Financial loan information encryption management method based on big data

Abstract

The invention relates to the technical field of information encryption, in particular to a financial loan information encryption management method based on big data, and the method comprises the following steps: S1, an encryption platform marks an authority level identifier on financial loan data at a collection end through a homomorphic encryption algorithm, carries out encryption processing, and outputs encrypted data; and S2, the encryption platform transforms a searchable encryption technology through a mathematical transformation mode, when the method is used, original data does not need to be decrypted, privacy security is ensured, permission level identification is attached to each field, only roles meeting conditions can use related data, abuse or unauthorized access is effectively prevented, and the security of the data is improved. A bank can provide encrypted data for a risk control model, a third-party credit investigation platform and the like for analysis and processing, original information leakage can be avoided, a traditional searchable encryption technology is compatible with a homomorphic encryption model through mathematical structure conversion, the whole search and weighting process is completed in an encryption state, and the method is high in practicability. Keywords and search intentions do not need to be exposed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of information encryption, and in particular to a financial loan information encryption management method based on big data. BACKGROUND

[0002] Financial loan information encryption is a key measure to ensure that all sensitive data involved in the loan process is fully protected. With the rapid development of financial technology, it is particularly important to protect user privacy and data security while improving service efficiency through the use of big data.

[0003] Patent No. CN113382021A, in the specification, records "The present application discloses a financial data processing method, which is executed by an encryption end respectively connected in communication with a first end and a second end, comprising: obtaining financial data to be processed, and obtaining a first number from the first end and a second number from the second end. According to the first number and the second number, a public key and a private key are generated. The financial data is encrypted using the private key to obtain encrypted data; the encrypted data and the public key are sent to the second end, so that the second end uses the public key to decrypt the encrypted data to obtain the financial data, and then the second end sends the financial data to the first end. The first end processes the financial data to obtain an intermediate result. The method in the specification is suitable for data processing systems or methods for administrative, commercial, financial, management, supervision or prediction purposes; and a communication network that selectively establishes one or more wireless communication links between a desired number of users or between a user and a network device in order to transmit information via these communication links", the above technology combines asymmetric encryption with distributed trust verification, and constructs a secure data processing framework suitable for the characteristics of the financial industry, which has significant application value in high-sensitive scenarios such as payment clearing and cross-institution reconciliation, but the current traditional financial loan information encryption management method is to encrypt and store the data first, and then decrypt and analyze when using, which exposes the data in the system memory, interface or log during the use stage, and there is a serious risk of leakage. The existing financial loan information encryption management method usually protects the entire data, but cannot flexibly control the decryption range according to the permissions of the calling party, resulting in the problem that the risk control system only needs to check the user risk score, but may have read the complete information of the user.

[0004] In summary, it is still a key problem in the field of information encryption technology to develop a financial loan information encryption management method based on big data. SUMMARY

[0005] The purpose of the present application is to solve the above-mentioned technical problems in the prior art. Although the asymmetric encryption is combined with the distributed trust verification, a secure data processing framework suitable for the characteristics of the financial industry is constructed, which has significant application value in high-sensitive scenes such as payment clearing and cross-institution reconciliation. However, the current traditional financial loan information encryption management method is to encrypt and store the data first, and then decrypt and analyze when using, which exposes the data in the system memory, interface or log during the use stage, and there is a serious risk of leakage. The existing financial loan information encryption management method usually protects the entire data, but cannot flexibly control the decryption range according to the permissions of the calling party, resulting in the problem that the risk control system only needs to check the user risk score, but may read the complete information of the user at the same time.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] The present application provides a financial loan information encryption management method based on big data, comprising the following steps: S1, the encryption platform encrypts the financial loan data after marking the permission level identifier on the collection end by using the homomorphic encryption algorithm, and outputs the encrypted data;

[0008] S2, the encryption platform transforms the searchable encryption technology by using the mathematical transformation method, and outputs the encryption model;

[0009] S3, the encrypted data is obtained by using the encryption calculation engine according to the encrypted data and the encryption model;

[0010] S4, the encrypted data is returned to the data owner, and the data owner obtains the calling permission search matching field by using the access control platform;

[0011] S5, the access control label and the operation log auditing mechanism are integrated in the homomorphic calculation process of the encryption model.

