Security record linking method and system based on circuit privacy set intersection

Abstract

The present invention relates to a secure record linking method and system based on the intersection of circuit privacy sets, and belongs to the field of privacy protection data linking. The method of the present invention comprises: the first participant and the second participant preprocess the data to be screened, and integrate and reduce the dimension of the features; the first participant and the second participant bucket the data to be screened, and calculate the pseudo-random function value of the bucketed data based on the inadvertent pseudo-random function; the first participant and the second participant implement the screening of the target data based on the inadvertent pseudo-random function protocol, and obtain the secret sharing share of whether the match is obtained; the first participant and the second participant respectively weightedly calculate the screening score share based on the feature error rate and the secret sharing share of whether the feature matches; the first participant and the second participant perform recovery calculation based on the screening score share to obtain the screening score, and complete the link screening based on the threshold. The present invention improves the screening accuracy and security of the existing privacy protection data linking method.

Description

Technical Field

[0001] The invention relates to a secure record linking method and system for circuit privacy set intersection, belonging to the technical field of privacy protection record linking. Background Art

[0002] Private Set Intersection (PSI) is one of the most important privacy protection technologies currently available. This technology enables the participants to obtain the intersection of the data held by each party without leaking any additional information. Not leaking any additional information means that each party will not obtain any information other than the intersection of the data of each party. The use scenarios of private set intersection technology are relatively rich, such as aligning training data in vertical federated learning, and realizing friend discovery and recommendation through address books in social software. It has been widely used in many fields such as finance, medical care, transportation, and government affairs.

[0003] Circuit-PSI refers to a privacy set intersection method based on obfuscated circuits. This method can output the sharing values ​​of 0 and 1 in the form of intersection vectors to the participants (where 0 indicates that the calculated element is outside the intersection, and 1 indicates that the calculated element is within the intersection), but does not output the intersection data. This can better protect the privacy security of the data in the intersection and is applicable to scenarios with high privacy protection requirements, such as privacy-preserving location sharing, genomic sequence paternity testing under privacy protection, and infectious disease contact tracing. Circuit-PSI technology can be divided into balanced circuit-PSI and unbalanced circuit-PSI. Balanced circuit-PSI technology is applicable to scenarios where the data volume held by the participants is balanced, while unbalanced circuit-PSI is applicable to scenarios where the data volume held by the participants is disparate.

[0004] Secure record linking technology aims to complete data linking in a way that protects sensitive and personally identifiable information that may exist in data records from being leaked. During the record linking process, such sensitive and personally identifiable information may appear in the form of direct identifiers, such as name, email address, telephone number, medical record number, etc., as well as quasi-identifiers, such as date of birth, gender, zip code, etc. These identifiers play an important role in record linking, supporting the location of records corresponding to the same person in different data sources. However, due to its high sensitivity, the record linking process must be performed in a way that protects sensitive personal information from being identified. In this context, the rapid development of secure record linking technology has been promoted. Currently, there are many secure record linking technologies based on secret sharing protocols and homomorphic encryption technologies. However, in actual application, it is found that when the linking results are required to be protected, the linking efficiency is greatly reduced after adding privacy protection technology, and it increases exponentially with the increase in the collective amount of data of the participating parties and the increase in the dimension of data features, which makes it difficult to apply to massive data and complex data linking scenarios. In addition, when the data sets of two or more parties to link data records are inaccurate, data fields are missing, wrong or repeated, the accuracy of the link will be greatly affected. Summary of the invention

[0005] In order to overcome the shortcomings of the prior art, the present invention provides a secure record linking method and system based on circuit privacy set intersection. The privacy set intersection based on obfuscated circuits solves the problem that both non-linked data and linked results require privacy protection, and designs a scoring function based on error rate to solve the problem of low screening accuracy in scenarios where the original data is missing, erroneous, and repeated.

[0006] The technical solution of the present invention is: a secure record linking method based on circuit privacy set intersection, comprising:

[0007] S1. The first participant and the second participant pre-process the data to be screened and integrate and reduce the dimensions of the features;

[0008] S2. The first participant and the second participant divide the data to be screened into buckets, and calculate pseudo-random function values ​​for the bucketed data based on an inadvertent pseudo-random function;

[0009] S3. The first participant and the second participant screen the target data based on the inadvertent pseudo-random function protocol to obtain a secret sharing share of whether the features match;

[0010] S4. The first participant and the second participant respectively weight and calculate screening score shares based on the feature error rate and the secret sharing share of whether the features match;

[0011] S5. The first party and the second party perform recovery calculation based on the screening score share to obtain a screening score, and complete the linked screening based on the threshold.

