Privacy protection data transaction order matching method and system based on cardinal number privacy set intersection

Abstract

The invention discloses a privacy protection data transaction order matching method and system based on cardinal number privacy set intersection, and mainly solves the problems of privacy disclosure, weak trust relationship and low efficiency in data transaction order matching in the prior art. According to the implementation scheme, data transaction order matching parameters are initialized; the data seller generates either-or casual transmission parameters according to the attribute values in the to-be-sold data set; the data buyer obtains the casual transmission result through random query, and generates a shared value according to the casual transmission result; the data seller performs casual key value pair storage OKVS coding on the attribute value; and solving an intersection cardinal number between the preference set and the data attribute set through OKVS decoding, and comparing the intersection cardinal number with a threshold value to judge whether the transaction order is matched or not. According to the method, it is guaranteed that no information except the intersection cardinal number is leaked in the order matching process, one-to-many efficient privacy protection transaction order matching is achieved, potential data sellers are screened for data buyers, the order matching privacy, credibility and effectiveness in transaction are improved, and the method can be applied to digital finance and data transaction.

Description

Technical Field

[0001] This invention belongs to the field of data security and privacy computing technology, and specifically relates to a privacy-preserving data transaction order matching method and system, which can be applied to fields such as digital finance and data transactions. Background Technology

[0002] With the rapid development of technologies such as big data, machine learning, and multimodal large models, the demand for large-scale datasets has surged. Data circulation and trading are of paramount importance for breaking down data silos, releasing the value of data elements, creating new advantages in the digital economy, and fostering a healthy digital ecosystem. In the cloud data market, the core objective is to facilitate transactions between data buyers and sellers, enabling buyers to enhance their internal datasets with external data, thereby significantly improving the performance of their trained machine learning models. However, the frequent flow and pricing of transaction data among market players in different fields and industries, coupled with the increasing number of participants and factors such as improper storage, unauthorized abuse, and insufficient anonymization during the flow process, easily leads to a "trilemma" of value transfer, privacy protection, and trustworthy regulation.

[0003] Specifically, because a large amount of data is currently traded in plaintext on public trading platforms such as data exchanges or between private institutions, involving a significant amount of personal privacy information and trade secrets, such as identity, funds, and models, data assets face severe security threats. On the one hand, greedy data trading platforms may profit by selling the privacy information of trading parties or by registering fake data with the same attributes for sale, thus inciting vicious competition and disrupting the data market order. On the other hand, trading platforms can not only obtain the preferences of data buyers but also understand the popularity of various types of data through the data sold by sellers.

[0004] To address this, some solutions have been designed with efficient and secure order matching mechanisms, enabling customers to effectively match their orders, reducing market impact while protecting data privacy, and ensuring no information is leaked even without a match, thus achieving dynamic and secure price negotiation and data transactions. However, these solutions still have privacy leakage issues during the interactive calculation of data sample similarity.

[0005] While existing solutions have achieved some success in protecting identity and content privacy in data transactions, several issues remain unresolved regarding content privacy. Firstly, in data transactions involving third-party intermediary platforms, attackers can indirectly infer participants' private information through publicly available information from the negotiation process, such as price details and demand quantities. This allows attackers to engage in malicious price-cutting and price gouging of data commodities, leading to frequent transaction irregularities and severely disrupting the normal operation of the market. Secondly, once data is transferred during a transaction, its subsequent flow and use are difficult to monitor. Data leaks or illegal resales can cause irreversible damage to the rights of data owners. Therefore, achieving privacy-protected pricing negotiation and data flow status control is crucial. Furthermore, most existing privacy-protected data transaction solutions are geared towards one-to-one transaction order matching, failing to consider the principle of "comparing prices before buying" in actual transactions and lacking screening of potential qualified sellers.