[0012] Further, in step S1, the encryption platform encrypts the financial loan data after marking the permission level identifier on the collection end by using the homomorphic encryption algorithm, and outputs the encrypted data by using the method of:

[0013] The collection end structures and arranges the user's financial loan data, the encryption platform assigns a permission level identifier to each dimension according to different data sensitivity, respectively representing low, medium and high sensitivity levels, and the homomorphic encryption algorithm introduces a Paillier homomorphic encryption scheme, and sets the public key as Wherein qk is the public key for encrypting data, n is an integer as a modulus in the encryption algorithm, and g is a generator for generating all element bases

[0014] Enc qk =g m ·En modn 2 ,

[0015] Quasi-number, expression formula: Where Y represents the output after the entire encryption operation, Enc qk (U) is an encryption operation that encrypts data using a public key qk, U1, U2, U3 is the original data item to be encrypted, is a random number, n 2 is a large prime number, modn 2 is a modulo operation, E is a randomly selected number in the encryption process, E must be selected from the multiplicative group of modulus n, E n is the n-th power of the random number E, g m represents the m-th power of the generator g, the final output is the ciphertext vector, the right to identify, form encrypted data, expression formula: Where I is the result set of encrypted data, U i is the original data item to be encrypted, represents the permission level related to U i , is a set representation.

[0016] Further, in step S2, the encryption platform transforms the searchable encryption technology through mathematical transformation, and the method for outputting the encryption model is:

[0017] The searchable encryption technology is transformed to construct the index structure of traditional symmetric searchable encryption, and the keyword set is set to: O={(o1,o2,...,o m )}, the fields in the encrypted data are mapped to the keyword space, the keyword label is generated through the pseudo-random function, and the query operation of traditional symmetric searchable encryption is transformed into the form of multiplicative nested expression through mathematical transformation, and the nested function is used to represent the Boolean value in the form of a polynomial structure, compatible with the Paillier homomorphic addition structure, expression formula: Where is a nested function used in an encryption scenario, k i is the input passed to the function as a parameter, is the sum symbol, α t is a weight coefficient related to t, is a binary function, β ij is an indicator function related to i and j, is the sum of all α t , a weight function is introduced to construct a homomorphic weighted expression, expression formula: wherein Ξ represents an encrypted output, is a summation symbol representing the summation of terms for j = 1 to j = 3, χ j is a weight coefficient associated with each j, U i is an original data item to be encrypted.

[0018] Further, in step S2, the encryption platform transforms the searchable encryption technique by mathematical transformation, and the method for outputting the encryption model is:

[0019] Since U j has been homomorphically encrypted as Enc qk (U j ), the homomorphic weight expression is transformed into a structure in the encryption domain, and the expression formula is: wherein represents the multiplication operation of the three terms for j = 1, 2, 3, Ξ represents an encrypted output, Enc qk (·) is an encryption operation, χ j is a weight coefficient associated with each j, and finally, the encryption model is output, and the expression formula is:

[0020] Further, in step S3, according to the encrypted data and the encryption model, the method for obtaining the encrypted data by the encryption calculation engine is:

[0021] The encryption calculation engine uses the Paillier additive homomorphism expression formula: wherein Enc qk (P1) represents the result of encrypting the message P1 using the public key qk, Enc qk (P2) represents the result of encrypting the message P2 using the public key qk, P1+P2 represents the plaintext addition operation of the two messages, modn 2 represents taking the result modulo, a·P represents the multiplication operation of the message P by the scalar a, a is an integer, Enc qk (a+P) represents the result of encrypting the message a·P after scalar a multiplication processing, Enc qk (P) a represents performing an exponential operation on the encrypted message P and calculating the structure in the encryption domain, the linear combination expression, represents embedding the permission weight and the keyword hit state into the weight coefficient, and realizing the fusion expression in the ciphertext.

[0022] Further, in step S3, according to the encrypted data and the encryption model, the method for obtaining the encrypted data by the encryption calculation engine is:

[0023] The encryption computing engine introduces a nonlinear model of polynomial approximation for predicting credit risk scores, sets the original score function as sigmoid form, and calculates the encrypted polynomial approximation function in the ciphertext domain with Horner structure, expressed as: where the ciphertext power Enc qk (Ξ t ) is iteratively estimated using repeated squaring and homomorphic multiplication, A(s) represents the original activation function in sigmoid form, s represents the input value, e -s represents the negative exponent operation in the exponential function for generating the output of the sigmoid function, ε t represents the weighted coefficient in the polynomial expansion, s t represents s raised to the power of t, and T represents the maximum order of the polynomial expansion, resulting in the encrypted data: Enc qk (f(Ξ))=φ soe , where φ soe is the encrypted data, and f(·) represents a function for scoring.

[0024] Further, in step S4, the encrypted data is returned to the data owner, and the data owner obtains the calling permission to search for matching fields through an access control platform by:

[0025] The encrypted data is returned to the data owner after being output by the encryption computing engine, and the data owner needs to submit a use declaration through the access control platform, the content of the use declaration is structured as a set of use label vectors, the use label vectors are generated into a use digest through a hash mapping function, the user submits an identity credential, the identity credential is verified by the access control platform, and an identity key digest is generated where is the generated identity feature value, H(·) is an encryption hash function, ID is the identity information of the user, || represents string concatenation, S is the current time stamp of the request initiation, the calling permission is generated by the access control platform according to the identity key digest, and the expression formula is: where γ represents the access permission control meta information of the ciphertext result, η ttl is the validity period of the calling permission, ι max is the maximum number of calling times, is the identity key digest, is the hash use digest.

