Transaction data personalization processing method and system based on privacy preference and trust degree

Abstract

The present application relates to data security technology, and discloses a transaction data personalized processing method and system based on privacy preference and trust degree, comprising: a data owner classifies the attributes of transaction data to obtain multiple levels of attributes; the privacy protection requirements of the attributes in the multiple levels of attributes are calculated, and multiple privacy preference levels are divided; the self information of a data purchaser is stored in a chain, and the trust degree score of the data owner to the data purchaser is calculated; a trusted smart contract generates multiple protection instructions; when the trusted smart contract monitors a first level protection instruction, the transaction data is issued to the data purchaser; when the trusted smart contract monitors a second level protection instruction, the transaction data is issued to the data purchaser after differential privacy difference is performed on the transaction data; and when the trusted smart contract monitors a third level protection instruction, the transaction data is issued to the data purchaser after the transaction data is encrypted by using hierarchical homomorphic encryption and proxy re-encryption for encryption processing. The present application can improve the privacy protection security of data transaction.

Description

Technical Field

[0001] This invention relates to the field of data security technology, and in particular to a method and system for personalized processing of transaction data based on privacy preferences and trust levels. Background Technology

[0002] In the digital economy era, data, as a core factor of production, has become a key engine driving industrial upgrading through its cross-entity circulation and value transformation. However, the conflict between value extraction and privacy protection is increasingly prominent in transactional data, which serves as the core carrier of data elements. The booming development of the data trading market has not only built a bridge for value exchange between data owners and buyers, but also placed higher demands on the coordinated balance between privacy protection and data utility.

[0003] Existing data processing solutions for data transactions generally suffer from three significant drawbacks: First, privacy protection is often applied in a one-size-fits-all manner, failing to classify data attributes into finer categories based on sensitivity, leakage risk, etc., and neglecting to fully consider the subjective protection intentions of data owners, making it difficult to adapt to the differentiated privacy needs of different users for different data attributes. Second, the trust relationship between the transacting parties is not incorporated into data processing decisions, ignoring differences in institutional qualifications, historical transaction compliance, and honesty among data buyers, resulting in a disconnect between protection strategies and the trust logic of real-world transaction scenarios. Third, privacy budget allocation lacks flexibility, often employing a fixed range of values ​​without dynamically adjusting to the privacy consumption differences in data usage scenarios. Furthermore, single perturbation techniques cannot adapt to different types of data, such as numerical and discrete data, easily leading to over-protection resulting in data utility loss, or insufficient protection causing privacy leaks, resulting in low privacy protection security in data transactions. Summary of the Invention

[0004] This invention provides a method and system for personalized processing of transaction data based on privacy preferences and trust levels, which can improve the privacy protection and security of data transactions.

[0005] To achieve the above objectives, the present invention provides a method for personalized processing of transaction data based on privacy preferences and trust levels, comprising:

[0006] Data owners classify the attributes of transaction data into multi-level attributes by multiple dimensions.

[0007] The privacy preference formula is used to calculate the privacy protection requirements of attributes in a multi-level attribute, and the attributes in the multi-level attribute are divided into multiple privacy preference levels according to the privacy protection requirements;

[0008] Blockchain notarization technology is used to store the data buyer's personal information on the blockchain, and a trusted smart contract is used to retrieve the data buyer's personal information and a trust score is calculated using a trust formula.

[0009] The trusted smart contract inputs the privacy preference level and trust score into a preset ternary mapping matrix to generate multiple protection instructions, including level 1 protection instructions, level 2 protection instructions and level 3 protection instructions;

[0010] When a trusted smart contract detects a Level 1 protection instruction, the transaction data is sent to the data buyer.

[0011] When a trusted smart contract detects a secondary protection instruction, it performs differential privacy processing on the transaction data before sending it to the data buyer.

[0012] When a trusted smart contract detects a Level 3 protection instruction, it encrypts the transaction data using hierarchical homomorphic encryption and proxy re-encryption before sending it to the data buyer.

[0013] Optionally, the data owner classifies the attributes of the transaction data into multi-level attributes based on multiple dimensions, including:

[0014] Extract all attributes from the transaction data and summarize all attributes into an attribute set;

[0015] The attribute set is classified based on the dimensions of data sensitivity, leakage risk, and compliance requirements to obtain multi-level attributes.