[0012] As a further solution of the present invention, in S1, the data to be screened are preprocessed, and the features are integrated and reduced in dimension, specifically including:

[0013] S11. The preprocessing of the data to be screened refers to checking the data to be screened and standardizing the missing, wrong and format problems of the data;

[0014] S12. The integration and dimensionality reduction of features refers to that the first party and the second party adopt the same rules to analyze multiple feature dimensions of each data and compress and integrate features with high correlation and small impact.

[0015] As a further solution of the present invention, in S2, the first participant and the second participant bucket the data to be screened, and calculate a pseudo-random function value for the bucketed data based on an inadvertent pseudo-random function, specifically including:

[0016] S21. The first participant and the second participant divide the data to be screened into buckets, specifically:

[0017] The first party uses a hash function , and , the first party to be screened is sent to the Data of The feature data is hashed into a cuckoo hash table, wherein the cuckoo hash table includes Buckets, specifically The value of The actual value of is determined to ensure that no hash collision occurs;

[0018] The second participant sends the second participant the first Data of The features are separated by hash functions , and Hash it to , and In the corresponding bucket, Indicates the second party to be screened Data No. Features , and Respectively represent the use of hash functions , and Calculate the data to be screened The hash value of

[0019] S22. The first participant and the second participant calculate a pseudo-random function value for the bucketed data based on an inadvertent pseudo-random function, specifically including:

[0020] The first participant to be screened The data is ,in i ∈[1,⋯, N 1 ] , Indicates the number of data to be screened held by the first participant, Represents the characteristic dimension of each piece of data; the data to be screened by the second participant is ,in Indicates the second participant to be screened Items of data, Indicates the number of data to be screened held by the second participant, each of which contains Characteristics of dimensions;

[0021] The first participant passes the first participant to be screened Data Input the random pseudo-random function protocol to generate the key vector , and the corresponding pseudo-random function , and Sending to the second party;

[0022] The first participant and the second participant are respectively based on a pseudo-random function Calculate the corresponding pseudo-random function value, where represents the independent variable of the pseudorandom function, Indicates a randomly generated key, the first participant based on a pseudo-random function Calculate the first participant to be screened Data The pseudo-random function value is , the second participant based on a pseudo-random function Calculate the data to be screened by the second party The pseudo-random function value is .

[0023] As a further solution of the present invention, in S3, the first participant and the second participant implement the screening of the target data based on the inadvertent pseudo-random function protocol to obtain the secret sharing share of whether the features match, which specifically includes:

[0024] S31. The first participant and the second participant implement the screening of target data based on the inadvertent pseudo-random function protocol, specifically:

[0025] The second participant randomly selects dimensional random vector , Represents a random number, and for , and The corresponding data calculation in the three buckets , and sends the calculation result to the first participant, , and Indicates that the second participant to be screened Data Through hash function , and Hash it into 3 buckets, represents the rth data of the jth data vector of the second participant;

[0026] The first participant calculates the pseudo-random function value of the data in the bucket XOR calculation with the data sent by the second participant , generate bucket vector;

[0027] The first participant inputs the first The first data The bucket vector value of the feature ,in Indicates that the first participant The first data The pseudo-random function value of the features, Indicates that the second party The first data The pseudo-random function value of the features, Indicates that the second party enters The data corresponding to A random vector;

[0028] The second party enters The data corresponding to A random vector , the first participant is realized by XOR calculation The first data The first characteristic of all data of the second party Screening of features and obtaining matching circuit password shares respectively;

[0029] S32. The first party and the second party obtain a secret sharing share of whether the characteristics match, which means;

[0030] when When the first participant and the second participant obtain and The matching circuit secret sharing share is obtained, otherwise the unmatched circuit secret sharing share is obtained, and the circuit secret sharing share includes the first participant's screening score share , Second Party Screening Score Share .