[0006] Patent document CN202111383713.7 discloses a method, apparatus, device, and storage medium for privacy data transactions. It proposes an interaction protocol based on secret sharing and smart contracts. The seller splits the original data into multiple data fragments using secret sharing technology, distributing them across multiple nodes, and storing the verification information hash value on the blockchain. The buyer uses a smart contract to perform predetermined computational tasks on the secretly shared data fragments without exposing the plaintext data, obtaining the computational results to verify the data value. Upon successful verification, the smart contract automatically triggers payment and transmits the data fragments stored on different nodes to the buyer. The buyer can only recover the original data after collecting all the fragments. However, this technology primarily addresses how the buyer can design a "verification computation task" that effectively reflects the overall value of the data without revealing too much data information. Some data value is reflected in complex correlations or human judgment; verification through predetermined algorithms alone may not fully reflect its value, raising questions about the credibility of the transaction data. Moreover, the verification process based on multi-party secure computation involves frequent communication between a large number of network nodes, which is several orders of magnitude slower than plaintext computation. For large-scale datasets or complex model training, the time consumption and cost of the verification phase can be very high. In addition, this technology is only designed for one-to-one transaction order binding and matching tasks, making it difficult to match potential sellers with data buyers. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the existing technology and propose a privacy data transaction order matching method and system based on the intersection of cardinality privacy sets, so as to improve the credibility and effectiveness of order matching in transactions, realize efficient one-to-many privacy-protected transaction order matching, and screen potential data sellers for data buyers.

[0008] To achieve the above objectives, the technical solution of the present invention includes:

[0009] 1. A privacy-preserving data transaction order matching method based on cardinality privacy set intersection PSI-CA, characterized in that it includes:

[0010] (1) Initialize data for the buyer , Data seller for Cloud data exchange is Data Buyer Purchase preferences set Data seller Set of data attributes to be sold Cryptographic hash functions Safety parameters and order matching threshold ;

[0011] (2) Data seller Select random number Embed it into the data attribute value From the hash value, to obtain a pseudo-random value and the generated random values ​​are paired As the selected value inadvertently transmitted OT (On-Time) data when choosing between two options, it is sent to the cloud data exchange. ;

[0012] (3) Data Buyer Select random number As an unintentional query sent to the cloud data exchange Cloud Data Exchange According to the query right Perform bit-by-bit selection and retrieve the query results. Returned to the data buyer , The query results are concatenated to obtain the unintentionally transmitted results. ;

[0013] (4) Data seller For each attribute in the metadata collection Generate key-value pairs And by using unintentional key-value pairs to encode the stored OKVS, the object is obtained. Send to cloud data exchange ;

[0014] (5) Data Buyer Based on their own preferred hash value And casual search results Calculate shared value Then, the shared value and the purchase preference hash value Send to cloud data exchange ;

[0015] (6) Cloud Data Exchange According to the received and Perform OKVS decoding to obtain the set of decoded values ​​for all data sellers. For each decoded value in the set With preference hash value Perform an XOR operation to filter eligible data sellers for data buyers.

[0016] Furthermore, the cloud data exchange mentioned in (6) according to and Perform OKVS decoding. Its implementation includes:

[0017] Input data seller Data attributes The corresponding OKVS encoding object ;

[0018] Input data buyer Shared values As the key for OKVS;

[0019] Using keys For encoded objects Obtain the decoded value by performing OKVS decoding. :

[0020] Furthermore, the cloud data exchange mentioned in (6) For each decoded value in the set With preference hash value Performing an XOR operation to filter data sellers who meet the criteria for data buyers can be achieved through the following methods:

[0021] For the set of decoded values Each decoded value in With preference hash value Perform an XOR operation to obtain the result. ;

[0022] Statistical XOR results The number of zero values ​​in the median is denoted as . This value is the cardinality of the intersection;

[0023] Will With threshold Compare the results to determine if the orders were successfully matched:

[0024] like Then the data seller With data buyers The transaction order has been successfully matched; proceed to the next step.

[0025] Otherwise, the order matching will fail;

[0026] For the matched sellers Calculate the matched gain This will serve as a basis for fair pricing in data transactions.