[0016] Optionally, the step of calculating the privacy protection requirements of attributes in a multi-level attribute system using the privacy preference formula includes:

[0017] The privacy requirements of attributes in a multi-level attribute system are calculated using the following formula. :

[0018]

[0019] in, For the first The desire to protect each attribute For the first Sensitive perception of each attribute For the first The scene adaptability of each attribute, To protect the weight of intention, For sensitive perception weights, Assigning weights to scene adaptability. For the first The attribute and the first The degree of association between attributes This represents the weight of attribute correlation.

[0020] Optionally, the step of using a trusted smart contract to access the data buyer's own information and calculating the data owner's trust score for the data buyer using a trust formula includes:

[0021] Obtain information on the qualifications, number of transactions, and honesty rate of data buyers;

[0022] Based on qualification information, transaction frequency information, and honesty rate information, a trust score is calculated using a trust formula. The following formula is used to calculate the trust score between the data owner and the data buyer. :

[0023]

[0024] in, For qualification information, For transaction frequency information, For honesty rate information, Weighting of qualification information Weighting of transaction frequency information, The information weight is the honesty rate.

[0025] Optionally, the step of performing differential privacy analysis on the transaction data before sending it to the data buyer when the trusted smart contract detects a secondary protection instruction includes:

[0026] Select a privacy budget range based on privacy preference level and trust score, and select a privacy budget value based on a preset ternary mapping matrix;

[0027] Identify the data type of the transaction data, select a preset differential privacy interference mechanism based on the data type to perform differential privacy processing on the transaction data to obtain encrypted interference data, and then send the encrypted interference data to the data buyer.

[0028] Optionally, the step of selecting a preset differential privacy interference mechanism based on the data type to perform differential privacy processing on the transaction data to obtain encrypted interference data includes:

[0029] When the data type is numerical, Laplace noise is used to interfere with the attributes of the transaction data;

[0030] When the data type is discrete, an exponential mechanism is used to interfere with the attributes of the transaction data.

[0031] Optionally, the step of encrypting the transaction data using hierarchical homomorphic encryption and proxy re-encryption before sending it to the data buyer when the trusted smart contract detects a Level 3 protection instruction includes:

[0032] The preset key generation center generates homomorphically encrypted data owner key pairs and sends them to the data owner, and generates data buyer key pairs and sends them to the data buyer;

[0033] For the one-way proxy re-encryption algorithm, the parameter space is a bilinear group. First generator Second generator , For prime numbers of order 1 The first cyclic group, For prime numbers of order 1 The second cyclic group, It is a bilinear mapping. The target group is a bilinear mapping; the data buyer randomly selects the buyer's private key. ,in, For model multiplication group For smart = from the multiplication group A uniformly random element is selected from the set, and the generator is used to... Calculate the buyer's public key Buyer's public key Send to the data owner;

[0034] The data owner receives the buyer's public key. Then, combined with the owner's private key Generate a re-encryption key ;

[0035] After all keys are generated, the data owner performs homomorphic encryption on the highly sensitive attribute values ​​corresponding to the attributes in the proposed data to obtain the ciphertext. ,in, The attribute value is the encrypted ciphertext. Sending it to the data buyer, because homomorphic encryption has the property that the decrypted result of ciphertext operation is completely identical to the result of direct operation on the original plaintext, the data buyer can... Perform the calculation directly on the top and then use the result. The message is sent to a trusted smart contract, which then performs proxy re-encryption to obtain the re-encrypted ciphertext. ,in, , will re-encrypt ciphertext The data is sent to the data buyer, who can then use their private key to decrypt the transaction data. The decrypted transaction data can then be obtained, including... For data buyers, the buyer's private key is used under homomorphic encryption.

[0036] To address the aforementioned issues, this invention also provides a system for personalized transaction data processing based on privacy preferences and trust levels, comprising a data owner, a data buyer, a trusted smart contract, and a blockchain evidence storage node. The data owner and the data buyer communicate bidirectionally with the trusted smart contract, and the data owner and the data buyer communicate unidirectionally with the blockchain evidence storage node. The blockchain evidence storage node also communicates unidirectionally with the trusted smart contract.

