A user privacy data encryption protection method of a commodity supply chain platform

Abstract

This application relates to the field of data encryption technology, specifically to a method for encrypting and protecting user privacy data on a commodity supply chain platform. The method includes: analyzing the membership vector of each user based on user privacy data obtained from the commodity supply chain platform; constructing a mapping function for each type of privacy data for each user in each cluster based on the membership degree between the user and each cluster, and the differences between each user and other users in each cluster in each type of privacy data, and combining this with a key obtained by encrypting all user privacy data within each cluster; performing blind mapping on each type of privacy data for each user in each cluster; and encrypting all blinded user privacy data in each cluster. This application achieves finer-grained access control and a stronger data protection mechanism without sacrificing excessive performance.

Description

Technical Field

[0001] This application relates to the field of data encryption technology, specifically to a method for encrypting and protecting user privacy data on a commodity supply chain platform. Background Technology

[0002] With the rapid development of e-commerce and globalized trade, commodity supply chain platforms play a crucial role in the modern economy. These platforms not only connect manufacturers, suppliers, distributors, and consumers, but also handle large amounts of user privacy data. Existing supply chain platforms typically utilize layered encryption to uniformly encrypt similar types of data or highly similar user data, ensuring platform data security while reducing platform overhead and improving operational efficiency.

[0003] However, these methods typically encrypt user data of the same type or with high similarity uniformly, thus sacrificing some security while improving efficiency, reducing the system's fine-grained access control capabilities, and increasing the possibility of data leakage. Theoretically, encrypting each user separately offers the highest security, but it significantly increases the system's burden, and maintaining a large key space also increases storage overhead, affecting the overall system performance.

[0004] Existing solutions struggle to achieve finer-grained access control and stronger data protection mechanisms without sacrificing too much performance, which to some extent limits the overall security of supply chain platforms. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a method for encrypting and protecting user privacy data on a commodity supply chain platform, thereby resolving the existing problems.

[0006] The method for encrypting and protecting user privacy data in a commodity supply chain platform according to this application adopts the following technical solution:

[0007] One embodiment of this application provides a method for encrypting and protecting user privacy data on a commodity supply chain platform, the method comprising the following steps:

[0008] S1, obtains user privacy data from the commodity supply chain platform;

[0009] S2, based on the feature vectors of all users' transaction records in the user privacy data, use the fuzzy clustering algorithm to cluster users and output the membership vector of each user;

[0010] S3. Based on the membership degree between the user and each cluster, and the difference between each user and other users in each cluster in each type of privacy data, determine the blinding coefficient vector of each user in each cluster for each type of privacy data.

[0011] S4, combining the key obtained by encrypting all user privacy data in each cluster, the membership vector of each user, and the blinding coefficient vector, construct a mapping function for each type of privacy data of each user in each cluster, and perform blinding mapping on each type of privacy data of the user;

[0012] S5 encrypts all user privacy data after blinding in each cluster.

[0013] Preferably, the user privacy data includes three categories of privacy data: basic user identity information, user transaction records, and sensitive user identity information.

[0014] Preferably, the user's basic identity information includes name, gender, age, and contact information; the user's sensitive identity information includes personal ID number and bank card number; and the user's transaction record includes transaction amount, category of goods traded, and identities of both parties to the transaction.

[0015] Preferably, the feature vector of the user transaction record consists of each user's total historical transaction amount, transaction product category, and the identities of both parties in the transaction.

[0016] Preferably, the clustering distance when using the fuzzy clustering algorithm is the distance between feature vectors.

[0017] Preferably, the method for determining the blinding coefficient vector for each user and each type of privacy data in each cluster is as follows:

[0018] The first Users within a cluster The first in the privacy data The blinding coefficient vector of privacy-preserving data is denoted as , ;

[0019] in, For users Belonging to the The membership degree of each cluster. For users The sum of the elements in the membership vector. For the first Users in each cluster The first with the other users A vector of differences between privacy-preserving data.

[0020] Preferably, ,in For the first The number of attributes in privacy-preserving data User With the The remaining users in the cluster The mean of the first attribute data difference measure in privacy-preserving data. , Corresponding to the 2nd and 3rd respectively The mean of the difference measure of attribute data.

[0021] Preferably, the method for constructing a mapping function for each type of privacy data for each user in each cluster is as follows:

[0022] The first Users within a cluster The first in the privacy data The mapping function for privacy-preserving data is denoted as , ;

[0023] in, For the first Users within a cluster The first in the privacy data Blinding coefficient vector of privacy-preserving data User The A column vector composed of privacy-preserving data. User The transpose of the membership vector. User The The dimension of the column vector composed of privacy-preserving data. User The The first column vector composed of privacy-preserving data Dimensional data, It is the first The key is obtained by encrypting the privacy data of all users within a cluster.

[0024] Preferably, the membership vector and the blinding coefficient vector are continuously updated as user privacy data changes in the supply chain platform.

