Data management system based on privacy computing

Abstract

The invention discloses a data management system based on privacy computing, and relates to the field of data management. Comprising a data receiving module for receiving a data sharing request and performing access authentication; the data preprocessing module is used for acquiring shared data based on the data sharing request and preprocessing the shared data; the privacy security calculation module is used for carrying out privacy security calculation on the preprocessed data based on a CATR decision tree; the cloud service module is used for sharing the calculated data to a user in a cloud service mode; and the data sharing detection module is used for detecting the data sharing process through a data sharing detection method. In the data preprocessing stage, data are encrypted and anonymized through Huffman coding and a center point clustering algorithm, so that the privacy and security of the data are ensured; the Huffman coding can reduce the storage space of the data by compressing the data, and meanwhile, the data is encrypted, so that the confidentiality of the data is protected.

Description

Technical Field

[0001] The present invention relates to the field of data management, and in particular to a data management system based on privacy computing. Background Art

[0002] In the current wave of digital transformation, data has become an important asset for enterprises and individuals. However, with the widespread sharing and transmission of data, the risk of data leakage and privacy violation has also increased.

[0003] Privacy computing allows multiple parties to jointly complete a computing task while protecting the privacy of their respective data, thereby achieving the goal of "data available but invisible". It involves multiple disciplines such as cryptography, security hardware, information theory, and distributed computing, and uses a variety of privacy protection technologies and privacy enhancement technologies to achieve encrypted storage and transmission of data, ensuring the security of data during processing and transmission.

[0004] Traditional data management systems mostly rely on simple encryption technology or access control mechanisms, which often fail to provide effective protection against advanced hacker attacks or insider leaks.

[0005] Therefore, a more advanced and comprehensive data management system is needed to ensure the security and privacy of data during sharing and transmission.

[0006] To this end, the present invention proposes a data management system based on privacy computing. Summary of the invention

[0007] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a data management system based on privacy computing.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] The data management system based on privacy computing includes:

[0010] A data receiving module, which receives a data sharing request and performs access authentication;

[0011] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0012] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0013] Cloud service module, which shares the calculated data with users through cloud services;

[0014] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0015] Preferably: the working method of the data receiving module comprises the following steps:

[0016] A1: Receive data sharing requests through the data sharing platform;

[0017] A2: Obtain the user's identity information based on the data sharing request;

[0018] A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

[0019] Preferably: the working method of the data preprocessing module comprises the following steps:

[0020] B1: extracting the shared data to be shared according to the user's data sharing request;

[0021] B2: Encrypt the shared data using the Huffman method and obtain the encrypted data;

[0022] B3: Use the center point clustering algorithm to anonymize the encrypted data.

[0023] Preferably: the step B2 comprises the following steps:

[0024] B21: Analyze the shared data and count the number of occurrences of each ASCII character;

[0025] B22: Use the number of occurrences of each ASCII character as a weight to construct a Huffman tree;

[0026] B23: Construct Huffman coding based on Huffman tree;

[0027] B24: compress the data to be shared according to Huffman coding to form encrypted data;

[0028] In step B23, the calculation formula of the average code length of Huffman coding is lavg=∑(W i *L i ) / ∑W i , where W i is the weight, which is the number of times the ASCII character appears.

[0029] Preferably: the step B3 comprises the following steps:

[0030] B31: Calculate the correlation between attribute columns in the shared data based on Cramer's correlation coefficient;

[0031] B32: Find the attribute column with the highest correlation with the sensitive attribute column and merge it with the sensitive attribute column into the same set;

[0032] B33: Calculate the silhouette coefficient of each attribute, and obtain the optimal number of divisions and the division results according to the maximum silhouette coefficient value;

[0033] B34: Use generalization technology to generalize each cluster and obtain anonymous data.

[0034] Preferably: in step B31, the formula for calculating the correlation is: Where W i is the attribute weight, d i (x, y) is the distance between x and y on attribute i.

[0035] Preferably: in the step B33, the calculation formula of the silhouette coefficient is s(i)=[b(i)-a(i) / max{a(i), b(i)}], where s(i) is the silhouette coefficient of the i-th sample, a(i) is the average distance between the i-th sample and other samples in its cluster, and b(i) is the average distance between the i-th sample and all samples in the nearest cluster.