[0031] As a further solution of the present invention, in S4, the first participant and the second participant respectively weightedly calculate the screening score share based on the feature error rate and the secret sharing share of whether the features match, specifically including:

[0032] S41. The error rate of the data characteristics corresponding to the first participant and the second participant refers to the error rate of a certain proportion of the same data when it is collected and stored by different participants. The error rate of a feature is ;

[0033] S42. The first participant and the second participant respectively weightedly calculate a screening score function based on the error rate of the corresponding data feature and the matching secret sharing share, specifically:

[0034] The first participant weights the The data corresponds to the second party Screening score share of data ,in Indicates the first party's The data of the second party The screening score share of the data, The calculation is to calculate the first party The data corresponding to The features are screened and calculated in step S3 to obtain screening score share and Screening score share multiplied by The error rate is calculated;

[0035] Similarly, the second participant calculates the weighted The data corresponds to the second party Screening score share of data ,in Indicates that the second party holds the first party's The data of the second party The screening score share of the data, is calculated by adding the second party The data corresponding to The features are screened and calculated in step S3 to obtain screening score share and Screening score share multiplied by The error rate is calculated.

[0036] As a further solution of the present invention, in S5, the first participant and the second participant perform recovery calculation based on the screening score share to obtain the screening score, and complete the link screening based on the threshold, specifically including:

[0037] S51. The first participant and the second participant perform a recovery calculation based on the screening score share to obtain a screening score, which means that the first participant performs a recovery calculation based on the screening score share. and the second party based on Through secret recovery calculation, the first participant's The data and the second party The filter score of the data ;

[0038] S52. The threshold refers to the threshold of the screening score. The value range is ;

[0039] S53. The link screening based on the threshold means that when the score exceeds or equals the threshold, it indicates that the data of the first participant matches the data of the second participant, otherwise it does not match; specifically, when , then it means that the first participant The data and the second party If the data matches, the link is successful, otherwise it does not match.

[0040] The present invention also provides a secure record linking system based on circuit privacy set intersection, comprising a module for executing the secure record linking method based on circuit privacy set intersection.

[0041] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned secure record linking method based on circuit privacy set intersection when executing the program.

[0042] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned secure record linking method based on circuit privacy set intersection is implemented.

[0043] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned secure record linking method based on circuit privacy set intersection.

[0044] The beneficial effects of the present invention are as follows: the present invention solves the problem of low screening and matching accuracy in existing privacy-preserving data linking methods when the original data is missing, erroneous, and duplicated, as well as the problem of weak privacy protection for intermediate screening results. The present invention ensures the privacy protection of screening data, screening process results, and final results based on the intersection of circuit privacy sets, and can be applied to a variety of multi-feature data screening scenarios. A scoring function is designed based on the error rate, which can improve the accuracy of screening links for multi-party data linking scenarios with data errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 It is a flow chart of a secure record linking method based on circuit privacy set intersection provided by the present invention;

[0047] Figure 2 It is a module connection diagram of a secure record linking system based on circuit privacy set intersection provided by the present invention;

[0048] Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention.

[0049] Reference numerals:

[0050] 110: data integration processing module; 120: pseudo-random function calculation module; 130: matching module; 140: calculation module; 150: data screening module

[0051] 310: processor; 320: communication interface; 330: memory; 340: communication bus. DETAILED DESCRIPTION

[0052] Example 1: Figure 1-Figure 3 As shown, the secure record linking method based on circuit privacy set intersection specifically includes:

[0053] S1. The first and second parties pre-process the data to be screened and integrate and reduce the dimensions of the features; specifically, the following steps are performed:

[0054] S11. The pre-processing of the data to be screened refers to the first participant and the second participant comprehensively checking the data to be screened, and both parties standardize the data for obvious omissions, errors and formats, so as to reduce the impact on the accuracy of the final screening link;

[0055] S12. The integration and dimensionality reduction of features refers to that the first participant and the second participant adopt the same rules to analyze multiple feature dimensions of each piece of data, compress and integrate features with high correlation and small impact, and consider not using features with high error rates as screening matching features during screening, thereby reducing the number of comparisons for each piece of data;

[0056] S2. The first participant and the second participant divide the data to be screened into buckets, and calculate pseudo-random function values ​​for the bucketed data based on an inadvertent pseudo-random function; specifically, including:

[0057] S21. The first participant and the second participant divide the data to be screened into buckets, specifically:

[0058] The first party uses a hash function , and , the first party to be screened is sent to the Data of The feature data is hashed into a cuckoo hash table, wherein the cuckoo hash table includes Buckets, specifically The value of The actual value of is determined to ensure that no hash collision occurs and that each piece of data held by the participant can be mapped to the corresponding bucket;

[0059] The second participant sends the second participant the first Data of The features are separated by hash functions , and Hash it to , and In the corresponding bucket, Indicates the second party to be screened Data No. Features , and Respectively represent the use of hash functions , and Calculate the data to be screened The hash value of

[0060] S22. The first participant and the second participant calculate a pseudo-random function value for the bucketed data based on an inadvertent pseudo-random function, specifically including:

[0061] The first participant is to be screened The data is ,in i ∈[1,⋯, N 1 ] , Indicates the number of data to be screened held by the first participant, Represents the characteristic dimension of each piece of data; the data to be screened by the second participant is ,in Indicates the second participant to be screened Items of data, Indicates the number of data to be screened held by the second participant, each of which contains Characteristics of dimensions;

[0062] The first participant passes the first participant to be screened Data Input the random pseudo-random function protocol to generate the key vector , and the corresponding pseudo-random function , and Sending to the second party;

[0063] The first participant and the second participant are respectively based on a pseudo-random function Calculate the corresponding pseudo-random function value, where represents the independent variable of the pseudorandom function, Indicates a randomly generated key, the first participant based on a pseudo-random function Calculate the first participant to be screened Data The pseudo-random function value is , the second participant based on a pseudo-random function Calculate the data to be screened by the second party The pseudo-random function value is .

[0064] S3. The first participant and the second participant screen the target data based on the inadvertent pseudo-random function protocol to obtain a secret sharing share of whether the features match; specifically, including:

[0065] S31. The first participant and the second participant implement the screening of target data based on the inadvertent pseudo-random function protocol, specifically:

[0066] The second participant randomly selects dimensional random vector , Represents a random number, and for , and The corresponding data calculation in the three buckets , and sends the calculation result to the first participant, , and Indicates that the second participant to be screened Data Through hash function , and Hash it into 3 buckets, represents the rth data of the jth data vector of the second participant;

[0067] The first participant calculates the pseudo-random function value of the data in the bucket XOR calculation with the data sent by the second participant , generate bucket vector;

[0068] The first participant inputs the first The first data The bucket vector value of the feature ,in Indicates that the first participant The first data The pseudo-random function value of the features, Indicates that the second party The first data The pseudo-random function value of the features, Indicates that the second party enters The data corresponding to A random vector;

[0069] The second party enters The data corresponding to A random vector , the first participant is realized by XOR calculation The first data The first characteristic of all data of the second party Screening of features and obtaining matching circuit password shares respectively;

[0070] S32. The first party and the second party obtain a secret sharing share of whether the characteristics match, which means;

[0071] when When the first participant and the second participant obtain and The matching circuit secret sharing share is obtained, otherwise the unmatched circuit secret sharing share is obtained, and the circuit secret sharing share includes the first participant's screening score share , Second Party Screening Score Share .

[0072] S4. The first participant and the second participant respectively weight and calculate the screening score share based on the feature error rate and the secret sharing share of whether the features match; specifically, including:

[0073] S41. The error rate of the data characteristics corresponding to the first participant and the second participant refers to the error rate of a certain proportion of the same data when it is collected and stored by different participants. The error rate of a feature is ;

[0074] S42. The first participant and the second participant respectively weightedly calculate a screening score function based on the error rate of the corresponding data feature and the matching secret sharing share, specifically:

[0075] The first participant weights the The data corresponds to the second party Screening score share of data ,in Indicates the first party's The data of the second party The screening score share of the data, The calculation is to calculate the first party The data corresponding to The features are screened and calculated in step S32 to obtain screening score share and Screening score share multiplied by The error rate is calculated;

[0076] Similarly, the second participant calculates the weighted The data corresponds to the second party Screening score share of data ,in Indicates that the second party holds the first party's The data of the second party The screening score share of the data, is calculated by adding the second party The data corresponding to The features are screened and calculated in step S32 to obtain screening score share and Screening score share multiplied by The error rate is calculated.

[0077] S5. The first and second participants perform recovery calculation based on the screening score share to obtain the screening score, and complete the link screening based on the threshold. Specifically, it includes:

[0078] S51. The first participant and the second participant perform a recovery calculation based on the screening score share to obtain a screening score, which means that the first participant performs a recovery calculation based on the screening score share. and the second party based on Through secret recovery calculation, the first participant's The data and the second party The filter score of the data ;

[0079] S52. The threshold refers to the threshold of the screening score. The value range is ;

[0080] S53. The link screening based on the threshold means that when the score exceeds or equals the threshold, it indicates that the data of the first participant matches the data of the second participant, otherwise it does not match; specifically, when , then it means that the first participant The data and the second party If the data matches, the link is successful, otherwise it does not match.