[0037] Optionally, the data owner and the data purchaser may send the transaction information to a blockchain notarization node for notarization.

[0038] Optionally, the trusted smart contract retrieves the data buyer's own information and uses a trust formula to calculate the data owner's trust score in the data buyer.

[0039] This invention allows data owners to classify transaction data attributes into multi-level attributes across multiple dimensions, enabling refined and structured hierarchical classification of transaction data attributes. Furthermore, it utilizes a privacy preference formula to calculate the privacy protection requirements of attributes within the multi-level attributes and divides these attributes into multiple privacy preference levels based on these requirements, allowing for personalized privacy configuration. Moreover, it employs blockchain notarization technology to store the data buyer's personal information on the blockchain and uses a trusted smart contract to retrieve this information and calculate the data owner's trust score for the data buyer using a trust degree formula. This ensures that the privacy protection strategy aligns with the trust logic of the transaction scenario, and the blockchain storage of the data buyer's information guarantees its originality. Finally, the trusted smart contract inputs the privacy preference levels and trust score into a preset ternary mapping matrix to generate multiple protection instructions, including first-level, second-level, and third-level protection instructions. This enables dynamic and flexible privacy protection strategies, improving the privacy security of data transactions. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating a method for personalized transaction data processing based on privacy preferences and trust levels, provided in an embodiment of the present invention.

[0041] Figure 2 A system schematic diagram of a personalized transaction data processing method based on privacy preferences and trust levels provided in an embodiment of the present invention;

[0042] Figure 3 A schematic diagram of a ternary mapping matrix for a personalized transaction data processing method based on privacy preferences and trust levels provided in an embodiment of the present invention;

[0043] Figure 4This is a schematic diagram illustrating the processing steps of a personalized transaction data processing method based on privacy preferences and trust levels according to an embodiment of the present invention.

[0044] Figure 5 This is a timing diagram of the processing steps of a personalized transaction data processing method based on privacy preferences and trust levels, provided in an embodiment of the present invention.

[0045] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0046] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0047] This application provides a method for personalized transaction data processing based on privacy preferences and trust levels. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, this method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0048] Reference Figure 1 The diagram shown is a flowchart illustrating a method for personalized transaction data processing based on privacy preferences and trust levels according to an embodiment of the present invention. In this embodiment, the method for personalized transaction data processing based on privacy preferences and trust levels includes:

[0049] S1. The data owner classifies the attributes of the transaction data into multi-level attributes by multiple dimensions.

[0050] Understandably, the data owner refers to the entity that stores the original transaction data and its attributes.

[0051] Understandably, transaction data refers to data held by the data owner, and transaction data can be medical data collected in accordance with laws and regulations.

[0052] Understandably, multi-dimensional attributes refer to the requirements that the attributes of transaction data should meet. For example, multi-dimensional attributes include, but are not limited to, data sensitivity, leakage risk, and compliance requirements.

[0053] Specifically, data owners classify the attributes of transaction data into multi-dimensional attributes to obtain multi-level attributes, including:

[0054] Extract all attributes from the transaction data and summarize all attributes into an attribute set;

[0055] The attribute set is classified based on the dimensions of data sensitivity, leakage risk, and compliance requirements to obtain multi-level attributes.

[0056] Understandably, an attribute refers to each specific field in the transaction data. For example, attributes include, but are not limited to, height, weight, and email address.

[0057] Understandably, the attributes of transaction data correspond to attribute values, such as the specific height corresponding to the attribute height mentioned above.

[0058] For example, the attributes and attribute values ​​of transaction data can be as follows:

[0059] The transaction data set is ,in The number of data samples, each data sample , Indicates the first The first sample One attribute, The number of attributes, and the corresponding attribute value. Attribute set Taken from data domain , For the first The value space of each attribute, i.e. .

[0060] Understandably, the multi-level attributes include a first attribute level (core sensitivity), a second attribute level (moderate sensitivity), and a third attribute level (basic sensitivity). The first attribute level, core sensitivity, primarily includes identity-identifying data such as ID card numbers, facial data, and fingerprint data. The second attribute level, moderate sensitivity, primarily includes sensitive transaction data such as phone numbers, email addresses, and home addresses. The third attribute level, basic information, primarily includes basic information such as height, weight, and disease type data. The attribute classification results provide a foundation for subsequent sensitivity perception quantification and privacy protection strategy selection.