[0026] Further, in step S4, the encrypted data is returned to the data owner, and the data owner obtains the calling permission to search for matching fields through an access control platform by:

[0027] The calling permission will be written into a pre-deployed smart contract, which automatically judges whether the request is legal, searches for matching fields after judging that it is legal, and constructs a response output ciphertext set according to the homomorphic mapping function introduced by the encrypted data, and the expression formula is: Wherein Π j is the permission weighted index of the jth field used to construct the exponential weight of the encrypted field, is the permission label , m is the number of users in the data user set, is an indicator function, if the user strategy q i matches the original data item U j , it is 1, otherwise it is 0, and κ ij represents the policy matching strength of user i to field j, F is the encrypted field response set returned to the user, Enc qk (U j ) is the encryption of the part of the original data item U j that the user is authorized to access using the public key qk, Π j is the permission weight index of the field constructed by the above formula, is the exponential operation on the ciphertext, |Π j ≠0 is to output only the fields with Π j ≠0, is the final ciphertext result, and γ represents the access permission control meta information of the ciphertext result.

[0028] Further, in step S5, the method of integrating access control labels and operation log auditing mechanisms in the homomorphic calculation process of the encryption model is:

[0029] The integrated access control label introduces a label integration mapping function, which is used to bind each piece of encrypted data with an access control label tuple, and the expression formula is: Wherein G represents the set of all access control labels, y i represents the access control label of the ith user, is the data owner identity hash, is the permission role code (evaluation, approval, risk control), is the field access scope, is the permission time parameter, the access control label is embedded in the encrypted data by the label encryption function TagEnc, and the label tracking mechanism is set, and the expression formula is: Wherein Enc qk (U j ) is the encryption of the part of the original data item U j that the user is authorized to access using the public key qk, y i represents the access control label of the ith user, and Ji is the ciphertext data of the ith item, TagEnc qk (·) indicates that the public key qk is used to encrypt the label y i The result after encryption is For any ciphertext data, Indicates that there is a label extraction function that can extract the label from the ciphertext, L lbl (J i )=y i Indicates that the result of the extraction function is the original label y i .

[0030] Furthermore, in step S5, the method of integrating access control tags and operation log auditing mechanisms in the homomorphic computation process of the encryption model is as follows:

[0031] The operation log audit mechanism, each homomorphic computation will be recorded as an operation log entry, denoted as: X t = <OP t ,J in ,J out ,y atie ,T t ′>, where X t The tth operation log entry represents a structured log record, OP t is the operation type, J in is the input ciphertext of this operation, J out is the output ciphertext after the operation, y atie Is the current operation visitor tag, T t ′ is the timestamp of the operation, which is used for audit and traceability verification. All operation logs are connected in a hash chain to form an unalterable log chain for audit. The expression formula is:

[0032] Where h0 is the hash value of the 0th node in the log hash chain, is a secure hash function, X0 is the 0th operation log entry, h t is the hash value of the tth log record, X t is the log content of the t-th operation, h t-1 is the hash value of the t-1th log, Is user X t Use its private key Operation log X t Digital signature made, is the entire log hash chain, {h0,h1,...,h t} is a set of all hash values from the 0th to the tth, and the encryption model generates the result ciphertext and the tracking label in each operation, and records the log hash, and the expression formula is:

[0033] V hm (J i ,J j )→<J c =(Enc qk (f(U i ,U j )),y c ), Wherein V hm is a homomorphic encryption calculation function, J i ,J j represent two input ciphertext data, J c represent the output result ciphertext, Enc qk (·) represents the operation of encryption using the public key qk, f(U i ,U j ) represents the function operation on the plaintext, y c represents a new label obtained after operation, and || represents string splicing.

[0034] Beneficial effects

[0035] Compared with the known prior art, the technical scheme provided by the present application has the following beneficial effects:

[0036] When the present application is used, the original data does not need to be decrypted, the privacy is ensured, the permission level identifier attached to each field is convenient for only the roles meeting the conditions to use the related data, the misuse or unauthorized access is effectively prevented, the bank can provide the encrypted data to the risk control model, the third-party credit investigation platform and the like for analysis and processing, which is beneficial to avoid the leakage of original information, the present method converts the traditional searchable encryption technology through mathematical structure and is compatible with the homomorphic encryption model, and the whole search and weighting process is completed in the encrypted state, without exposing the keyword and search intention.

