Cipher state data independence test method

Abstract

The invention relates to encryption operation, in particular to an operation method of encrypted data. The invention relates to a secret state data independence test method, which comprises the steps of user side encryption uploading, cloud computing and user result obtaining and decryption, and is characterized in that the encryption uploading comprises the steps of data arrangement, statistical result binarization, data encryption and secret state data sending to a cloud server, and the data is finally uploaded in a data table form. Comprising a serial number, a group and a type, the type data is 0 or 1, and the data is encrypted; the cloud computing comprises the steps of firstly establishing a connection list by using secret state data, then computing a theoretical frequency number, and finally respectively computing a statistical magnitude and a degree of freedom; and the user side decrypts the statistical magnitude and the degree-of-freedom data returned by the cloud side after obtaining the statistical magnitude and the degree-of-freedom data, looks up a table by using a decryption result, and judges whether the two variables are related according to a table lookup result. Data security of the user can be guaranteed, and the user can enjoy low-cost computing power provided by the cloud platform.

Description

Technical Field

[0001] The present invention relates to encryption operations, and in particular to a method for computing encrypted data. Background Art

[0002] Independence analysis is a crucial part of data analysis. It's simple and easy to use, and can quickly reveal potential relationships between data. The chi-square test is a commonly used test method, primarily used to analyze the independence of categorical variables.

[0003] With the rapid development of cloud computing, more and more companies are choosing to use the computing power provided by cloud platforms to complete data analysis. However, due to the uncontrollable nature of cloud platforms, many companies are worried about data leakage on cloud platforms, resulting in the loss of core corporate assets.

[0004] Homomorphic encryption technology supports computations on encrypted data and is suitable for addressing data leakage issues when using cloud platforms for data analysis. When using this technology, users encrypt their data locally to generate homomorphic ciphertext. The cloud platform then uses homomorphic operators to perform independence analysis based on the homomorphic ciphertext and sends the results to the user, who then decrypts them to obtain the final result. Summary of the Invention

[0005] The present invention aims to address the aforementioned shortcomings of the prior art by providing a method for verifying the independence of confidential data. From the moment data leaves the user until the user finally receives the result, no one else in the entire business process can access the original data, thereby ensuring data security and enabling users to benefit from the low-cost computing power provided by the cloud platform.

[0006] The present invention is implemented as follows: a method for verifying the independence of secret data, which includes user-side encrypted upload, cloud computing, and user obtaining and decrypting the results. The encrypted upload includes: arranging the data, binarizing the statistical results, and then encrypting the data. The secret data is sent to the cloud server. The final upload is in the form of a data table, including three columns: sequence number, group, and type. The type data is 0 or 1, and the data is encrypted.

[0007] Cloud computing involves first using encrypted data to build a linked list, then calculating theoretical frequencies, and finally calculating statistics and degrees of freedom respectively;

[0008] The client obtains the statistics and degrees of freedom data returned by the cloud, decrypts them, and uses the decrypted results to look up the table. Based on the results of the table lookup, it determines whether the two variables are correlated.

[0009] The method for testing the independence of secret data includes the following steps: the user can sort the statistical results of group 1g1 and group 2g2 according to his / her own testing needs. 1iand r 2i Binarization; when encrypting, the user encrypts the binarized data and the total number of each group, and sends the encryption result together with other data to the cloud server.

[0010] In the aforementioned method for verifying the independence of encrypted data, during the link table calculation phase, the cloud platform completes the creation of the link table based on the ciphertext. The link table is a table with four columns and four rows. ij is the observation frequency, R i For the line, C i is a column, and N is the total sample size.

[0011] The independence test method of the secret state data, after receiving the data, the cloud server first sums the results of all classes of group 1 to obtain s 11 , put it in (group 1, class 1), and then use the total number t1 of group 1 minus s11 to get s 12 Put it in (group 1, class 2). Calculate the total number of class 1 l1=s 11 +s 21 The other cells are calculated in a similar way. Finally, the totals of each column are calculated. After the calculation of the connected table is completed, the theoretical frequency E is calculated. ig , the calculation formula of theoretical frequency is:

[0012]

[0013] R in the formula i ·C j In the ciphertext state, it corresponds to t i ·l j ,,Directly use homomorphic multiplication of homomorphic encryption itself.

[0014] The aforementioned independence test method for encrypted data, the calculation method of the theoretical frequency in the plain text is:

[0015]

[0016] (O ij -E ij ) 2 In ciphertext state, it is replaced by (s ij -E ij ) 2 ; Addition, subtraction and multiplication in calculations are directly calculated using homomorphic addition and multiplication, and division is calculated using fractions;

[0017] The degrees of freedom are calculated as df = (r-1)(c-1), where R is the number of rows and c is the number of columns;

[0018] After the calculation is completed, the cloud server will 2 Sent to the user.