[0061] S2. Calculate the privacy protection requirements of attributes in the multi-level attributes using the privacy preference formula, and divide the attributes in the multi-level attributes into multiple privacy preference levels according to the privacy protection requirements.

[0062] Specifically, the privacy protection requirements of attributes in a multi-level attribute hierarchy are calculated using a privacy preference formula, including:

[0063] The privacy requirements of attributes in a multi-level attribute system are calculated using the following formula. :

[0064]

[0065] in, For the first The desire to protect each attribute For the first Sensitive perception of each attribute For the first The scene adaptability of each attribute, To protect the weight of intention, For sensitive perception weights, Assigning weights to scene adaptability. For the first The attribute and the first The degree of association between attributes This represents the weight of attribute correlation.

[0066] Understandable is the desire to protect. (Value range 1-5): Determined by the data owner based on the attribute. Self-assessment of protection needs; higher scores indicate a stronger willingness to proactively protect; sensitive perception (Value range 1-5): Attribute The objective privacy sensitivity is determined by the above attribute classification results, with the first attribute level having a value of 5, the second attribute level having a value of 3, and the third attribute level having a value of 1; scenario adaptability. (Value range 1-5): Identifies the attribute In the current transaction scenario, the lower the score, the greater the necessity for the data buyer, and the less noise interference there is.

[0067] For example, the attribute association degree is calculated as follows:

[0068] Attribute correlation (Value range 1-5): Used to uncover potential linear correlations between common attributes and core sensitive attributes, avoiding indirect privacy leaks caused by attribute isolation protection. The Pearson correlation coefficient is used for measurement, and its expression is defined as follows: Among them, molecular covariance Used to measure the consistency of the changing trends of two variables, denominator variance Used for measurement The degree of dispersion of values, denominator variance Used for measurement The degree of dispersion of the value. The higher the attribute correlation value, the stronger the correlation. The final result is normalized to 1-5.

[0069] Understandably, multiple privacy preference levels refer to levels categorized based on privacy protection needs, for example, according to a privacy preference function. The calculation results (value range 1-5) divide the privacy preference level into four levels: insensitive (1-2), moderately sensitive (2-3), highly sensitive (3-4), and extremely sensitive (4-5).

[0070] S3. Utilize blockchain notarization technology to store the data buyer's personal information on the blockchain, and use trusted smart contracts to retrieve the data buyer's personal information and use a trust formula to calculate the data owner's trust score in the data buyer.

[0071] Understandably, blockchain evidence storage technology refers to using the immutability, distributed consensus, and timestamp mechanism of blockchain to record data or events on the chain, thereby achieving trusted evidence storage.

[0072] Understandably, a trusted smart contract refers to an automated program deployed on a blockchain that automatically executes preset logic when preset conditions are met.

[0073] Understandably, the data buyer's personal information refers to the data buyer's own attribute information. For example, personal information includes, but is not limited to, qualification information, transaction frequency information, and honesty rate information. Among them, qualification information refers to the data buyer's compliance qualification certificates. The institutional qualification of an individual data buyer is valued at 1, the institutional qualification of an enterprise-level data buyer is valued at 2, and the institutional qualification of a national-level research institution is valued at 3. Different institutions at the same level can be further subdivided. Transaction frequency information refers to the result of normalization of the historical number of transactions between the data owner and the data buyer. The more transactions, the higher the score. Honesty rate information refers to the compliance rate of the data buyer's historical data processing, such as whether there are any records of illegal disclosure and abuse. No violations are valued at 3, and violations are normalized to 1-3 according to their severity.

[0074] Specifically, a trusted smart contract is used to access the data buyer's personal information, and a trust score is calculated using a trust formula, including:

[0075] Obtain information on the qualifications, number of transactions, and honesty rate of data buyers;

[0076] Based on qualification information, transaction frequency information, and honesty rate information, a trust score is calculated using a trust formula. The following formula is used to calculate the trust score between the data owner and the data buyer. :

[0077]

[0078] in, For qualification information, For transaction frequency information, For honesty rate information, Weighting of qualification information Weighting of transaction frequency information, The information weight is the honesty rate.