[0037] When the present application is used, the present method enables the data owner to obtain the encrypted field calling permission through the use declaration and identity authentication by using the access control platform without decrypting the data, and the encrypted data is automatically judged and returned by the smart contract, the data always exists in the form of ciphertext, and even in the authorization process, the data does not need to be decrypted, the sensitive data leakage is effectively prevented, the access permission is controlled according to the use, the role and the identity, and the strict limitation of "who can access, access what and access for how long" is achieved, and the unauthorized access and misuse are prevented. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1This is a flow chart of a financial loan information encryption management method based on big data of the present invention. DETAILED DESCRIPTION

[0039] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0040] It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the numbers used in this manner are interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or apparatuses.

[0041] The present invention is described in further detail below with reference to the accompanying drawings:

[0042] Example:

[0043] like Figure 1 As shown, the present invention provides a financial loan information encryption management method based on big data, comprising the following steps: S1, an encryption platform uses a homomorphic encryption algorithm to mark the financial loan data with an authority level mark at the collection end, performs encryption processing, and outputs the encrypted data;

[0044] Furthermore, in step S1, the encryption platform uses a homomorphic encryption algorithm to mark the financial loan data with an authority level mark at the collection end, and then encrypts the data. The method for outputting the encrypted data is as follows:

[0045] The collection end structures and organizes the user's financial loan data. The encryption platform assigns permission level identifiers to each dimension according to different data sensitivity levels, representing low, medium and high sensitivity levels respectively. The homomorphic encryption algorithm introduces a Paillier homomorphic encryption scheme, assuming that the public key is Where qk is the public key used to encrypt data, n is an integer used as the modulus in the encryption algorithm, and g is the generator used to generate the basis of all elements.

[0046] Enc qk =g m ·E n modn 2 ,

[0047] Quasi number, expression formula: Where Y represents the output after the entire encryption operation, Enc qk (U) is an encryption operation that uses the public key qk to encrypt data. U1, U2, and U3 are the original data items to be encrypted. is a random number, n 2 is a large prime number, modn 2 It is a modular operation, E is a number randomly selected during the encryption process, means that E must be selected from the multiplicative group modulo n, E n is the random number E to the power of n, g m It represents the mth power of the generator g, and finally outputs the ciphertext vector, retains the permission identifier, and forms encrypted data. The expression formula is: Where I is the result set of encrypted data, U i is the original data item to be encrypted, Indicates that i The relevant permission levels, It is a set representation.

[0048] In this embodiment, after adopting homomorphic encryption, the bank can perform legal operations on the ciphertext (such as statistics, summation, etc.) without decrypting the original data, ensuring privacy and security. The permission level identification attached to each field ensures that only qualified roles can use the relevant data, effectively preventing abuse or unauthorized access. The bank can provide encrypted data to risk control models, third-party credit reporting platforms, etc. for analysis and processing, which is conducive to avoiding the leakage of original information.

[0049] S2. The encryption platform transforms the searchable encryption technology through mathematical transformation and outputs the encryption model;

[0050] Furthermore, in step S2, the encryption platform transforms the searchable encryption technology through mathematical transformation, and the method of outputting the encryption model is:

[0051] The searchable encryption technology is transformed to construct the index structure of traditional symmetric searchable encryption, and the keyword set is set as: O = {(o1, o2, ..., o m )}, map the fields in the encrypted data to the keyword space, generate keyword tags through pseudo-random functions, transform the query operation of traditional symmetric searchable encryption into a multiplicative nested expression form through mathematical transformation, and design a nested function For the Boolean value is expressed in polynomial structure, compatible with Paillier homomorphic addition structure, expression formula: Wherein Is a nested function for encryption scenario, k i Is the input passed to the function As a parameter, Is the sum symbol, alpha t Is the weight coefficient related to t, Is a binary function, beta ij Is the indicator function related to i and j, Is the sum of all alpha t Summed for all t, the weight function is introduced, and the homomorphic weighted expression is constructed, expression formula: Wherein, represents an encrypted output, Is the sum symbol representing the sum of the terms from j=1 to j=3, chi j Is the weight coefficient related to each j, U i Is the original data item to be encrypted.

[0052] Further, in step S2, the encryption platform transforms the searchable encryption technology by mathematical transformation, and the method for outputting the encryption model is:

[0053] Since U j Has been homomorphically encrypted as Enc qk (U j ), the homomorphic weighted expression is transformed into a structure in the encrypted domain, expression formula: Wherein Represents the product operation of the three terms j=1, 2, 3, represents an encrypted output, Enc qk (·) is an encryption operation, chi j Is the weight coefficient related to each j, and finally, the encryption model is output, expression formula:

[0054] In this embodiment, the method transforms the traditional searchable encryption technology by mathematical structure transformation and compatible with the homomorphic encryption model, and constructs an encryption model that supports privacy protection, search and calculation. The entire search and weighting process is completed in an encrypted state without exposing the keyword and search intent.

