial data (for example, the financial data of the th customer), and calculate the credit score differences between two adjacent time periods in the target data respectively. For example, the credit score of the th time period of the th customer, and the credit score of the +1th time period of the th customer, so as to determine the credit score differences between two adjacent time periods in the target data .

[0064] Third, perform a summation process according to each credit score difference and the number of credit score differences to obtain the summation result, that is, in the financial data of the th customer, by taking different values, various credit score differences can be determined , and then based on the number of credit score differences Perform a summation process to obtain the result of the summation process: , where is the number of credit scores. Additionally, based on the first ratio of the result of the summation process to the number of differences in credit scores, obtain the local sensitivity corresponding to the target data:

[0065]

[0066] where is the local sensitivity corresponding to the target data (currently representing the financial data of the th customer, for example = 1).

[0067] Then, return the above steps of determining the target data from the financial data, that is, determine the financial data of other customers as the target data until the local sensitivity corresponding to each financial data is determined based on the above principle , at this time represents the local sensitivity corresponding to the financial data of the ( can take any value)th customer.

[0068] S202. Respectively perform global sensitivity analysis on the financial data to obtain the global sensitivity corresponding to each financial data.

[0069] First, respectively determine the target credit score difference between the latest two time periods in each financial data , that is, the target credit score difference of the financial data of the th customer. For example, assume the latest two time periods are and (at this time may be 5), and at this time the corresponding latest two time periods are 5 and 6, and in all financial data, that is, in the financial data of all customers, determine the minimum credit score difference and the maximum credit score difference .

[0070] Secondly, respectively determine the first difference between the target credit score difference and the minimum credit score difference , and determine the second difference between the maximum credit score difference and the minimum credit score difference , and respectively obtain the global sensitivity corresponding to each financial data according to the second ratio of the first difference to the second difference:

[0071]

[0072] where represents the The global sensitivity corresponding to the financial data of each customer.

[0073] S203. Respectively, according to the local sensitivity and the global sensitivity to obtain the data sensitivity corresponding to each financial data.

[0074] Specifically, the formula is:

[0075]

[0076] Among them, represents the data sensitivity corresponding to the financial data of the th customer.

[0077] It should be noted that the local sensitivity characterizes the fluctuation degree of the credit score of the th customer in the credit dimension, that is, the instability of the credit score of the th customer. When the local sensitivity is larger, it indicates that the degree of sharing among different data providers may be relatively high, that is, the importance of the financial data of this customer is relatively high and key attention needs to be paid; while the global sensitivity can quantify the impact of the change in the credit score of a single th customer in the credit dimension on the overall credit score distribution of customers. When the change in the credit score of a single th customer is too large or too small compared with the overall change, it indicates that the change rule of the financial data of this th customer is quite different from the overall change, that is, the stability is poor, and key attention needs to be paid to this th customer; and the data sensitivity is comprehensively determined by the local sensitivity and the global sensitivity . Therefore, the data sensitivity can be used to characterize the stability of the credit dimension in financial data. The greater the stability, the smaller the value.

[0078] In one implementation, in step S300, according to the data sensitivity, determining the importance degree of the data characteristics of each financial data includes steps S301 - S303:

[0079] S301. Determine the life cycle of the customer corresponding to each financial data, the number of data providers involved in the financial data of each customer, and the total number of data providers.

[0080] Optionally, determine the number of data providers involved in the financial data of each customer and the total number of data providers , i.e., the number of providers of the data providers corresponding to the financial data of the th customer. It should be noted that, for example, there are a total of 4 data providers now, namely Financial Institution A, Financial Institution B, Financial Institution C, and Financial Institution D. Since a customer may have participated in business in different financial institutions, for example, participated in business in both Financial Institution A and Financial Institution B, the number of providers of the data providers corresponding to the financial data of this customer is 2, and the total number of data providers is 4.

[0081] Optionally, determine the life cycle of the customer corresponding to each financial data, including:

[0082] First, respectively determine the basic information of each customer from each financial data. The basic information includes but is not limited to customer age, occupation, historical loan records, and transaction frequency.

[0083] Secondly, input the basic information of each customer, such as customer age, occupation, historical loan records, and transaction frequency, etc., into the life cycle model respectively. The life cycle model outputs the life cycle corresponding to the financial data of each customer , representing the life cycle corresponding to the financial data of the th customer.

[0084] It should be noted that the life cycle model is obtained by training a neural network in advance based on data such as customer age, occupation, transaction history, and loan records as a data set. The training method can be based on existing methods and will not be elaborated here; the life cycle model can help infer the change of customer needs at different time periods. For example, for young people, financial business may mainly focus on education loans and gradually transition to consumer loans or housing loans; for middle-aged people, it may mainly involve housing loans, car loans, etc., and as they get older, they gradually turn to demand for pension products, etc.