[0027] 2. A privacy-preserving data transaction order matching system based on cardinality privacy set intersection PSI-CA, characterized in that it includes:

[0028] The initialization module is used to initialize the data buyer. , Data sellers Cloud Data Exchange Data Buyer Purchase preferences set Data seller The set of data attributes for sale Cryptographic hash functions Safety parameters and order matching threshold ;

[0029] Unintentionally transmitted parameter generation module, used by the seller Generate attribute values The two-choice unintentional transmission of OT random value pairs ;

[0030] Unintentional transmission module, for the buyer Based on random query For random value pairs Perform bit-by-bit inadvertent selection to obtain inadvertent query results. ;

[0031] Shared value generation module, for the buyer Results from an unintentional search Generate shared values ;

[0032] Unintentional key-value pair storage module, used by sellers For attribute values Perform unintentional key-value pair storage OKVS encoding to obtain the encoded object. ;

[0033] The transaction order matching module is used by the seller. Data attribute values OKVS decoding, buyer Preference set With the seller The set of data attributes for sale Cardinality of the intersection Solve for the cardinality. With threshold The comparison is used to determine whether the transaction orders match.

[0034] Compared with the prior art, the present invention has the following advantages:

[0035] Firstly, in transactions, this invention protects the privacy of transmitted data and transaction order matching by employing a cardinality privacy set intersection method that combines unintentional transmission of OT with unintentional key-value pair storage OKVS technology. It does not disclose any information other than the intersection cardinality to the exchange, thus making up for the shortcomings of existing technologies in privacy protection, improving the credibility of data transactions, and promoting the compliant circulation and supervision of transaction data.

[0036] Secondly, this invention employs cardinality privacy set intersection technology to efficiently calculate the cardinality of the intersection between the seller's data attribute set and the buyer's purchase preference set, and compares it with the order matching degree threshold to screen qualified data sellers for data buyers. Compared with existing one-to-one privacy data transaction order binding matching technology, this invention fully considers the idea of ​​"comparing three options before buying" in actual transactions, and can achieve flexible and efficient screening of potential qualified sellers. Attached Figure Description

[0037] Figure 1 The flowchart below shows the implementation of the privacy-preserving data transaction order matching method of the present invention, which involves finding the intersection of cardinality privacy sets (PSI-CA).

[0038] Figure 2 This is a schematic diagram illustrating the interaction between the data seller, the data buyer, and the cloud data exchange in the method of the present invention;

[0039] Figure 3 The diagram below shows the privacy-preserving data transaction order matching system of the PSI-CA, which is based on the intersection of the cardinality privacy set of this invention. Detailed Implementation

[0040] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0041] Example 1: A privacy-preserving data transaction order matching method based on the intersection of cardinality privacy sets (PSI-CA).

[0042] Reference Figure 1 The implementation steps of this example include the following:

[0043] Step 1: Initialize data transaction order matching parameters.

[0044] Initial data buyer is , Data seller for and cloud data exchange for Data Buyer The set of purchase preferences is Data seller The attribute set of the data to be sold is ;

[0045] Set the cryptographic hash function to And set a threshold for the cardinality of the intersection of privacy sets PSI. As the order matching threshold, where , , For safety parameters.

[0046] In this embodiment, the security parameters are set according to the GM / T 0028-2024 cryptographic security standard. The cryptographic hash function is set to the secure hash algorithm SHA-256. Seller Data attribute set and buyer The cardinality of the purchase preference set is Based on the cardinality of the data attribute set and preference set Set the order matching threshold Set as .

[0047] In this embodiment, the data seller The dataset used is the MedMNISTv2 medical image dataset, where each sample in the dataset is... Images, data buyer A lightweight convolutional neural network (CNN) with 232,089 model parameters was trained to perform a medical image classification task. The model consists of three convolutional layers and two fully connected layers. Each convolutional layer contains a stride of 1 and padding of 1. Convolution kernel, Square activation function, and stride of 2 Average pooling, which also introduces in the first layer The convolutional kernel and residual connections with a stride of 2 are used to adapt to the size after the first convolutional layer. The input layer feature dimension is... Each fully connected layer outputs the category dimension. .

[0048] Step 2: Generate the two-option unintentional transmission parameters.