[0079] S4. The trusted smart contract inputs the privacy preference level and trust score into a preset ternary mapping matrix to generate multiple protection instructions, including level 1 protection instructions, level 2 protection instructions and level 3 protection instructions.

[0080] Understandably, the pre-defined ternary mapping matrix refers to a privacy budget value matrix constructed using multiple privacy preference levels and trust scores.

[0081] S5. When the trusted smart contract detects a Level 1 protection instruction, it will send the transaction data to the data buyer.

[0082] For example, when a trusted smart contract detects a Level 1 protection instruction, the transaction data can be sent to the data buyer, as shown in the following steps:

[0083] The detection of the first protection command indicates that the data attribute is not sensitive. At the same time, data owners have a high degree of trust in data buyers. At this point, the data privacy risk is extremely low, and the utility of the data can be maximized. That is, no privacy processing is performed on the attribute values; the original attribute values ​​are traded directly. In other words, the attribute values ​​of the data to be traded are... .

[0084] S6. When a trusted smart contract detects a secondary protection instruction, it performs differential privacy processing on the transaction data and then sends it to the data buyer.

[0085] Understandably, differential privacy refers to adding random noise to data analysis or query results, making it impossible for attackers to determine whether a particular individual appears in the dataset.

[0086] Specifically, when a trusted smart contract detects a secondary protection instruction, it performs differentiated privacy differentials on the transaction data before sending it to the data purchaser, including:

[0087] Select a privacy budget range based on privacy preference level and trust score, and select a privacy budget value based on a preset ternary mapping matrix;

[0088] Identify the data type of the transaction data, select a preset differential privacy interference mechanism based on the data type to perform differential privacy processing on the transaction data to obtain encrypted interference data, and then send the encrypted interference data to the data buyer.

[0089] Furthermore, based on the data type, a preset differential privacy interference mechanism is selected to perform differential privacy processing on the transaction data to obtain encrypted interference data, including:

[0090] When the data type is numerical, Laplace noise is used to interfere with the attributes of the transaction data;

[0091] When the data type is discrete, an exponential mechanism is used to interfere with the attributes of the transaction data.

[0092] Understandably, Laplace noise refers to random perturbation values ​​drawn from the Laplace distribution.

[0093] Understandably, the index mechanism refers to...

[0094] For example, when a trusted smart contract detects a secondary protection instruction, it performs differential privacy processing on the transaction data before sending it to the data buyer, as shown in the following steps:

[0095] Upon detecting the second protection instruction, a corresponding privacy budget range will be selected based on different privacy preferences and levels of trust. A differentiated noise perturbation strategy will then be implemented. For example, if the attribute is extremely sensitive data, even with high trust in the data buyer, a lower privacy budget range is needed to protect the attribute value. The corresponding privacy budget range in this case is... After selecting the corresponding privacy budget range based on the ternary mapping matrix, TSC will randomly select a value from the range as the final privacy budget value. Next, TSC will determine the data type of the attribute and generate adaptive noise based on the different data types of the transaction data attributes. In this invention, noise is mainly divided into numerical data and discrete data. Laplace noise is used for numerical data to ensure that the statistical characteristics of the data are not affected. For discrete data, an exponential mechanism is used to select the alternative category by probability to ensure that the noise matches the data distribution.

[0096] After determining the privacy budget value If the data is numerical, then Laplace noise should be used to adjust the attributes. To generate interference, the Laplace noise follows a Laplace distribution, and its probability density function is: ,in For position parameters, The scale parameter determines the degree of noise dispersion. Location parameter... That is, the mean and scale parameter of the noise. Directly from the privacy budget With attribute sensitivity The determination is the core indicator for controlling noise intensity, and the calculation formula is... Strictly follow The theoretical requirements of differential privacy include, among which For the range of attribute values, the privacy budget value This means that the privacy budget range is randomly determined by the TSC. The smaller the privacy budget value, the stronger the generated noise, indicating a higher demand for attribute protection. In determining... and Then, based on the probability characteristics of the Laplace distribution, noise values ​​are generated through uniform distribution sampling and inverse transformation. Specifically, two independent noise values ​​are first generated. Uniformly distributed random numbers in intervals and The noise value is calculated using the inverse transform formula. ,when hour, ;when hour, That is, the final data attribute value for the prepared transaction is... + .