[0055] S3, according to the encryption data and the encryption model, the encryption data is obtained through the encryption calculation engine;

[0056] Further, in step S3, according to the encryption data and the encryption model, the method for obtaining the encryption data through the encryption calculation engine is:

[0057] The encryption calculation engine uses a Paillier additive homomorphism expression formula: Wherein, Enc qk (P1) represents the result of encrypting the message P1 using the public key qk, Enc qk (P2) represents the result of encrypting the message P2 using the public key qk, P1+P2 represents the plaintext addition operation of the two messages, mod n 2 represents the modulus operation on the result, a·P represents the scalar a multiplication operation on the message P, a is an integer, Enc qk (a+P) represents the result of encrypting the message a·P after scalar a multiplication processing, Enc qk (P) a represents the exponential operation on the encrypted message P, and calculates the structure in the encryption domain, the linear combination expression, represents embedding the permission weight and the keyword hit state into the weighted coefficient, and realizing the fusion expression in the ciphertext.

[0058] Further, in step S3, according to the encryption data and the encryption model, the method for obtaining the encryption data by the encryption calculation engine is:

[0059] The encryption calculation engine introduces a nonlinear model of polynomial approximation for predicting credit risk scores, sets the original score function to be in the form of sigmoid, and calculates the encrypted polynomial approximation function in the ciphertext domain in the Horner structure, and the expression formula is: Wherein, the ciphertext power Enc qk (Ξ t ) is iteratively estimated using repeated squaring and homomorphic multiplication, A(s) represents an original activation function in the form of sigmoid, s represents an input value, e -s represents the negative exponential operation in the exponential function for generating the output of the sigmoid function, ε t represents the weighted coefficient in the polynomial expansion, s t represents the t-th power of s, and T represents the maximum order number of the polynomial expansion, and the encryption data obtained is: Enc qk (f(Ξ))=φ soe , wherein φ soe is the encryption data, and f(·) represents a function for scoring.

[0060] In the present embodiment, the method utilizes the homomorphism of Paillier encryption and the polynomial approximation function to complete complex encryption calculation in the ciphertext domain to generate encrypted encrypted data, so that all score calculations are completed in the ciphertext domain, sensitive data does not need to be decrypted, user privacy security is improved, regulatory requirements are met, the calculation process has mathematical verifiability, the risk of human intervention or data leakage is avoided, the credibility of the score result is improved, the bank headquarters and the external risk control platform can share the encrypted data for collaborative evaluation without exposing the original data, and the cross-institutional data collaboration demand is met.

[0061] S4, return the encrypted data to the data owner, and the data owner obtains a calling permission to search a matching field through an access control platform;

[0062] Further, in step S4, the method for returning the encrypted data to the data owner and the data owner obtaining a calling permission to search a matching field through an access control platform is:

[0063] The encrypted data is returned to the data owner after being output by the encryption calculation engine, and the data owner needs to submit a use declaration through an access control platform, the content of the use declaration is structured as a set of use label vectors, a use abstract is generated from the use label vectors through a hash mapping function, a user submits an identity credential, the identity credential is verified by the access control platform, and an identity key abstract is generated Wherein is the generated identity feature value, H(·) is an encryption hash function, ID is the identity information of the user, || represents string splicing, S is the timestamp of the current request initiation, the calling permission is generated according to the identity key abstract through the access control platform, and the expression formula is: Wherein γ represents the access permission control meta information of the ciphertext result, η ttl is the validity period of the calling permission, ι max is the maximum number of calling times, is the identity key abstract, is the hash use abstract.

[0064] Further, in step S4, the method for returning the encrypted data to the data owner and the data owner obtaining a calling permission to search a matching field through an access control platform is:

[0065] The calling permission will be written into a previously deployed smart contract, the smart contract automatically judges whether the request is legal, searches for a matching field after judging that it is legal, constructs a response output ciphertext set according to the encrypted data and introduces a homomorphic mapping function, and the expression formula is: Wherein Π jis the authority weighted index of the jth field, which is used to construct the exponential weight of the encrypted field. Is the permissions tag The corresponding weight function, m is the number of users in the data user set, is an indicator function if the user policy q i With the original data item U j If it matches, it is 1; otherwise, it is 0. ij Indicates the policy matching strength of user i for field j, F is the encrypted field response set returned to the user, Enc qk (U j ) is the use of the public key qk to j The part that the user is authorized to access is encrypted, Π j is the permission weight index of the field constructed by the above formula, is the exponential operation on the ciphertext, |Π j ≠0 means only output π j ≠0 field, is the final ciphertext result, and γ represents the access permission control metadata of the ciphertext result.