[0049] Data seller Select random number Embed it into the seller's data attribute value hash value To calculate pseudo-random values and pair the random values As an unintentional transmission of OT value in a choice between two options, it is uploaded to the cloud data exchange. .

[0050] In this embodiment, the data seller Employing a two-choice OT extension method for batch transmission of attribute values ​​from data to be sold. Generated pseudo-random value pairs By hiding seller data attribute values Protect data privacy and achieve efficient, unintentional transmission through parallel data processing.

[0051] Step 3: Obtain the results of the unintentional transmission.

[0052] Data Buyer Select random number This was treated as an unintentional query for a pseudo-random value and sent to the cloud data exchange. During each unintentional search, Based on random query For random value pairs Perform bit-by-bit selection and retrieve the query results. Returned to the data buyer ;

[0053] Data Buyer Results of cascading queries Calculate the result of intentional transmission :

[0054] .

[0055] Reference Figure 2 In this embodiment, the data seller and data buyers Interacting via unintentional transmission of the OT protocol through a choice between two options:

[0056] If data exchange Or data buyer It can only receive the selected value of the OT protocol inadvertently transmitted between two options. and random query ,and If all values ​​are pseudo-random, then the data buyer... Only random numbers can be used Perform random bit selection: when The dimensional bits are When, choose The corresponding bit position can be obtained ;when The dimensional bits are When, then choose The corresponding bit position can be obtained Regardless of the choice, the result is an indistinguishable random value; therefore, the data buyer... Unable to spy on the data seller Data attribute values This can protect the seller. Security of data attributes;

[0057] If the data buyer Through random query Bit-by-bit selection of data sellers pseudo-random value pairs Get query results If so, the query process is unintentional. Unable to distinguish data buyers The query result received bit by bit still This can protect the data buyer. Privacy of query results.

[0058] Step 4: Generate shared values.

[0059] Data Buyer Using casual search results Calculate shared value :

[0060] ;

[0061] Data Buyer Share its value Used as the key for subsequent unintentional key-value pair storage of OKVS decoding, and hashes of shared values ​​and preference values. Send to cloud data exchange Used for matching data transaction orders.

[0062] In this embodiment, shared values There are two scenarios:

[0063] If the data seller's attribute value Purchase preferences of data buyers Consistency, that is ,but The shared value A correct OKVS key is required to enable correct OKVS decoding.

[0064] Otherwise, the attribute values ​​of the data seller Purchase preferences of data buyers Inconsistency, i.e. ,but The value is random, which prevents OKVS from decoding correctly.

[0065] Step 5: Perform unintentional key-value pair storage encoding.

[0066] 5.1) Seller For data attribute collection Each attribute value in Generate key-value pairs And use unintentional key-value pair storage for OKVS pairs Encode to obtain the encoded object :

[0067] ,

[0068] in, For OKVS encoding functions, random values hash value For the input key and attribute value of the encoding function hash value The input value for the encoding function;

[0069] 5.2) Seller Encoded object Send to cloud data exchange .

[0070] In this embodiment, the cloud data exchange Only OKVS encoded objects can be received and hashes of shared values ​​and preference values Because of this transmitted value They are all random, and cannot distinguish between OKVS encoded objects from different sellers. Therefore, OKVS encoding satisfies both randomness and inadvertent security.

[0071] Step 6: Match data transaction orders.

[0072] This step is performed by the Cloud Data Exchange. Based on the received encoded object and hashes of shared values ​​and preference values The process of matching transaction orders and selecting suitable data sellers includes:

[0073] 6.1) Cloud Data Exchange Calculate the decoded value using the decoding function. , obtain the value set :

[0074] ,

[0075] in, This is the OKVS decoding function. This is the input key for the decoding function. The input encoded value for this decoding function, ;

[0076] 6.2) Cloud Data Exchange For data sellers OKVS decoded value set Each decoded value Calculate its decoded value With preference hash value XOR result The number of values ​​with zeros is counted, and this number is used as the PSI base, denoted as . ;

[0077] 6.3) Cloud Data Exchange PSI base With the set threshold Compare the results to determine if the orders were successfully matched:

[0078] like Then the data seller With the buyer The data transaction order was successfully matched; proceed to step 6.4).