[0097] After determining the privacy budget value If the data is discrete, then an exponential mechanism is used to apply the attributes. The interference effect of the exponential mechanism depends on four core parameters: the discrete attribute range. Privacy Budget utility function and utility function sensitivity The discrete attribute value range It is the set of all legal values ​​for this attribute, utility function It is a measure of the raw data The attribute value perturbation is for the category The utility, the sensitivity of the utility function It is the maximum change in the utility function, that is , To and Adjacent datasets differing by only one record, with the most critical parameter being the privacy budget. This determines the strength of privacy protection; its value is randomly determined by the TSC within the corresponding privacy budget range. Value. Specifically, first according to the formula Calculate each category The index utility value is used to determine the weight of a category being selected as noise. Then, the sum of all index utility values ​​is calculated, and the index utility value of each category is divided by the sum to obtain the final probability. Finally, the categories are sorted in descending order of probability, and the cumulative probability is calculated to form... The probability segment within the interval is finally generated. Uniform random numbers in intervals ,judge The probability range to which the noise belongs, and the category corresponding to the probability range, is the final noise category. That is, the attribute value of the final prepared transaction data (Replace the original) ).

[0098] S7. When the trusted smart contract detects a Level 3 protection instruction, the transaction data is encrypted using hierarchical homomorphic encryption and proxy re-encryption before being sent to the data buyer.

[0099] Understandably, hierarchical homomorphic encryption refers to a special homomorphic encryption technique that allows specific forms of mathematical operations (such as addition and multiplication) to be performed directly on encrypted data. After the operation is decrypted, the result is consistent with the result of performing the same operation on plaintext data. Hierarchical homomorphic encryption means that the encryption scheme has a multi-level key structure, and keys at different levels can decrypt ciphertext at different "depths" (i.e., after different numbers of operations).

[0100] Understandably, proxy re-encryption refers to a cryptographic mechanism that allows a semi-trusted third party to convert ciphertext encrypted with one party's public key into ciphertext with another party's public key, enabling the data recipient to decrypt it, but preventing the proxy from obtaining the plaintext.

[0101] Specifically, when a trusted smart contract detects a Level 3 protection instruction, the transaction data is encrypted using hierarchical homomorphic encryption and proxy re-encryption before being sent to the data purchaser, including:

[0102] The preset key generation center generates homomorphically encrypted data owner key pairs and sends them to the data owner, and generates data buyer key pairs and sends them to the data buyer;

[0103] For the one-way proxy re-encryption algorithm, the parameter space is a bilinear group. First generator Second generator , For prime numbers of order 1 The first cyclic group, For prime numbers of order 1 The second cyclic group, It is a bilinear mapping. The target group is a bilinear mapping; the data buyer randomly selects the buyer's private key. ,in, For model multiplication group For smart = from the multiplication group A uniformly random element is selected from the set, and the generator is used to... Calculate the buyer's public key Buyer's public key Send to the data owner;

[0104] The data owner receives the buyer's public key. Then, combined with the owner's private key Generate a re-encryption key ;

[0105] After all keys are generated, the data owner performs homomorphic encryption on the highly sensitive attribute values ​​corresponding to the attributes in the proposed data to obtain the ciphertext. ,in, The attribute value is the encrypted ciphertext. Sending it to the data buyer, because homomorphic encryption has the property that the decrypted result of ciphertext operation is completely identical to the result of direct operation on the original plaintext, the data buyer can... Perform the calculation directly on the top and then use the result. The message is sent to a trusted smart contract, which then performs proxy re-encryption to obtain the re-encrypted ciphertext. ,in, , will re-encrypt ciphertext The data is sent to the data buyer, who can then use their private key to decrypt the transaction data. The decrypted transaction data can then be obtained, including... For data buyers, the buyer's private key is used under homomorphic encryption.