[0036] Preferably: the calculation method of the privacy security calculation module includes the following steps:

[0037] C1: Set the differential privacy budget;

[0038] C2: Allocate the budget to each level of the decision tree according to the arithmetic difference privacy budget allocation mechanism, and calculate the privacy budget of each node in each level;

[0039] C3: Divide the privacy budget allocated to each node into two equal parts;

[0040] C4: If the data on the node meets the preset stop condition, the node is set as a leaf node and the creation of sub-nodes is stopped. Then, the Laplace mechanism is used to add noise to the count value of the node, and the most common classification in the sample is selected as the label of the leaf node;

[0041] C5: If the data on the node does not meet the stopping condition, the number of data sets contained in the node is calculated and noise is added to it using the Laplace mechanism.

[0042] Preferably: in the C4 step, the formula resulting from the addition is Noise＝Lap(0 , b / ε), where b is the sensitivity and ε is the privacy budget.

[0043] The beneficial effects of the present invention are:

[0044] 1. In the data preprocessing stage of the present invention, Huffman coding and center point clustering algorithm are used to encrypt and anonymize the data to ensure the privacy and security of the data; Huffman coding reduces the storage space of the data by compressing the data, and encrypts the data to protect the confidentiality of the data; the center point clustering algorithm hides sensitive information in the data through anonymization processing, making it difficult for attackers to extract sensitive information from the data; in addition, the combination of Huffman coding and center point clustering algorithm further enhances the privacy protection capability of the data.

[0045] 2. The privacy security calculation stage of the present invention uses the CATR decision tree to process data. By setting the differential privacy budget and allocation mechanism, it ensures that sensitive information will not be leaked during the data processing process. The differential privacy technology protects privacy by adding noise to the data. The amount of noise added is adjusted according to the privacy budget, achieving a balance between privacy protection and data accuracy. The introduction of the CATR decision tree enables the system to effectively analyze and utilize data while protecting privacy.

[0046] 3. In the data sharing stage of the present invention, the processed data is shared with users through cloud services, which is convenient for users to analyze and use the data; in the data sharing detection stage, the security and compliance of the data in the sharing process are ensured by monitoring the data usage, access rights and data flow; the implementation of these detection measures helps to timely discover and deal with potential security risks and ensure the privacy and security of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is the architecture diagram of the data management system based on privacy computing proposed by the present invention;

[0048] Figure 2 This is a flow chart of the working method of the data receiving module of the data management system based on privacy computing proposed by the present invention;

[0049] Figure 3 This is a flow chart of the working method of the data preprocessing module of the data management system based on privacy computing proposed by the present invention;

[0050] Figure 4 This is a flow chart of the calculation method of the privacy security calculation module of the data management system based on privacy computing proposed by the present invention. DETAILED DESCRIPTION

[0051] The technical solution of the present invention is further described in detail below in conjunction with specific implementation methods.

[0052] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", and "set" should be understood in a broad sense, for example, they can be fixedly connected or set, or detachably connected or set, or integrally connected or set. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0053] Embodiment 1:

[0054] The data management system based on privacy computing includes:

[0055] A data receiving module, which receives a data sharing request and performs access authentication;

[0056] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0057] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0058] Cloud service module, which shares the calculated data with users through cloud services;

[0059] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0060] Embodiment 2:

[0061] The data management system based on privacy computing includes:

[0062] A data receiving module, which receives a data sharing request and performs access authentication;

[0063] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0064] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0065] Cloud service module, which shares the calculated data with users through cloud services;

[0066] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0067] The working method of the data receiving module comprises the following steps:

[0068] A1: Receive data sharing requests through the data sharing platform;

[0069] A2: Obtain the user's identity information based on the data sharing request;

[0070] A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

[0071] Embodiment 3:

[0072] The data management system based on privacy computing includes:

[0073] A data receiving module, which receives a data sharing request and performs access authentication;

[0074] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0075] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0076] Cloud service module, which shares the calculated data with users through cloud services;

[0077] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0078] The working method of the data receiving module comprises the following steps:

[0079] A1: Receive data sharing requests through the data sharing platform;

[0080] A2: Obtain the user's identity information based on the data sharing request;

[0081] A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

[0082] The working method of the data preprocessing module comprises the following steps:

[0083] B1: extracting the shared data to be shared according to the user's data sharing request;

[0084] B2: Encrypt the shared data using the Huffman method and obtain the encrypted data;

[0085] B3: Use the center point clustering algorithm to anonymize the encrypted data.