[0019] The method for verifying the independence of secret data, the user receives the result and decrypts χ 2 And look up the table, and determine whether the two variables are independent based on the table lookup results.

[0020] The homomorphic encryption technology of the present invention supports the calculation of encrypted data and is suitable for solving the problem of data leakage when using cloud platforms for data analysis. When using this technology, users need to complete data encryption locally to generate homomorphic ciphertext. The cloud platform uses homomorphic operators based on the homomorphic ciphertext to complete the independence analysis calculation and send the calculation results to the user, who decrypts the calculation results to obtain the final result. From the time the data leaves the user to the time the user finally receives the result, only the user can obtain the original data, thereby ensuring the user's data security and enjoying the cheap computing power provided by the cloud platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention will be further described below in conjunction with the accompanying drawings:

[0022] Figure 1 It is a calculation flow chart of the cloud server. DETAILED DESCRIPTION

[0023] See Figure 1 The invention discloses a method for verifying the independence of secret data, which includes encrypted uploading by a user end, cloud computing, and user obtaining results and decrypting them. The encrypted uploading includes: arranging data, binarizing statistical results, and then encrypting the data. The secret data is sent to a cloud server, and the final upload is in the form of a data table, including three columns: sequence number, group, and type. The type data is 0 or 1, and the data is encrypted. The cloud computing includes first using the secret data to establish a connected table, then calculating the theoretical frequency, and finally calculating the statistics and degrees of freedom respectively. The user end obtains the statistics and degrees of freedom data returned by the cloud, decrypts them, and uses the decrypted results to look up the table, and determines whether two variables are correlated based on the results of the table lookup.

[0024] The present invention is further described below with reference to research related to AIDS and gender:

[0025] When arranging data, users can combine the statistical results of group 1g1 and group 2g2 into statistical results according to their own testing needs. 1i and r 2i Binarization. For example, in a scenario where whether AIDS is related to gender, a user has collected data related to AIDS, including gender and AIDS status. In this case, the "yes" field for AIDS status can be binarized to 1, and the "no" field to 0. Here, the "AIDS status" field is called the class, and the gender field is called the group. The data table now looks like this:

[0026] Serial number Gender (group) Whether the disease (type) 1 g1 1 2 g2 0 … … … i … … … … … N … … .

[0027] During encryption, the user encrypts the binary data and the total number of each group, and sends the encryption result together with other data to the cloud server.

[0028] During the link list calculation phase, the cloud platform creates the link list based on the ciphertext. The link list is a table with four columns and four rows:

[0029] Table Typical Linked List

[0030] Class 1 Class 2 total Group 1 <![CDATA[O 11 ]]> <![CDATA[O 12 ]]> <![CDATA[R1]]> Group 2 <![CDATA[O 21 ]]> <![CDATA[O 22 ]]> <![CDATA[R2]]> total <![CDATA[C1]]> <![CDATA[C2]]> N ;

[0031] Among them, O ij represents the observation frequency, R i Indicates the total number of rows, C i represents the total number of columns, and N represents the total number of samples.

[0032] After receiving the data, the cloud server first sums the results of all classes in group 1 to obtain s 11 , put it in (group 1, class 1), and then use the total number t1 of group 1 minus s11 to get s 12 Put it in (group 1, class 2). Calculate the total number of class 1 l1=s 11 +s 21 The other cells are calculated in a similar way. Finally, the totals of each column are calculated to form the following table:

[0033] The connected list formed under the dense state calculation

[0034] Class 1 Class 2 total Group 1 <![CDATA[s 11 ]]> <![CDATA[s 12 =t1-s 11 ]]> <![CDATA[t1]]> Group 2 <![CDATA[s 21 ]]> <![CDATA[s 22 =t2-s 21 ]]> <![CDATA[t2]]> total <![CDATA[l1=s 11 +s 21 ]]> <![CDATA[l2=s 12 +s 22 ]]> <![CDATA[N=t1+t2]]> ;

[0035] In the scenario of determining whether HIV infection is related to gender, the corresponding linked list is as follows:

[0036] Table showing whether smoking is related to gender

[0037] Suffering from AIDS Not suffering from AIDS total male <![CDATA[s 11 ]]> <![CDATA[s 12 =t1-s 11 ]]> <![CDATA[t1]]> female <![CDATA[s 21 ]]> <![CDATA[s 22 =t2-s 21 ]]> <![CDATA[t2]]> total <![CDATA[l1=s 11 +s 21 ]]> <![CDATA[l2=s 12 +s 22 ]]> <![CDATA[t1+t2]]> ;

[0038] After the connected list is calculated, the theoretical frequency is calculated. The calculation formula for the theoretical frequency is:

[0039]

[0040] R in the formula i ·C j In the ciphertext state, it corresponds to t i ·lj , we can directly use the homomorphic multiplication of homomorphic encryption itself. In particular, since the homomorphic encryption algorithm does not support division calculation, the division calculation here is not actually performed. Instead, it is treated as a fraction, with the dividend as the numerator and the divisor as the denominator. The fraction calculation method is used in subsequent calculations, and the numerator and denominator are decrypted separately during the final decryption, and the division calculation is performed in the plaintext state. The calculation method for the theoretical frequency in the plaintext is:

[0041]

[0042] (O ij -E ij ) 2 In ciphertext state, it is replaced by (s ij -E ij ) 2 The addition, subtraction and multiplication in the calculation can be directly calculated using the same addition and multiplication method, and the division calculation still uses the fraction method.

[0043] The degree of freedom is calculated as df = (r-1)(c-1), where r is the number of rows and c is the number of columns. In the case of 2 rows and 2 columns, df = 1.

[0044] After the calculation is completed, the cloud server will 2 Sent to the user.

[0045] Finally, the user receives the result and decrypts χ 2 And look up the table, and determine whether the two variables are independent based on the table lookup results.

[0046] Cloud server calculation process reference Figure 1 The cloud server calculates the sum of the type fields of group 1 and group 2 from the received data and obtains s 11 =HeAdd(r 1i ) and s 21 =HeAdd(r 2i ),s 11 and s 21 As the data in (group 1, class 1) and (group 2, class 1) in the table respectively. In the second step, calculate the data s in (group 1, class 2) 12 =t1-s 11 and data s in (group 2, class 2) 22 =t2-s 21 , where t1 and t2 are the number of data in group 1 and the number of data in group 2. In the third step, the server calculates l1=s 11 +s 21 , l2=s 12 +s 22And N = t1 + t2. In the fourth stage, the server calculates E ij =(t i ·l j ) / N and χ 2 =∑(s ij -E ij ) 2 / E ij Ultimately, cloud servers will 2 and df is sent to the user.

Claims

1. A method for verifying the independence of secret data, which includes user-side encryption upload, cloud computing, and user obtaining and decrypting the results, is characterized by: The encrypted upload includes: arranging the data, binarizing the statistical results, and then encrypting the data, sending the encrypted data to the cloud server, and finally uploading the data in the form of a data table, including three columns: sequence number, group, and type. The type data is 0 or 1, and the data is encrypted; Cloud computing involves first using encrypted data to build a linked list, then calculating theoretical frequencies, and finally calculating statistics and degrees of freedom respectively; The client obtains the statistics and degrees of freedom data returned by the cloud, decrypts them, and uses the decrypted results to look up the table. Based on the results of the table lookup, it determines whether the two variables are correlated.

2. The method for verifying the independence of secret data according to claim 1, wherein: The data sorting includes the user sorting the statistical results of group 1g1 and group 2g2 according to their own detection needs. 1i and r 2i Binarization; when encrypting, the user encrypts the binarized data and the total number of each group, and sends the encryption result together with other data to the cloud server.

3. The method for verifying the independence of secret data according to claim 1, wherein: In the link table calculation phase, the cloud platform completes the establishment of the link table based on the ciphertext. The link table is a table with four columns and four rows. ij is the observation frequency, R i For the line, C i is a column, and N is the total sample size.

4. The method for verifying the independence of secret data according to claim 1, wherein: After receiving the data, the cloud server first sums the results of all classes in group 1 to obtain s 11 , put it in (group 1, class 1), and then use the total number t1 of group 1 minus s11 to get s 12 Put it in (group 1, class 2). Calculate the total number of class 1 l1=s 11 +s 21 The other cells are calculated in a similar way. Finally, the totals of each column are calculated. After the calculation of the connected table is completed, the theoretical frequency E is calculated. ig , the calculation formula of theoretical frequency is: R in the formula i ·C j In the ciphertext state, it corresponds to t i ·l j ,,Directly use homomorphic multiplication of homomorphic encryption itself.

5. The method for verifying the independence of secret data according to claim 1, wherein: The calculation method of theoretical frequency in plain text is: (O ij -E ij ) 2 In ciphertext state, it is replaced by (s ij -E ij ) 2 ; Addition, subtraction, and multiplication in calculations are directly performed using homomorphic addition and multiplication, while division is performed using fractions. The degrees of freedom are calculated as df = (r-1)(c-1), where r is the number of rows and c is the number of columns; After the calculation is completed, the cloud server will 2 Sent to the user.

6. The method for verifying the independence of secret data according to claim 1, wherein: The user receives the result and decrypts χ 2 And look up the table, and determine whether the two variables are independent based on the table lookup results.

Images