[0081] like Figure 2 As shown, the present invention also provides a secure record linking system based on circuit privacy set intersection, comprising:

[0082] The data integration processing module 110 is used by the first participant and the second participant to pre-process the data to be screened and integrate and reduce the dimension of the features;

[0083] The pseudo-random function value calculation module 120 is used for the first participant and the second participant to bucket the data to be screened, and calculate the pseudo-random function value of the bucketed data based on the inadvertent pseudo-random function;

[0084] A matching module 130 is used for the first participant and the second participant to screen the target data based on an inadvertent pseudo-random function protocol to obtain a secret sharing share of whether the features match;

[0085] A calculation module 140, configured for the first participant and the second participant to respectively weight and calculate screening score shares based on the feature error rate and the secret sharing share of whether the features match;

[0086] The data screening module 150 is used for the first participant and the second participant to perform recovery calculation based on the screening score share to obtain the screening score, and complete the link screening based on the threshold.

[0087] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned secure record linking method based on circuit privacy set intersection when executing the program.

[0088] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned secure record linking method based on circuit privacy set intersection is implemented.

[0089] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned secure record linking method based on circuit privacy set intersection.

[0090] In order to illustrate the effect of the present invention, the present invention conducted a comparative experiment to verify the improvement of the accuracy and safety of the method of the present invention.

[0091] The first participant holds 1.1 million data items; the second participant holds 1.2 million data items; the two data sets have 15 features respectively, 20% of the data in the two data sets are repeated, and except for 3 features without errors, the other 12 features may have different error rates (5%~25%) and different degrees of missing values.

[0092] The data of the above embodiment are respectively screened under plain text conditions, traditional link screening based on secure multi-party computing, and screening based on the method of the present invention, and the results are shown in Table 1. The link method under plain text conditions needs to perform aggregation links, does not protect privacy, and the screening accuracy is 100%; the traditional link screening method based on secure multi-party computing does not consider the real data and will have a certain error rate, which will reduce the accuracy of the link. The accuracy is only 87.37% under the above data holding situation and error rate scenario; the method of the present invention is based on the intersection of circuit privacy sets and the design of error score function. Under the premise of ensuring data privacy security, the accuracy of the link is improved, and the accuracy can reach 99.53% under the above data holding situation and error rate scenario.

[0093] Table 1: Link accuracy of different methods

[0094]

[0095] The specific implementation modes of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above implementation modes, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A secure record linking method based on circuit privacy set intersection, characterized in that: include: S1. The first participant and the second participant pre-process the data to be screened and integrate and reduce the dimensions of the features; S2. The first participant and the second participant divide the data to be screened into buckets, and calculate pseudo-random function values ​​for the bucketed data based on an inadvertent pseudo-random function; S3. The first participant and the second participant screen the target data based on the inadvertent pseudo-random function protocol to obtain a secret sharing share of whether the features match; S4. The first participant and the second participant respectively weight and calculate screening score shares based on the feature error rate and the secret sharing share of whether the features match; S5. The first participant and the second participant perform recovery calculation based on the screening score share to obtain a screening score, and complete the linked screening based on the threshold; In S4, the first participant and the second participant respectively weightedly calculate screening score shares based on the feature error rate and the secret sharing share of whether the features match, specifically including: S41. The error rate of the data characteristics corresponding to the first participant and the second participant refers to the error rate of a certain proportion of the same data when it is collected and stored by different participants. The error rate of a feature is ; S42. The first participant and the second participant respectively weightedly calculate a screening score function based on the error rate of the corresponding data feature and the matching secret sharing share, specifically: The first participant weights the The data corresponds to the second party Screening score share of data ,in Indicates the first party's The data of the second party The screening score share of the data, The calculation is to calculate the first party The data corresponding to The features are screened and calculated in step S3 to obtain screening score share and Screening score share multiplied by The error rate is calculated; Similarly, the second participant calculates the weighted The data corresponds to the second party Screening score share of data ,in Indicates that the second party holds the first party's The data of the second party The screening score share of the data, is calculated by adding the second party The data corresponding to The features are screened and calculated in step S3 to obtain screening score share and Screening score share multiplied by The error rate is calculated.

2. The secure record linking method based on circuit privacy set intersection according to claim 1 is characterized in that: In S1, the data to be screened is preprocessed, and the features are integrated and reduced in dimension, specifically including: S11. The preprocessing of the data to be screened refers to checking the data to be screened and standardizing the missing, wrong and format problems of the data; S12. The integration and dimensionality reduction of features refers to that the first participant and the second participant adopt the same rules to analyze multiple feature dimensions of each data and compress and integrate features with high correlation and small impact.