[0086] Embodiment 4:

[0087] The data management system based on privacy computing includes:

[0088] A data receiving module, which receives a data sharing request and performs access authentication;

[0089] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0090] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0091] Cloud service module, which shares the calculated data with users through cloud services;

[0092] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0093] The working method of the data receiving module comprises the following steps:

[0094] A1: Receive data sharing requests through the data sharing platform;

[0095] A2: Obtain the user's identity information based on the data sharing request;

[0096] A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

[0097] The working method of the data preprocessing module comprises the following steps:

[0098] B1: extracting the shared data to be shared according to the user's data sharing request;

[0099] B2: Encrypt the shared data using the Huffman method and obtain the encrypted data;

[0100] B3: Use the center point clustering algorithm to anonymize the encrypted data.

[0101] The step B2 comprises the following steps:

[0102] B21: Analyze the shared data and count the number of occurrences of each ASCII character;

[0103] B22: Use the number of occurrences of each ASCII character as a weight to construct a Huffman tree;

[0104] B23: Construct Huffman coding based on Huffman tree;

[0105] B24: Compress the data to be shared according to Huffman coding to form encrypted data.

[0106] In step B23, the calculation formula of the average code length of Huffman coding is lavg=∑(W i *L i ) / ∑W i , where W iis the weight, which is the number of times the ASCII character appears.

[0107] Embodiment 5:

[0108] The data management system based on privacy computing includes:

[0109] A data receiving module, which receives a data sharing request and performs access authentication;

[0110] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0111] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0112] Cloud service module, which shares the calculated data with users through cloud services;

[0113] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0114] The working method of the data receiving module comprises the following steps:

[0115] A1: Receive data sharing requests through the data sharing platform;

[0116] A2: Obtain the user's identity information based on the data sharing request;

[0117] A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

[0118] The working method of the data preprocessing module comprises the following steps:

[0119] B1: extracting the shared data to be shared according to the user's data sharing request;

[0120] B2: Encrypt the shared data using the Huffman method and obtain the encrypted data;

[0121] B3: Use the center point clustering algorithm to anonymize the encrypted data.

[0122] The step B2 comprises the following steps:

[0123] B21: Analyze the shared data and count the number of occurrences of each ASCII character;

[0124] B22: Use the number of occurrences of each ASCII character as a weight to construct a Huffman tree;

[0125] B23: Construct Huffman coding based on Huffman tree;

[0126] B24: Compress the data to be shared according to Huffman coding to form encrypted data.

[0127] In step B23, the calculation formula of the average code length of Huffman coding is lavg=∑(W i *L i ) / ∑W i , where W i is the weight, which is the number of times the ASCII character appears.

[0128] The step B3 comprises the following steps:

[0129] B31: Calculate the correlation between attribute columns in the shared data based on Cramer's correlation coefficient;

[0130] B32: Find the attribute column with the highest correlation with the sensitive attribute column and merge it with the sensitive attribute column into the same set;

[0131] B33: Calculate the silhouette coefficient of each attribute, and obtain the optimal number of divisions and the division results according to the maximum silhouette coefficient value;

[0132] B34: Use generalization technology to generalize each cluster and obtain anonymous data.

[0133] Embodiment 6:

[0134] The data management system based on privacy computing includes:

[0135] A data receiving module, which receives a data sharing request and performs access authentication;

[0136] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0137] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0138] Cloud service module, which shares the calculated data with users through cloud services;

[0139] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0140] The working method of the data receiving module comprises the following steps:

[0141] A1: Receive data sharing requests through the data sharing platform;

[0142] A2: Obtain the user's identity information based on the data sharing request;

[0143] A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

[0144] The working method of the data preprocessing module comprises the following steps:

[0145] B1: extracting the shared data to be shared according to the user's data sharing request;

[0146] B2: Encrypt the shared data using the Huffman method and obtain the encrypted data;

[0147] B3: Use the center point clustering algorithm to anonymize the encrypted data.