[0079] Otherwise, the data transaction order matching will fail;

[0080] 6.4) Cloud Data Exchange For the matched seller Calculate the matched gain It is used for fair pricing in the data transaction process.

[0081] In this embodiment, the correctness of OKVS decoding depends on the input key. The correctness;

[0082] If you press the key Random value selected by the seller hash ,Right now

[0083] The corresponding OKVS decoding result This value is the correct decoded value;

[0084] Otherwise, the input key is The decoded value obtained by OKVS decoding is a random value. It is a random value.

[0085] During the decoding process, due to The decoded values ​​are indistinguishable from those of other random key outputs, therefore the data seller... Aside from knowing the matching result, no other buyers can be obtained. Purchase preference information; because all transmitted values ​​between the two parties in a data transaction are pseudo-random, cloud data exchanges... In addition to obtaining the PSI base of the matching results Furthermore, it is impossible to obtain any private information about the two parties involved in the transaction.

[0086] It should be noted that the step numbers in the specification and claims of this invention are only for the purpose of clearly describing the embodiments of this invention and facilitating the understanding of the invention, and their order is not limited.

[0087] Example 2: Privacy-Preserving Data Transaction Order Matching System Based on Cardinality Privacy Set Intersection (PSI-CA)

[0088] Reference Figure 3 This example includes: an initialization module 1, an unintentional transmission parameter generation module 2, an unintentional transmission module 3, an unintentional key-value pair storage encoding module 4, a shared value generation module 5, and a transaction order matching module 6. The initialization module 1 includes: a data set initialization submodule 11 and a system parameter initialization submodule 12; the transaction order matching module 6 includes: an OKVS decoding submodule 61, a PSI-CA solving submodule 62, and a threshold comparison submodule 63.

[0089] The working principle of the entire system is as follows:

[0090] The initialization module 1 is used to initialize the data buyer. , Data seller for Cloud data exchange is Purchase Preference Set Set of attributes for data to be sold Cryptographic hash functions Safety parameters and order matching threshold The data set initialization submodule 11 is used to initialize the data for the buyer. Initialize the purchase preference set and for data sellers Initialize the set of data attributes to be sold System parameter initialization submodule 12 is used to set the cryptographic hash function. Safety parameters and order matching threshold These parameters are then transmitted to the following modules: Unintentional Parameter Generation Module 2, Unintentional Transmission Module 3, Shared Value Generation Module 4, Unintentional Key-Value Pair Storage Encoding Module 5, and Transaction Order Matching Module 6.

[0091] The unintentional transmission parameter generation module 2 is used by the seller. According to the set given in Module 11 Attribute values ​​in and the cryptographic hash function given in Module 12 and safety parameters Generate random value pairs for accidental transmission of OT (On-The-Air) data. and pair the random value with Send to the unintentional transmission module 3;

[0092] The unintentional transmission module 3 is used by the buyer. According to the safety parameters given in module 12 Generate random query The random value pair given by module 2 is selected bit by bit. and the unintentional query results obtained The random value pairs are transmitted to the shared value generation module 4. Transmitted to the unintentional key-value pair storage encoding module 5;

[0093] The shared value generation module 4 is used by the buyer. Using the cryptographic hash function given in Module 12 With safety parameters and the unintentional query results transmitted by module 3 Generate shared values The hash set of shared values ​​and preference values Transmitted to transaction order matching module 6;

[0094] The unintentional key-value pair storage encoding module 5 is used by the seller. Using the cryptographic hash function given in Module 12 With safety parameters and random values ​​in module 3 hash For the seller Attribute value hash Unintentionally storing key-value pairs using OKVS encoding yields the encoded object. , encoding object Send to transaction order matching module 6;