[0066] In this embodiment, this method allows the data owner to use the access control platform to obtain the right to call the encrypted field through purpose declaration and identity authentication without decrypting the data, and the smart contract automatically determines and returns the encrypted data. The data always exists in ciphertext form and does not need to be decrypted even during the authorization process, effectively preventing the leakage of sensitive data. Access rights are controlled according to factors such as purpose, role, and identity, so that "who can access, what to access, and how long to access" are strictly restricted to prevent unauthorized access and abuse. The legality of the request is automatically determined by the smart contract, reducing the manual review process and improving the degree of automation.

[0067] S5. Integrating access control tags and operation log auditing mechanisms into the homomorphic computation process of the encryption model;

[0068] Furthermore, in step S5, the method of integrating access control tags and operation log auditing mechanisms in the homomorphic computation process of the encryption model is as follows:

[0069] The supporting integrated access control tag introduces a tag integration mapping function for binding each encrypted data to an access control tag tuple, expressed as: Where G represents the set of all access control labels, y i represents the access control label of the i-th user, is the data owner identity hash, It is the permission role code (assessment, approval, risk control), is the field access scope, is the access control label, which is embedded into the encrypted data by a tag encryption function TagEnc and sets up a tag tracing mechanism, expressed as: where Enc qk (U j ) is the encryption of the part of the original data item U j that the user is authorized to access using the public key qk, y i represents the access control label of the i-th user, J i is the i-th piece of ciphertext data, and TagEnc qk (·) represents the result of encrypting the label y i using the tag encryption public key qk, represents that for any piece of ciphertext data, there exists a label extraction function L lbl (J i ) = y i , which can extract the label y i from the ciphertext.

[0070] Further, in step S5, the access control label and the operation log auditing mechanism are integrated in the homomorphic computation process of the encryption model as follows:

[0071] The operation log auditing mechanism records each homomorphic computation as an operation log entry, denoted as: t X t = <OP in , J out , J atie , y t , T t ′>, where X t is the t-th operation log entry representing a structured log record, OP in is the operation type, J out is the input ciphertext of the current operation, J atie is the output ciphertext after the operation, y t is the current operation visitor label, and T t ′ is the timestamp of the current operation for auditing and traceability verification. All operation logs are connected in a hash chain manner to form an audit immutable log chain, expressed as:

[0072] where h0 is the hash value of the 0-th node of the log hash chain, is a secure hash function, X0 is the 0-th operation log entry, h t is the hash value of the t-th log record, X t is the log content of the t-th operation, and ht-1 is the hash value of the t-1th log, is user X t uses its private key to make a digital signature on the operation log X t , is the entire log hash chain, {h0, h1,..., ht-1, ht} t} is the set of all hash values from the 0th to the tth, and the encryption model generates the result ciphertext and the tracking label in each operation, and records the log hash, and the expression formula is:

[0073] V hm (J i ,J j )→<J c =(Enc qk (f(U i ,U j )),y c ), wherein V hm is a homomorphic encryption calculation function, J i ,J j represent two input ciphertext data, J c represents the output result ciphertext, Enc qk (·) represents an encryption operation using the public key qk, f(U i ,U j ) represents a function operation on the plaintext, y c represents a new label obtained after the operation, and || represents string concatenation.

[0074] In the embodiment, the method effectively integrates the access control label and the operation log auditing mechanism to realize the mechanism of data security, accurate permission control and operation traceability, facilitates the log record and signature at each access and operation, meets the compliance and auditing requirements, and through the integration of the label and the log, the method has the "self-explaining ability", and can automatically judge the legality even in a complex scene.

[0075] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit it; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements will not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A financial loan information encryption management method based on big data, characterized in that: The following steps are involved: S1. The encryption platform uses a homomorphic encryption algorithm to mark the financial loan data with permission levels at the collection end, encrypts it, and outputs the encrypted data. S2. The encryption platform transforms the searchable encryption technology through mathematical transformation and outputs the encryption model; S3. Obtaining encrypted data through an encryption calculation engine based on the encrypted data and the encryption model; S4. Return the encrypted data to the data owner, and the data owner obtains the call permission through the access control platform to search for matching fields; S5. Integrate access control tags and operation log audit mechanisms in the homomorphic computing process of the encryption model.