[0106] This invention allows data owners to classify transaction data attributes into multi-level attributes across multiple dimensions, enabling refined and structured hierarchical classification of transaction data attributes. Furthermore, it utilizes a privacy preference formula to calculate the privacy protection requirements of attributes within the multi-level attributes and divides these attributes into multiple privacy preference levels based on these requirements, allowing for personalized privacy configuration. Moreover, it employs blockchain notarization technology to store the data buyer's personal information on the blockchain and uses a trusted smart contract to retrieve this information and calculate the data owner's trust score for the data buyer using a trust degree formula. This ensures that the privacy protection strategy aligns with the trust logic of the transaction scenario, and the blockchain storage of the data buyer's information guarantees its originality. Finally, the trusted smart contract inputs the privacy preference levels and trust score into a preset ternary mapping matrix to generate multiple protection instructions, including first-level, second-level, and third-level protection instructions. This enables dynamic and flexible privacy protection strategies, improving the privacy security of data transactions.

[0107] like Figure 2 The figure shown is a system schematic diagram of a transaction data personalization processing method based on privacy preferences and trust levels provided in an embodiment of the present invention.

[0108] Specifically, the system includes a data owner, a data buyer, a trusted smart contract, and a blockchain evidence storage node. The data owner and the data buyer communicate bidirectionally with the trusted smart contract, and the data owner and the data buyer communicate unidirectionally with the blockchain evidence storage node. The blockchain evidence storage node also communicates unidirectionally with the trusted smart contract.

[0109] Furthermore, the data owner and the data purchaser send the transaction information to the blockchain notarization node for notarization.

[0110] Furthermore, the trusted smart contract retrieves the data buyer's own information and uses a trust formula to calculate the data owner's trust score in the data buyer.

[0111] like Figure 3 The diagram shown is a ternary mapping matrix schematic of a transaction data personalization processing method based on privacy preferences and trust levels provided in an embodiment of the present invention.

[0112] like Figure 4 The diagram shown illustrates the processing steps of a personalized transaction data processing method based on privacy preferences and trust levels, according to an embodiment of the present invention.

[0113] like Figure 5 The diagram shown is a sequence diagram of the processing steps of a personalized transaction data processing method based on privacy preferences and trust levels provided in an embodiment of the present invention.

[0114] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0115] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0116] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0117] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for personalized processing of transaction data based on privacy preferences and trust levels, characterized in that, The method includes: Data owners classify the attributes of transaction data into multi-level attributes by multiple dimensions. The privacy preference formula is used to calculate the privacy protection requirements of attributes in a multi-level attribute, and the attributes in the multi-level attribute are divided into multiple privacy preference levels according to the privacy protection requirements; The data purchaser's personal information is stored on the blockchain using blockchain notarization technology. A trusted smart contract is then used to retrieve this information, and a trust score is calculated based on the data owner's trust in the purchaser. The process of retrieving the data purchaser's personal information using a trusted smart contract and calculating the trust score includes: Obtain information on the qualifications, number of transactions, and honesty rate of data buyers; Based on qualification information, transaction frequency information, and honesty rate information, a trust score is calculated using a trust formula. The following formula is used to calculate the trust score between the data owner and the data buyer. : ； in, For qualification information, For transaction frequency information, For honesty rate information, Weighting of qualification information Weighting of transaction frequency information, Weighting of honesty rate information; The trusted smart contract inputs the privacy preference level and trust score into a preset ternary mapping matrix to generate multiple protection instructions, including level 1 protection instructions, level 2 protection instructions and level 3 protection instructions; When a trusted smart contract detects a Level 1 protection instruction, the transaction data is sent to the data buyer. When a trusted smart contract detects a secondary protection instruction, it performs differential privacy processing on the transaction data before sending it to the data buyer. When a trusted smart contract detects a Level 3 protection instruction, it encrypts the transaction data using hierarchical homomorphic encryption and proxy re-encryption before sending it to the data buyer.

2. The personalized transaction data processing method based on privacy preferences and trust levels as described in claim 1, characterized in that, The data owner classifies the attributes of the transaction data into multi-level attributes based on multiple dimensions, including: Extract all attributes from the transaction data and summarize all attributes into an attribute set; The attribute set is classified based on the dimensions of data sensitivity, leakage risk, and compliance requirements to obtain multi-level attributes.