[0095] The transaction order matching module 6 is used by the cloud data exchange to match orders according to the cryptographic hash function provided by module 12. Safety parameters and order matching threshold OKVS decoding and purchase preference set With the set of data attributes for sale Cardinality of the intersection Solve for the cardinality and assign it to the cardinality. With threshold A comparison is performed to determine whether the transaction orders match. Specifically, the OKVS decoding submodule 61 determines the match based on the shared value given by module 4 and the hash of the preference value. and the encoding object given in module 5 For each attribute value of the seller data Decode the value and then convert the decoded value. The data is transmitted to module 62; the PSI-CA solving submodule 62 then uses the decoded value provided by submodule 62... The preference hash value given in Module 4 Calculate the XOR value And count the number of zero values ​​to obtain the buyer's information. Preference set With the seller The set of data attributes for sale Cardinality of the intersection And transmit it to module 63; threshold comparison submodule 63, and use the intersection cardinality given by submodule 62. and the order matching threshold given by submodule 12 The comparison is performed to determine whether the transaction orders match.

[0096] It should be noted that the above functional modules can be implemented entirely or partially through software, hardware, or any combination thereof. When implemented in software, they can be implemented entirely or partially in the form of program instruction products. A program instruction product includes one or a set of program instructions. When a computer loads and executes the program instructions, all or part of the described process or function is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Program instructions can be stored in a computer-readable and writable storage medium, or transferred from one computer's readable and writable storage medium to another.

[0097] In this embodiment, the direct coupling or communication connection between the modules can be achieved through indirect coupling or communication connection via interfaces, devices, or modules. The functional modules and sub-modules in this embodiment can dynamically reside within a single processing unit, or each module can exist physically independently, or two or more modules can dynamically reside within a single processing unit. When these dynamic components are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable and writable storage medium. This storage medium can be a memory, disk, or optical disc, etc.

[0098] The above descriptions are merely two specific examples of the present invention and do not constitute any limitation on the present invention. Obviously, those skilled in the art, after understanding the content and principles of the present invention, may make various modifications and changes in form and detail without departing from the principles and structure of the present invention. For example, the unintentional transmission OT technology used in steps 2 and 3 above can be replaced with the unintentional pseudo-random function OPRF method; the unintentional key-value pair OKVS encoding method used in step 5 above can be replaced with the vector unintentional linear estimation VOLE method; the calculation of matching gain in step 6 above... It can be replaced with other cardinality that intersects with it. and threshold The associated normalization expression, however, these modifications and changes based on the ideas of this invention are still within the scope of protection of the claims of this invention.

Claims

1. A privacy-preserving data transaction order matching method based on cardinality privacy set intersection PSI-CA, characterized in that, include: (1) Initialize data buyer , Data sellers Cloud Data Exchange Data Buyer Purchase preferences set Data seller Set of data attributes for sale Cryptographic hash functions Safety parameters and order matching threshold ; (2) Data seller Select random number Embed it into the data attribute value From the hash value, to obtain a pseudo-random value and the generated random values ​​are paired As the selected value inadvertently transmitted OT (On-Time) data when choosing between two options, it is sent to the cloud data exchange. ; (3) Data Buyer Select random number As an unintentional query sent to the cloud data exchange Cloud Data Exchange According to the query right Perform bit-by-bit selection and retrieve the query results. Returned to the data buyer , The query results are concatenated to obtain the unintentionally transmitted results. ; (4) Data seller For each attribute in the metadata collection Generate key-value pairs And by using unintentional key-value pairs to encode the stored OKVS, the object is obtained. Send to cloud data exchange ; (5) Data Buyer Based on their own preferred hash value And casual search results Calculate shared value Then, the shared value and the purchase preference hash value Send to cloud data exchange ; (6) Cloud Data Exchange According to the received and Perform OKVS decoding to obtain the set of decoded values ​​for all data sellers. For each decoded value in the set With preference hash value Perform an XOR operation to filter eligible data sellers for data buyers.

2. The method according to claim 1, characterized in that, The initialization data buyer in (1) , Data sellers It is the buyer of the initial data. Purchase preferences set and data sellers Set of data attributes for sale They are represented as follows: ， ， in, , , .