[0148] The step B2 comprises the following steps:

[0149] B21: Analyze the shared data and count the number of occurrences of each ASCII character;

[0150] B22: Use the number of occurrences of each ASCII character as a weight to construct a Huffman tree;

[0151] B23: Construct Huffman coding based on Huffman tree;

[0152] B24: Compress the data to be shared according to Huffman coding to form encrypted data.

[0153] In step B23, the calculation formula of the average code length of Huffman coding is lavg=∑(W i *L i ) / ∑W i , where W i is the weight, which is the number of times the ASCII character appears.

[0154] The step B3 comprises the following steps:

[0155] B31: Calculate the correlation between attribute columns in the shared data based on Cramer's correlation coefficient;

[0156] B32: Find the attribute column with the highest correlation with the sensitive attribute column and merge it with the sensitive attribute column into the same set;

[0157] B33: Calculate the silhouette coefficient of each attribute, and obtain the optimal number of divisions and the division results according to the maximum silhouette coefficient value;

[0158] B34: Use generalization technology to generalize each cluster and obtain anonymous data.

[0159] In step B31, the formula for calculating the correlation is: Where Wi is the attribute weight, d i (x, y) is the distance between x and y on attribute i.

[0160] In step B33, the calculation formula of the silhouette coefficient is s(i)=[b(i)-a(i) / max{a(i), b(i)}], where s(i) is the silhouette coefficient of the ith sample, a(i) is the average distance between the ith sample and other samples in its cluster, and b(i) is the average distance between the ith sample and all samples in the nearest cluster.

[0161] Embodiment 7:

[0162] The data management system based on privacy computing includes:

[0163] A data receiving module, which receives a data sharing request and performs access authentication;

[0164] A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request;

[0165] The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree;

[0166] Cloud service module, which shares the calculated data with users through cloud services;

[0167] The data sharing detection module detects the data sharing process through the data sharing detection method.

[0168] The working method of the data receiving module comprises the following steps:

[0169] A1: Receive data sharing requests through the data sharing platform;

[0170] A2: Obtain the user's identity information based on the data sharing request;

[0171] A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

[0172] The working method of the data preprocessing module comprises the following steps:

[0173] B1: extracting the shared data to be shared according to the user's data sharing request;

[0174] B2: Encrypt the shared data using the Huffman method and obtain the encrypted data;

[0175] B3: Use the center point clustering algorithm to anonymize the encrypted data.

[0176] The step B2 comprises the following steps:

[0177] B21: Analyze the shared data and count the number of occurrences of each ASCII character;

[0178] B22: Use the number of occurrences of each ASCII character as a weight to construct a Huffman tree;

[0179] B23: Construct Huffman coding based on Huffman tree;

[0180] B24: Compress the data to be shared according to Huffman coding to form encrypted data.

[0181] In step B23, the calculation formula of the average code length of Huffman coding is lavg=∑(W i *L i ) / ∑W i , where W i is the weight, which is the number of times the ASCII character appears.

[0182] The step B3 comprises the following steps:

[0183] B31: Calculate the correlation between attribute columns in the shared data based on Cramer's correlation coefficient;

[0184] B32: Find the attribute column with the highest correlation with the sensitive attribute column and merge it with the sensitive attribute column into the same set;

[0185] B33: Calculate the silhouette coefficient of each attribute, and obtain the optimal number of divisions and the division results according to the maximum silhouette coefficient value;

[0186] B34: Use generalization technology to generalize each cluster and obtain anonymous data.

[0187] In step B31, the formula for calculating the correlation is: Where W i is the attribute weight, d i (x, y) is the distance between x and y on attribute i.

[0188] In step B33, the calculation formula of the silhouette coefficient is s(i)=[b(i)-a(i) / max{a(i), b(i)}], where s(i) is the silhouette coefficient of the ith sample, a(i) is the average distance between the ith sample and other samples in its cluster, and b(i) is the average distance between the ith sample and all samples in the nearest cluster.