[0025] Preferably, when decrypting user privacy data, it is necessary to input the decryption key, membership vector and blinding coefficient vector of the corresponding cluster unified encryption, first use the decryption key to decrypt and obtain the blinded user privacy data, and then use the mapping function to restore the blinded user privacy data to the original user privacy data.

[0026] This application has at least the following beneficial effects:

[0027] This application firstly utilizes fuzzy clustering and membership calculation to simultaneously process the privacy data of all users, avoiding the overhead of generating keys for each user and the storage burden caused by managing and updating keys, thus ensuring system efficiency. Secondly, this protection method also has unique advantages in terms of timeliness. As user privacy data changes in the system, the membership vector and blinding coefficient vector also change in real time. Therefore, the membership vector and blinding coefficient vector obtained by the user can only play a role within a limited time. After fuzzy clustering is performed again, the membership vector and blinding coefficient vector will also be updated, thus realizing the functions of permission revocation and key update. Furthermore, the blinding coefficient vector can adaptively adjust the blinding coefficient vector according to the membership between the user and each cluster, controlling the degree of blinding to further reduce the probability of successful adversary attacks. Finally, the membership vector of each user can be uniformly obtained through fuzzy clustering, avoiding the overhead caused by processing each user individually, and also avoiding the huge key space management when encrypting each user separately. In this method, the system does not need key management and updates, but only needs to use mapping functions to achieve similar functions, thereby reducing the system's storage overhead. In summary, this application achieves finer-grained access control and stronger data protection mechanisms without sacrificing too much performance. Attached Figure Description

[0028] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 A flowchart illustrating a method for encrypting and protecting user privacy data on a commodity supply chain platform, as provided in this application. Detailed Implementation

[0030] To further illustrate the technical means and effects adopted by this application to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a user privacy data encryption protection method for a commodity supply chain platform proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0032] The following description, in conjunction with the accompanying drawings, details a specific scheme for encrypting and protecting user privacy data on a commodity supply chain platform provided in this application.

[0033] One embodiment of this application provides a method for encrypting and protecting user privacy data on a commodity supply chain platform.

[0034] Specifically, the following method for encrypting and protecting user privacy data on a commodity supply chain platform is provided. Please refer to [link / reference]. Figure 1 The method includes the following steps:

[0035] S1 obtains user privacy data from the commodity supply chain platform.

[0036] Commodity supply chain platforms contain various types of user privacy data, which can be roughly divided into basic user identity information, user transaction records, and sensitive user identity information. Among them, basic user identity information includes, but is not limited to, name, gender, age, and contact information. This type of information can be directly quantified in text form. For example, the name can be converted into a string, and 0 or 1 can be used to represent different genders. The specific conversion methods are well-known in the field of data processing and are not subject to specific restrictions, as long as various data can be converted into quantifiable numerical forms.

[0037] Sensitive user identity information includes, but is not limited to, personal ID numbers and bank card numbers; user transaction records include the user's transaction amount, the category of goods traded, and the identities of both parties in the transaction. Because basic user identity information and sensitive identity information are unique, they cannot be used for user classification on the supply chain platform; however, user transaction records can be used for user classification.

[0038] In the user's transaction record, the user's transaction amount is stored in numerical form. At the same time, the user's transaction product category is quantified using a one-hot encoding method, and different product category codes can be assigned to each product category in sequence.

[0039] A user's transaction amount is arranged in order of the size of the corresponding product category code to form a sequence of the user's transaction amount;

[0040] Secondly, the identities of the two parties to the transaction include, but are not limited to, suppliers, supplier accounts, sellers, and seller accounts. A supplier and its account together constitute a supplier vector, and a seller and its account together constitute a seller vector. All supplier vectors in a user's transaction records are arranged in order of account size to form the user's supplier sequence, and all seller vectors are arranged in order of account size to form the user's seller sequence.

[0041] S2, based on the feature vectors of all users' transaction records in the user privacy data, uses a fuzzy clustering algorithm to cluster users and outputs the membership vector of each user.

[0042] Large supply chain platforms possess vast amounts of user privacy data, which is highly real-time—meaning the content of each user's privacy data is constantly changing as transactions proceed. To maintain data privacy, supply chain platforms need to frequently encrypt user data. Therefore, supply chain platforms typically employ layered encryption, encrypting different types of data separately. To ensure data consistency and efficient data management, they may also uniformly encrypt related or similar data to improve system efficiency. However, this weakens granularity; when a user gains access to a specific set of data, they actually gain access to a whole group of data, violating the principle of least privilege and threatening system information security. Therefore, it is necessary to further enhance fine-grained access control based on current encryption management methods to improve the security of supply chain platforms.

[0043] When encrypting privacy data on a supply chain platform, the first step is to categorize users, managing those with certain relevance or similarity to improve platform efficiency. However, traditional data categorization methods have limitations. The variability of user data during categorization can further threaten system security. For example, a user who is both a supplier and a distributor might be associated with both. Grouping all data related to this user together would result in overly large categories, allowing each user to access a large amount of irrelevant or weakly related data – a serious security vulnerability. Conversely, if a user is fixed to a single category, users in other categories might not be able to associate with them, leading to vulnerabilities in data management.