2. A financial loan information encryption management method based on big data according to claim 1, characterized in that: In step S1, the encryption platform uses a homomorphic encryption algorithm to mark the financial loan data with permission levels at the collection end, and then encrypts the data. The method for outputting the encrypted data is as follows: The collection end structures and organizes the user's financial loan data. The encryption platform assigns permission level identifiers to each dimension according to different data sensitivity levels, representing low, medium and high sensitivity levels respectively. The homomorphic encryption algorithm introduces a Paillier homomorphic encryption scheme, assuming that the public key is Where qk is the public key used to encrypt data, n is an integer used as the modulus in the encryption algorithm, and g is the base number used by the generator to generate all elements. The expression formula is: Where Y represents the output after the entire encryption operation, Enc qk (U) is an encryption operation that uses the public key qk to encrypt data. U1, U2, and U3 are the original data items to be encrypted. is a random number, n 2 is a large prime number, modn 2 It is a modular operation, E is a number randomly selected during the encryption process, means that E must be selected from the multiplicative group modulo n, E n is the random number E to the power of n, g m It represents the mth power of the generator g, and finally outputs the ciphertext vector, retains the permission identifier, and forms encrypted data. The expression formula is: Where I is the result set of encrypted data, U i is the original data item to be encrypted, l i Indicates that i The relevant permission levels, It is a set representation.

3. A financial loan information encryption management method based on big data according to claim 2, characterized in that: In step S2, the encryption platform transforms the searchable encryption technology through mathematical transformation, and the method of outputting the encryption model is: The searchable encryption technology is transformed to construct the index structure of traditional symmetric searchable encryption, and the keyword set is set as: O = {(o1, o2, ..., o m )}, map the fields in the encrypted data to the keyword space, generate keyword tags through pseudo-random functions, transform the query operation of traditional symmetric searchable encryption into a multiplicative nested expression form through mathematical transformation, and design a nested function Used to represent Boolean values ​​as polynomial structures, compatible with Paillier homomorphic addition structures, expression formula: in It is a nested function used in encryption scenarios, k i is passed as a parameter to the function Input, is the summation symbol, α t is the weight coefficient associated with t, is a binary function, β ij is the indicator function associated with i and j, is to sum over all t to get all α t The sum of , introduce the weight function, construct a homomorphic weighted expression, and express the formula: Where Ξ represents an encrypted output, is the summation symbol indicating the summation of the items from j = 1 to j = 3, j is the weight coefficient associated with each j, U i is the original data item to be encrypted.

4. A financial loan information encryption management method based on big data according to claim 3, characterized in that: In step S2, the encryption platform transforms the searchable encryption technology through mathematical transformation, and the method of outputting the encryption model is: Because U j Homomorphically encrypted to Enc qk (U j ), transform the homomorphic weighted expression into a structure in the encryption domain, and express the formula: in represents the product operation of the three items j=1,2,3, Ξ represents an encrypted output, Enc qk (·) is the encryption operation, χ j is the weight coefficient associated with each j, and finally, the output encryption model is expressed as:

5. The method for encrypting and managing financial loan information based on big data according to claim 4, characterized in that: In step S3, based on the encrypted data and the encryption model, the method for obtaining the encrypted data through the encryption calculation engine is: The encryption calculation engine uses the Paillier addition homomorphism expression formula: Among them, Enc qk (P1) represents the result of encrypting the message P1 using the public key qk, Enc qk (P2) represents the result of encrypting message P2 using public key qk, P1+P2 represents the plaintext addition operation of two messages, modn 2 Indicates taking the modulus of the result, a·P means multiplying the message P by scalar a, a is an integer, Enc qk (a+P) represents the result of encrypting the message a·P after the scalar a multiplication process. Enc qk (P) a It represents performing exponential operation on the encrypted message P and calculating the structure in the encryption domain. The linear combination expression represents embedding the authority weight and the keyword hit status into the weighting coefficient to achieve fusion expression in the ciphertext.

6. A financial loan information encryption management method based on big data according to claim 5, characterized in that: In step S3, based on the encrypted data and the encryption model, the method for obtaining the encrypted data through the encryption calculation engine is: The encryption calculation engine introduces a nonlinear model of polynomial approximation to predict credit risk scores. The original scoring function is set to a sigmoid form. The encryption calculation engine calculates the encryption polynomial approximation function in the ciphertext domain using the Horner structure, and the expression formula is: The ciphertext power Enc qk (Ξ t ) Iterative estimation is performed using repeated squares and homomorphic multiplications, A(s) represents the original activation function in the form of sigmoid, s represents the input value, e -s Indicates that the negative exponential operation in the exponential function is used to generate the output of the sigmoid function, ε t represents the weight coefficient in the polynomial expansion, s t represents t-th power of s, T represents the maximum order of polynomial expansion, and the encrypted data is: Enc qk (f(Ξ))=φ soe , where φ soe is the encrypted data, and f(·) represents a function used for scoring.