3. The secure record linking method based on circuit privacy set intersection according to claim 1 is characterized in that: In S2, the first participant and the second participant bucket the data to be screened, and calculate pseudo-random function values ​​for the bucketed data based on an inadvertent pseudo-random function, specifically including: S21. The first participant and the second participant divide the data to be screened into buckets, specifically: The first party uses a hash function , and , the first party to be screened is sent to the Data of The feature data is hashed into a cuckoo hash table, wherein the cuckoo hash table includes Buckets, specifically The value of The actual value of is determined to ensure that no hash collision occurs; The second participant sends the second participant the first Data of The features are separated by hash functions , and Hash it to , and In the corresponding bucket, Indicates the second party to be screened Data No. Features , and Respectively represent the use of hash functions , and Calculate the data to be screened The hash value of S22. The first participant and the second participant calculate a pseudo-random function value for the bucketed data based on an inadvertent pseudo-random function, specifically including: The first participant is to be screened The data is ,in , Indicates the number of data to be screened held by the first participant, Represents the characteristic dimension of each piece of data; the data to be screened by the second participant is ,in Indicates the second participant to be screened Items of data, Indicates the number of data to be screened held by the second participant, each of which contains Characteristics of dimensions; The first participant passes the first participant to be screened Data Input the random pseudo-random function protocol to generate the key vector , and the corresponding pseudo-random function , and Sending to the second party; The first participant and the second participant are respectively based on a pseudo-random function Calculate the corresponding pseudo-random function value, where represents the independent variable of the pseudorandom function, Indicates a randomly generated key, the first participant based on a pseudo-random function Calculate the first participant to be screened Data The pseudo-random function value is , the second participant based on a pseudo-random function Calculate the data to be screened by the second party The pseudo-random function value is .

4. The secure record linking method based on circuit privacy set intersection according to claim 1 is characterized in that: In S3, the first participant and the second participant implement the screening of the target data based on the inadvertent pseudo-random function protocol to obtain the secret sharing share of whether the features match, which specifically includes: S31. The first participant and the second participant implement the screening of target data based on the inadvertent pseudo-random function protocol, specifically: The second participant randomly selects dimensional random vector , Represents a random number, and for , and The corresponding data calculation in the three buckets , and sends the calculation result to the first participant, , and Indicates that the second participant to be screened Data Through hash function , and Hash it into 3 buckets, represents the rth data of the jth data vector of the second participant; The first participant calculates the pseudo-random function value of the data in the bucket XOR calculation with the data sent by the second participant , generate bucket vector; The first participant inputs the first The first data The bucket vector value of the feature ,in Indicates that the first participant The first data The pseudo-random function value of the features, Indicates that the second party The first data The pseudo-random function value of the features, Indicates that the second party enters The data corresponding to A random vector; The second party enters The data corresponding to A random vector , the first participant is realized by XOR calculation The first data The first characteristic of all data of the second party Screening of features and obtaining matching circuit password shares respectively; S32. The first participant and the second participant obtain a secret sharing share of whether the characteristics match, which means: when When the first participant and the second participant obtain and The matching circuit secret sharing share is obtained, otherwise the unmatched circuit secret sharing share is obtained, and the circuit secret sharing share includes the first participant's screening score share , Second Party Screening Score Share .

5. The secure record linking method based on circuit privacy set intersection according to claim 1 is characterized in that: In S5, the first participant and the second participant perform recovery calculation based on the screening score share to obtain a screening score, and complete link screening based on a threshold, specifically including: S51. The first participant and the second participant perform a recovery calculation based on the screening score share to obtain a screening score, which means that the first participant performs a recovery calculation based on the screening score share. and the second party based on Through secret recovery calculation, the first participant's The data and the second party The filter score of the data ; S52. The threshold refers to the threshold of the screening score. The value range is ; S53. The link screening based on the threshold means that when the score exceeds or equals the threshold, it indicates that the data of the first participant matches the data of the second participant, otherwise it does not match; specifically, when , then it means that the first participant The data and the second party If the data matches, the link is successful, otherwise it does not match.

6. A secure record linking system based on circuit privacy set intersection, characterized in that: include: A module for executing the secure record linking method based on circuit privacy set intersection as claimed in any one of claims 1 to 5.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the secure record linking method based on circuit privacy set intersection as described in any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the secure record linking method based on circuit privacy set intersection as claimed in any one of claims 1 to 5 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the secure record linking method based on circuit privacy set intersection as claimed in any one of claims 1 to 5 is implemented.

Images