3. The method for personalized transaction data processing based on privacy preferences and trust levels as described in claim 1, characterized in that, The calculation of privacy protection requirements for attributes in a multi-level attribute system using the privacy preference formula includes: The privacy requirements of attributes in a multi-level attribute system are calculated using the following formula. : ； in, For the first The desire to protect each attribute For the first Sensitive perception of each attribute, For the first The scene adaptability of each attribute, To protect the weight of intention, For sensitive perception weights, Assigning weights to scene adaptability. For the first The attribute and the first The degree of association between attributes This represents the weight of attribute correlation.

4. The method for personalized transaction data processing based on privacy preferences and trust levels as described in claim 1, characterized in that, When a trusted smart contract detects a secondary protection instruction, the transaction data is processed with differential privacy features before being sent to the data purchaser, including: Select a privacy budget range based on privacy preference level and trust score, and select a privacy budget value based on a preset ternary mapping matrix; Identify the data type of the transaction data, select a preset differential privacy interference mechanism based on the data type to perform differential privacy processing on the transaction data to obtain encrypted interference data, and then send the encrypted interference data to the data buyer.

5. The method for personalized transaction data processing based on privacy preferences and trust levels as described in claim 4, characterized in that, The step of selecting a preset differential privacy interference mechanism based on the data type to perform differential privacy processing on the transaction data to obtain encrypted interference data includes: When the data type is numerical, Laplace noise is used to interfere with the attributes of the transaction data; When the data type is discrete, an exponential mechanism is used to interfere with the attributes of the transaction data.

6. The method for personalized transaction data processing based on privacy preferences and trust levels as described in claim 1, characterized in that, When a trusted smart contract detects a Level 3 protection instruction, the transaction data is encrypted using hierarchical homomorphic encryption and proxy re-encryption before being sent to the data buyer, including: The preset key generation center generates homomorphically encrypted data owner key pairs and sends them to the data owner, and generates data buyer key pairs and sends them to the data buyer; For the one-way proxy re-encryption algorithm, the parameter space is a bilinear group. First generator Second generator , For prime numbers of order 1 The first cyclic group, For prime numbers of order 1 The second cyclic group, It is a bilinear mapping. The target group is a bilinear mapping; the data buyer randomly selects the buyer's private key. ,in, For model multiplication group To the multiplication group A uniformly random element is selected from the set, and the generator is used to... Calculate the buyer's public key Buyer's public key Send to the data owner; The data owner receives the buyer's public key. Then, combined with the owner's private key Generate a re-encryption key ; After all keys are generated, the data owner performs homomorphic encryption on the highly sensitive attribute values ​​corresponding to the attributes in the proposed data to obtain the ciphertext. ,in, The attribute value is the encrypted ciphertext. Sending it to the data buyer, because homomorphic encryption has the property that the decrypted result of ciphertext operation is completely identical to the result of direct operation on the original plaintext, the data buyer can... Perform the calculation directly on the top and then use the result. The message is sent to a trusted smart contract, which then performs proxy re-encryption to obtain the re-encrypted ciphertext. ,in, , will re-encrypt ciphertext The data is sent to the data buyer, who can then use their private key to decrypt the transaction data. The decrypted transaction data can then be obtained, including... For data buyers, the buyer's private key is used under homomorphic encryption.

7. A system based on the personalized transaction data processing method based on privacy preferences and trust levels as described in any one of claims 1-6, characterized in that, It includes a data owner, a data buyer, a trusted smart contract, and a blockchain evidence storage node. The data owner and the data buyer communicate bidirectionally with the trusted smart contract, and the data owner and the data buyer communicate unidirectionally with the blockchain evidence storage node. The blockchain evidence storage node also communicates unidirectionally with the trusted smart contract.

8. The system for personalized transaction data processing based on privacy preferences and trust levels as described in claim 7, characterized in that, The data owner and the data purchaser send the transaction information to the blockchain notarization node for notarization.

9. The system for personalized transaction data processing based on privacy preferences and trust levels as described in claim 7, characterized in that, The trusted smart contract calls upon the data buyer's personal information and uses a trust formula to calculate the data owner's trust score in the data buyer.

Images