[0085] In the embodiments of the present application, the life cycle of a customer may include the initial stage (when the customer just starts to take a loan), the growth stage (when the customer is repaying the loan stably and the loan amount is gradually increasing), the maturity stage (when the customer has completed the main loan and starts to plan long-term financial management, etc.), and the decline stage (when the customer's loan demand decreases and starts to turn to other financial products); as the life cycle changes in different stages, the more stable the life cycle is, that is, the decline stage is the most stable and the initial stage is the least stable. Among them, the value of the life cycle is between [0,1]. For example, in one implementation manner, when the life cycle is in the initial stage, the life cycle Take values from 0.75 to 1. When the life cycle is in the growth stage, the life cycle Take values from 0.5 to 0.75. When the life cycle is in the mature stage, the life cycle Take values from 0.25 to 0.5. When the life cycle is in the decline stage, the life cycle Take values from 0 to 0.25.

[0086] S302. Determine the data sharing range value of each customer according to the number of providers and the total number.

[0087] It should be noted that the privacy requirements of data determine the security weighting when sharing data among multiple data providers. The more data providers there are for a customer's data, the higher the weighting should be given to the customer's privacy characteristics, and it is necessary to ensure that it is more protected under encryption and access control. That is, for data with a high sharing frequency and a wide sharing range, because the data is more likely to be read, the requirements for data protection are higher, so their weights should be increased accordingly.

[0088] Specifically, determine the number of providers and the total number of the third ratio, and determine the data sharing range value of each customer according to the third ratio and the normalization function :

[0089] where represents the data sharing range value of the th customer.

[0090] S303. Determine the importance degree of the data characteristics of each financial data according to the data sharing range value, the life cycle, and the data sensitivity respectively.

[0091] It should be noted that the life cycle of a customer's financial data means the change in the customer's behavior pattern, showing the instability of the customer's financial data; the data sharing range means the degree of flow and interaction of the customer's financial data among multiple data providers, showing the breadth of data sharing among multiple data providers. By adding the life cycle and the data sharing range value, the risk level of the customer's financial data when sharing data among multiple data providers can be comprehensively evaluated; therefore, by combining the data risk level with the customer's data sensitivity, the importance degree of the data characteristics of the customer's financial data can be comprehensively evaluated.

[0092] First, determine the sum value of the life cycle and the data sharing range value respectively, and determine the normalization result according to the sum value and the normalization function .

[0093] Secondly, according to the normalization results respectively and the data sensitivity of the product, determine the importance degree of the data characteristics of each financial data:

[0094]

[0095] Among them, represents the importance degree of the data characteristics of the th customer. It should be noted that when the sensitivity changes of the financial data of customers are similar, the more unstable the stage of the life cycle of the financial data of customers, the larger the scope of data sharing, the greater the degree of attention required for the financial data of customers, and the more significant the importance degree of the data characteristics of customers. Therefore, by obtaining , as the weight of the data sensitivity of the th customer, it shows the risk level when the financial data of the customer is shared among multiple data providers, so as to perform weighted processing on the data sensitivity of the customer and improve the discrimination of the importance degree of the data characteristics of the customer when the sensitivity changes of the financial data of the customer are similar.

[0096] In one implementation manner, in step S300, the financial data is stratified according to the importance degree of the data characteristics respectively, and the target stratification corresponding to each financial data is obtained, including steps S304 - S306:

[0097] S304. When the importance degree of the data characteristics corresponding to the financial data is greater than or equal to the first threshold, determine that the target stratification of the financial data is the first stratification.

[0098] Exemplarily, set the first threshold to , and the second threshold to . Therefore, when the importance degree of the data characteristics corresponding to the financial data is greater than or equal to the first threshold , determine that the target stratification of the financial data is the first stratification, indicating that the financial data belongs to important data.

[0099] S305. When the importance degree of the data characteristics corresponding to the financial data is greater than or equal to the second threshold and less than the first threshold, determine that the target stratification of the financial data is the second stratification.

[0100] Optionally, when the importance degree of the data characteristics corresponding to the financial data is greater than or equal to the second threshold and less than the first threshold , that is, , determine that the target stratification of the financial data is the second stratification, indicating that the financial data belongs to general data.

[0101] S306. When the importance level of the data features corresponding to the financial data is less than the second threshold, determine that the target layer of the financial data is the third layer;

[0102] Optionally, when the importance level of the data features corresponding to the financial data is less than the second threshold, that is , determine that the target layer of the financial data is the third layer, indicating that the financial data belongs to unimportant data.