3. The method according to claim 1, characterized in that, The cryptographic hash function in (1) Safety parameters and order matching threshold The settings are as follows: According to the GM / T 0028-2024 cryptographic security standard, the security parameters are set as follows: , The cryptographic hash function is set to the secure hash algorithm SHA-256. , Based on the cardinality of the data attribute set and preference set Set the order matching threshold Set as If the cardinality of the intersection exceeds the threshold, then the order is matched.

4. The method according to claim 1, characterized in that: The pseudo-random value obtained in (2) It is a random value selected randomly. With data attribute values hash value Calculated, and expressed as: ,in, and It is a random bit string of length 128; The data buyer in (3) The query results for each bit are concatenated to obtain the unintentional query result. The formula is: ,in, It is a random bit string of length 128.

5. The method according to claim 1, characterized in that: The data seller in (4) OKVS will use unintentional key-value pair storage The encoding is performed using the following formula: ,in, and It is a bit string of length 128; The data buyer in (5) Based on their own preferred hash value And casual search results Calculate shared value Its formula is: .

6. The method according to claim 1, characterized in that, The cloud data exchange mentioned in (6) According to the received and OKVS decoding is implemented by including: 6a) OKVS encoded object of the seller's data attributes of the input data ; 6b) Input data: Shared values ​​of the buyer As the key for OKVS; 6c) Using keys For encoded objects Obtain the decoded value by performing OKVS decoding. : ， in, , For data sellers The set of decoded values; Should The decoded value contains two cases: If the data seller's attribute value Purchase preferences of data buyers Consistency, that is ,but The shared value The correct key for OKVS is used to obtain the correct decoding result. ; otherwise, The value returned by OKVS decoding is a random value. It is a random value.

7. The method according to claim 1, characterized in that, The cloud data exchange mentioned in (6) For each decoded value in the set With preference hash value Performing an XOR operation to filter data sellers who meet the criteria for data buyers can be achieved through the following methods: 6d) For the set of decoded values Each decoded value in With preference hash value Perform an XOR operation to obtain the result. ; 6e) Statistical XOR results The number of zero values ​​in the median is denoted as . This value is the cardinality of the intersection; 6f) Will With threshold Compare the results to determine if the orders were successfully matched: like Then the data seller With data buyers The transaction order was successfully matched and will be executed in 6 hours. Otherwise, the order matching will fail; 6g) for the matched seller Calculate the matched gain: This will serve as a basis for fair pricing in data transactions.

8. A privacy-preserving data transaction order matching system based on cardinality privacy set intersection PSI-CA, characterized in that, include: The initialization module is used to initialize the data for the buyer. , Data seller for Cloud data exchange is Data Buyer Purchase preferences set Data seller The set of data attributes for sale Cryptographic hash functions Safety parameters and order matching threshold ; Unintentionally transmitted parameter generation module, used by the seller Generate attribute values The two-choice unintentional transmission of OT random value pairs ; Unintentional transmission module, for the buyer Based on random query For random value pairs Perform bit-by-bit inadvertent selection to obtain inadvertent query results. ; Shared value generation module, for the buyer Results from an unintentional search Generate shared values ; Unintentional key-value pair storage encoding module, used by the seller For attribute values Perform unintentional key-value pair storage OKVS encoding to obtain the encoded object. ; The transaction order matching module is used by the seller. Data attribute values OKVS decoding, buyer Preference set With the seller The set of data attributes for sale Cardinality of the intersection Solve for the cardinality. With threshold The comparison is used to determine whether the transaction orders match.

9. The system according to claim 8, characterized in that, The initialization module includes: The data set initialization submodule is used to initialize the data buyer. Purchase preferences set and data sellers Set of data attributes for sale ; The system parameter initialization submodule is used to initialize the cryptographic hash function in the system. Safety parameters and order matching threshold .

10. The system according to claim 8, characterized in that, The transaction order matching module includes: The OKVS decoding submodule is used to decode each attribute value of the seller data. Perform OKVS decoding; The PSI-CA solver submodule is used to solve the buyer's... Preference set and the seller The set of data attributes for sale Cardinality of the intersection ; The threshold comparison submodule is used for intersection cardinality. With threshold The comparison between them is used to determine whether the transaction orders match.

Images