[0189] The calculation method of the privacy security calculation module includes the following steps:

[0190] C1: Set the differential privacy budget;

[0191] C2: Allocate the budget to each level of the decision tree according to the arithmetic difference privacy budget allocation mechanism, and calculate the privacy budget of each node in each level;

[0192] C3: Divide the privacy budget allocated to each node into two equal parts;

[0193] C4: If the data on the node meets the preset stop condition, the node is set as a leaf node and the creation of sub-nodes is stopped. Then, the Laplace mechanism is used to add noise to the count value of the node, and the most common classification in the sample is selected as the label of the leaf node;

[0194] C5: If the data on the node does not meet the stopping condition, the number of data sets contained in the node is calculated and noise is added to it using the Laplace mechanism.

[0195] In the C4 step, the formula resulting from the addition is N oise=Lap(0,b / ε), where b is the sensitivity and ε is the privacy budget.

[0196] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A data management system based on privacy computing, characterized in that: include: A data receiving module, which receives a data sharing request and performs access authentication; A data preprocessing module, which obtains and preprocesses shared data based on a data sharing request; The privacy-safety calculation module performs privacy-safety calculations on the preprocessed data based on the CATR decision tree; Cloud service module, which shares the calculated data with users through cloud services; The data sharing detection module detects the data sharing process through the data sharing detection method.

2. The data management system based on privacy computing according to claim 1, characterized in that: The working method of the data receiving module comprises the following steps: A1: Receive data sharing requests through the data sharing platform; A2: Obtain the user's identity information based on the data sharing request; A3: Perform biometric authentication based on the obtained user identity information, confirm the user's permission level, and determine whether the user has access rights.

3. The data management system based on privacy computing according to claim 1, characterized in that: The working method of the data preprocessing module comprises the following steps: B1: extracting the shared data to be shared according to the user's data sharing request; B2: Encrypt the shared data using the Huffman method and obtain the encrypted data; B3: Use the center point clustering algorithm to anonymize the encrypted data.

4. The data management system based on privacy computing according to claim 3 is characterized in that: The step B2 comprises the following steps: B21: Analyze the shared data and count the number of occurrences of each ASCII character; B22: Use the number of occurrences of each ASCII character as a weight to construct a Huffman tree; B23: Construct Huffman coding based on Huffman tree; B24: compress the data to be shared according to Huffman coding to form encrypted data; In step B23, the calculation formula of the average code length of Huffman coding is lavg=∑(Wi*Li) / ∑Wi, where Wi is a weight value, which is the number of times the ASCII character appears.

5. The data management system based on privacy computing according to claim 3 is characterized in that: The step B3 comprises the following steps: B31: Calculate the correlation between attribute columns in the shared data based on Cramer's correlation coefficient; B32: Find the attribute column with the highest correlation with the sensitive attribute column and merge it with the sensitive attribute column into the same set; B33: Calculate the silhouette coefficient of each attribute, and obtain the optimal number of divisions and the division results according to the maximum silhouette coefficient value; B34: Use generalization technology to generalize each cluster and obtain anonymous data.

6. The data management system based on privacy computing according to claim 5, characterized in that: In step B31, the formula for calculating the correlation is: Where Wi is the attribute weight, and di(x, y) is the distance between x and y on attribute i.

7. The data management system based on privacy computing according to claim 5, characterized in that: In step B33, the calculation formula of the silhouette coefficient is: s(i )=[ b(i )-a(i) / max{ a(i ), b(i)}], where s(i) is the silhouette coefficient of the ith sample, a(i) is the average distance between the ith sample and other samples in its cluster, and b(i) is the average distance between the ith sample and all samples in the nearest cluster.

8. The data management system based on privacy computing according to claim 1, characterized in that: The calculation method of the privacy security calculation module includes the following steps: C1: Set the differential privacy budget; C2: Allocate the budget to each level of the decision tree according to the arithmetic difference privacy budget allocation mechanism, and calculate the privacy budget of each node in each level; C3: Divide the privacy budget allocated to each node into two equal parts; C4: If the data on the node meets the preset stop condition, the node is set as a leaf node and the creation of sub-nodes is stopped. Then, the Laplace mechanism is used to add noise to the count value of the node, and the most common classification in the sample is selected as the label of the leaf node; C5: If the data on the node does not meet the stopping condition, the number of data sets contained in the node is calculated and noise is added to it using the Laplace mechanism.

9. The data management system based on privacy computing according to claim 8, characterized in that: In the C4 step, the formula resulting from the addition is N oise=Lap(0,b / ε), where b is the sensitivity and ε is the privacy budget.

Images