[0103] Therefore, from the above content, it can be known that the importance levels of the data represented by the first layer, the second layer, and the third layer decrease in sequence.

[0104] In one implementation, in step S400, according to the target layer, use the differential privacy algorithm to process the financial data to obtain the differential privacy result corresponding to each financial data, including steps S401 - S403:

[0105] S401. When the target layer is the first layer, determine that the initial privacy budget is the first value; when the target layer is the second layer, determine that the initial privacy budget is the second value; when the target layer is the third layer, determine that the initial privacy budget is the third value.

[0106] In the embodiments of the present application, the magnitudes of the first value, the second value, and the third value increase in sequence. Exemplarily, the first value is , the second value is , the third value is .

[0107] Specifically, when the target layer is the first layer, determine that the initial privacy budget is the first value , when the target layer is the second layer, determine that the initial privacy budget is the second value , when the target layer is the third layer, determine that the initial privacy budget is the third value .

[0108] It should be noted that the purpose of layering is to reduce the risk of information leakage between different data and provide a basis for the subsequent implementation of the differential privacy mechanism. For financial data, highly sensitive features such as users' credit scores and borrowing records usually require stronger privacy protection, while low - sensitive features (such as transaction amounts, consumption categories, etc.) can be protected moderately.

[0109] S402. Determine the target privacy budget according to the fourth ratio of the importance level of the data features corresponding to each financial data and the initial privacy budget.

[0110] Optionally, according to the importance level of the data features corresponding to each financial data and the initial privacy budget the fourth ratio, to determine the target privacy budget .

[0111] S403. According to the target privacy budget, use the differential privacy algorithm to process the financial data to obtain the differential privacy result corresponding to each financial data.

[0112] It should be noted that differential privacy protects privacy by adding noise to the data. The addition of noise is based on the privacy budget. The smaller the value of the privacy budget, the stronger the privacy protection, but it may lead to data distortion; the larger the value of the privacy budget, the more accurate the data, but the weaker the privacy protection. Choosing an appropriate privacy budget is crucial. In the embodiments of the present application, the differential privacy mechanism is applied at different layers, and corresponding noise is added to the data, so that no specific information of any single user can be inferred from the data.

[0113] Optionally, determine the Laplace value constraint in the differential privacy algorithm , that is, the final privacy budget is the final privacy budget for the financial data of the th customer . Then, respectively based on the final privacy budget , use the differential privacy algorithm to process the corresponding financial data to obtain the differential privacy result corresponding to each financial data, which is equivalent to adding different random noises to the financial data of different layers, and obtaining different protection measures for data with different degrees of importance of data features, meeting the dual requirements of different data providers for the processing speed and security of financial data. In addition, the calculation of the differential privacy algorithm can adopt existing methods, which will not be elaborated here. The focus of the embodiments of the present application is to specifically determine the final privacy budget of the financial data of different customers , implement targeted differential privacy algorithm processing, and obtain more accurate, more targeted, and more reflective differential privacy results of the financial data of different customers; and use the differential privacy mechanism to ensure that the financial data to be shared is protected by privacy while maintaining its original value, ensuring that the individual data at the sending end cannot be accurately inferred, significantly reducing the risk of data leakage in the industry, protecting the interests of enterprises and users, enhancing the sharing willingness among different data providers, and effectively avoiding the emergence of data information islands.

[0114] In the embodiments of the present application, after determining the differential privacy results corresponding to each financial data, the data sharing platform can select an appropriate homomorphic encryption algorithm, such as Paillier encryption, according to the above differential privacy results, and further encrypt the differential privacy results to obtain encrypted data and store it. Then, each data provider can extract the encrypted data from the data sharing platform according to its own needs by using the public key distributed by the data sharing platform, so that the encrypted data can be shared among the data providers; among them, after each data provider obtains the encrypted data, it processes the encrypted data in an encrypted state. In the embodiments of the present application, key processing technology is used to convert plaintext data into an encrypted format, ensuring the data concealment and security of different data providers such as financial data providers, data calculation parties, and result acquirers during the information sharing process.