7. A financial loan information encryption management method based on big data according to claim 6, characterized in that: In step S4, the encrypted data is returned to the data owner, and the method for the data owner to obtain the calling authority to search for matching fields through the access control platform is: The encrypted data is returned to the data owner. The encrypted data is output by the encryption calculation engine and then returned to the data owner. The data owner needs to submit a purpose statement through the access control platform. The content of the purpose statement is structured into a set of purpose label vectors. The purpose label vectors are used to generate a purpose summary through a hash mapping function. The user submits identity credentials, which are verified by the access control platform. After passing the verification, an identity key summary is generated. in is the generated identity feature value, H(·) is a cryptographic hash function, ID is the user's identity information, || represents string concatenation, S is the timestamp of the current request, and the call permission is generated by the access control platform based on the identity key digest, expressed as follows: Where γ represents the access control metadata of the ciphertext result, η ttl Is the validity period of the call permission, ι max Is the maximum number of calls limit, is the identity key digest, is the hash usage digest.

8. The method for encrypting and managing financial loan information based on big data according to claim 7, characterized in that: In step S4, the encrypted data is returned to the data owner, and the method for the data owner to obtain the calling authority to search for matching fields through the access control platform is: The call permission will be written into a pre-deployed smart contract. The smart contract automatically determines whether the request is legal. After determining whether it is legal, it searches for matching fields and introduces a homomorphic mapping function based on the encrypted data to construct a response output ciphertext set. The expression formula is: where π j is the authority weighted index of the jth field, which is used to construct the exponential weight of the encrypted field. j ) is the permission tag l j The corresponding weight function, m is the number of users in the data user set, is an indicator function if the user policy q i With the original data item U j If it matches, it is 1; otherwise, it is 0. ij Indicates the policy matching strength of user i for field j, F is the encrypted field response set returned to the user, Enc qk (U j ) is the use of the public key qk to j The part that the user is authorized to access is encrypted, Π j is the permission weight index of the field constructed by the above formula, is the exponential operation on the ciphertext, |Π j ≠0 means only output π j ≠0 field, is the final ciphertext result, and γ represents the access permission control metadata of the ciphertext result.

9. A financial loan information encryption management method based on big data according to claim 8, characterized in that: In step S5, the method of integrating access control tags and operation log auditing mechanisms in the homomorphic computation process of the encryption model is as follows: The supporting integrated access control tag introduces a tag integration mapping function for binding each encrypted data to an access control tag tuple, expressed as: Where G represents the set of all access control labels, y i represents the access control label of the i-th user, is the data owner identity hash, It is the permission role code (assessment, approval, risk control), is the field access scope, It is a parameter of permission validity. The access control tag is embedded in the encrypted data through the tag encryption function TagEnc, and a tag tracking mechanism is set. The expression formula is: Among them, Enc qk (U j ) is the use of the public key qk to j The part that the user is authorized to access is encrypted, i represents the access control label of the i-th user, J i is the ciphertext data of the ith item, TagEnc qk (·) indicates that the public key qk is used to encrypt the label y i The result after encryption is For any ciphertext data, Indicates that there is a label extraction function that can extract the label from the ciphertext, L lbl (J i )=y i Indicates that the result of the extraction function is the original label y i .

10. The method for encrypting and managing financial loan information based on big data according to claim 8, characterized in that: In step S5, the method of integrating access control tags and operation log auditing mechanisms in the homomorphic computation process of the encryption model is as follows: The operation log audit mechanism, each homomorphic computation will be recorded as an operation log entry, denoted as: X t = <OP t ,J in ,J out ,y atie ,T t ′>, where X t The tth operation log entry represents a structured log record, OP t is the operation type, J in is the input ciphertext of this operation, J out is the output ciphertext after the operation, y atie Is the current operation visitor tag, T t ′ is the timestamp of the operation, which is used for audit and traceability verification. All operation logs are connected in a hash chain to form an unalterable log chain for audit. The expression formula is: Where h0 is the hash value of the 0th node in the log hash chain, is a secure hash function, X0 is the 0th operation log entry, h t is the hash value of the tth log record, X t is the log content of the t-th operation, h t-1 is the hash value of the t-1th log, Is user X t Use its private key Operation log X t Digital signature made, is the entire log hash chain, {h0,h1,...,h t } is the set of all hash values ​​from item 0 to item t. The encryption model generates the result ciphertext and tracking tag in each operation, and records the log hash at the same time. The expression formula is: Where V hm is the homomorphic encryption calculation function, J i ,J j Represents the two ciphertext data input, J c Indicates the output ciphertext, Enc qk (·) represents the encryption operation using the public key qk, f(U i ,U j ) represents the function operation on the plaintext, y c Represents the new label obtained after the operation, and || represents string concatenation.

Images