[0115] It should be noted that the above sequence of the embodiments of the present application is only for description and does not represent the advantages or disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0116] The embodiments in the specification of the present application are all described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A data encryption method for realizing multi-party secure sharing of financial data, characterized in that, The method includes: Obtaining financial data of different customers from each data provider; Performing sensitivity analysis on the financial data respectively to obtain the data sensitivity corresponding to each piece of financial data, where the data sensitivity is used to characterize the stability of the credit dimension in the financial data; Among them, the method for obtaining the data sensitivity is: performing local sensitivity analysis on the financial data respectively to obtain the local sensitivity corresponding to each piece of financial data; performing global sensitivity analysis on the financial data respectively to obtain the global sensitivity corresponding to each piece of financial data; respectively obtaining the data sensitivity corresponding to each piece of financial data according to the product of the local sensitivity and the global sensitivity; Among them, the method for obtaining the local sensitivity corresponding to the financial data is: segmenting the financial data respectively based on a preset time interval to obtain financial sub-data for several time periods corresponding to each piece of financial data, and determining the credit score of each piece of financial sub-data; determining target data from the financial data, and calculating the credit score difference between two adjacent time periods in the target data respectively; performing a summation process according to each credit score difference and the number of credit score differences, and obtaining the local sensitivity corresponding to the target data according to the first ratio of the summation result to the number of credit score differences; returning to the step of determining target data from the financial data until the local sensitivity corresponding to each piece of financial data is obtained; Among them, the method for obtaining the global sensitivity corresponding to the financial data is: respectively determining the target credit score difference between the latest two time periods in each piece of financial data, and determining the minimum credit score difference and the maximum credit score difference among all financial data; respectively determining the first difference between the target credit score difference and the minimum credit score difference, and determining the second difference between the maximum credit score difference and the minimum credit score difference, and respectively obtaining the global sensitivity corresponding to each piece of financial data according to the second ratio of the first difference to the second difference; determining the importance degree of the data characteristics of each piece of financial data according to the data sensitivity, and respectively stratifying the financial data according to the importance degree of the data characteristics to obtain the target stratification corresponding to each piece of financial data; Among them, the method for obtaining the importance degree of the data characteristics is: determining the life cycle of the customer corresponding to each piece of financial data, the number of data providers involved in the financial data of each customer, and the total number of data providers; determining the data sharing range value of each customer according to the number of data providers and the total number of data providers; respectively determining the importance degree of the data characteristics of each piece of financial data according to the data sharing range value, the life cycle, and the data sensitivity; Among them, the method for obtaining the life cycle of the customers corresponding to the financial data is as follows: respectively determine the basic information of each customer from each piece of the financial data, where the basic information includes the customer's age, occupation, historical loan record, and transaction frequency; respectively input the basic information of each customer into the life cycle model to obtain the life cycle of the customers corresponding to each piece of the financial data; Among them, the method for obtaining the data sharing range value is as follows: determine the third ratio of the number of the providers to the total number; determine the data sharing range value of each customer according to the third ratio and the normalization function; Among them, the method for obtaining the importance degree of the data characteristics of the financial data is as follows: respectively determine the sum value of the life cycle and the data sharing range value, and determine the normalization result according to the sum value and the normalization function; respectively determine the importance degree of the data characteristics of each piece of the financial data according to the product of the normalization result and the data sensitivity; According to the target stratification, use the differential privacy algorithm to process the financial data to obtain the differential privacy result corresponding to each piece of the financial data, and obtain the encrypted data according to the differential privacy result and the homomorphic encryption algorithm, where the encrypted data is used for sharing among the data providers.

2. The data encryption method for realizing multi-party secure sharing of financial data according to any one of claims 1, wherein: The stratifying the financial data according to the importance degree of the data characteristics respectively to obtain the target stratification corresponding to each piece of the financial data includes: When the importance degree of the data characteristics corresponding to the financial data is greater than or equal to the first threshold, determine that the target stratification of the financial data is the first stratification; When the importance degree of the data characteristics corresponding to the financial data is greater than or equal to the second threshold and less than the first threshold, determine that the target stratification of the financial data is the second stratification; When the importance degree of the data characteristics corresponding to the financial data is less than the second threshold, determine that the target stratification of the financial data is the third stratification; Among them, the data importance degrees represented by the first stratification, the second stratification, and the third stratification decrease in sequence.

3. The data encryption method for realizing multi-party secure sharing of financial data according to claim 2, wherein: The using the differential privacy algorithm to process the financial data according to the target stratification to obtain the differential privacy result corresponding to each piece of the financial data includes: When the target stratification is the first stratification, determine that the initial privacy budget is the first value, when the target stratification is the second stratification, determine that the initial privacy budget is the second value, when the target stratification is the third stratification, determine that the initial privacy budget is the third value, where the magnitudes of the first value, the second value, and the third value increase in sequence; Determine the target privacy budget according to the fourth ratio of the importance degree of the data characteristics corresponding to each piece of the financial data and the initial privacy budget; According to the target privacy budget, use the differential privacy algorithm to process the financial data to obtain the differential privacy result corresponding to each piece of the financial data.

Images