[0044] The unique hot code of the transaction product category, the transaction amount sequence, the supplier sequence, and the seller sequence in each user's transaction record are combined to form the feature vector of a user's transaction record.

[0045] All user feature vectors, and the set number of clusters. and fuzzy factor As input, a fuzzy clustering algorithm is used to calculate membership degrees, and the output is a membership degree matrix, in which the number of clusters is... The fuzzy factor can be set according to system needs or determined using the elbow method. The more clusters there are, the greater the system load. The recommended value range is [1,2]. Fuzzy clustering algorithms and elbow methods are well-known techniques, and their specific details will not be elaborated here. Among different implementation methods, fuzzy clustering algorithms can be selected from fuzzy C-means clustering, fuzzy K-means clustering, and Gustafson-Kessel (GK) clustering.

[0046] It is important to note that the one-hot encoding of a transaction category is a string of binary data. The size of the data itself has no meaning, but each bit, whether 0 or 1, indicates whether there is a corresponding transaction category, which is meaningful. Therefore, in different implementations, the one-hot encoding of the transaction category needs to use edit distance or Hamming distance when calculating distance. The supplier sequence and the seller sequence may have different lengths, so edit distance or DTW distance can be used in different implementations.

[0047] Different distance formulas yield different distance values. In this embodiment, the distance between data of the same dimension in two user feature vectors is calculated separately, and the calculated distances are normalized. Finally, the sum of all normalized distances is the distance formula for fuzzy clustering. Edit distance, Hamming distance, DTW distance, and normalization are well-known techniques and will not be elaborated further.

[0048] S3. Based on the membership degree between the user and each cluster, and the differences between each user and other users in each cluster in each type of privacy data, determine the blinding coefficient vector for each user in each cluster for each type of privacy data.

[0049] After obtaining the membership matrix in step S2, the membership vector for each user is obtained. This vector measures the degree to which each user's information belongs to each cluster, and the membership degree changes as user information changes throughout the user space. In other words, the membership vector is time-sensitive; and it remains relatively secure unless an adversary has cracked all the clusters to which a user belongs, and all the user privacy data within those clusters.

[0050] Therefore, user data can be blinded using each user's membership vector, thereby further ensuring the security of user privacy data. At the same time, membership vectors can be used to achieve more granular access control.

[0051] Secondly, since each user's membership vector indicates the degree of correlation between the user's data and the elements in each cluster, a low membership degree between a user and a cluster indicates a low correlation between the user and other elements in that cluster. Therefore, a smaller blinding coefficient can be used during blinding to control the difference in user data before and after blinding. Conversely, a high membership degree between a user and a cluster indicates a high correlation between the user's data and other elements in that cluster. In this case, using a small blinding coefficient during blinding might result in a smaller difference in data before and after blinding, thus increasing the probability that an adversary could obtain the user's data characteristics or privacy information by analyzing the data features of other elements in the cluster, thereby increasing the possibility of privacy leakage to some extent.

[0052] Therefore, it is necessary to set a corresponding blinding coefficient vector for each type of privacy data for each user in each cluster, based on different membership situations.

[0053] In this embodiment, the blinding coefficient vector is calculated as follows:

[0054] : For the first Users within a cluster The first in the privacy data Blinding coefficient vector of privacy-preserving data For users Belonging to the The membership degree of each cluster. For users The sum of the elements in the membership vector. For the first Users in each cluster The first with the other users A vector of differences between privacy-preserving data.

[0055] In this embodiment, ,in For the first The number of attributes in privacy-preserving data User With the The remaining users in the cluster The mean of the first attribute data difference measure in privacy-preserving data, and so on. , Corresponding to the 2nd and 3rd respectively The mean of the difference measure of attribute data.

[0056] It should be noted that if the corresponding attribute data is in numerical form, the difference measure is the reciprocal of the normalized difference; if the corresponding data is in vector form, the difference measure is cosine similarity.

[0057] It is understandable that the higher the membership degree of a user in a cluster, the greater the data similarity, thus requiring a larger blinding coefficient to ensure greater data difference before and after blinding. Similarly, for each data point in a class of data, the smaller the difference between it and other elements in the same cluster, the larger the blinding coefficient is required to ensure greater difference before and after blinding, thus the corresponding blinding coefficient is relatively large.

[0058] S4. Combining the key obtained by encrypting all user privacy data within each cluster, the membership vector of each user, and the blinding coefficient vector, a mapping function is constructed for each type of privacy data of each user in each cluster, and blinding mapping is performed on each type of privacy data of the user.

[0059] In this embodiment, the mapping function is constructed as follows:

[0060] : For the first Users within a cluster The first in the privacy data Mapping functions for privacy-related data, For the first Users within a cluster The first in the privacy data Blinding coefficient vector of privacy-preserving data User The A column vector composed of privacy-preserving data. User The transpose of the membership vector. User The The dimension of the column vector composed of privacy-preserving data. User The The first column vector composed of privacy-preserving data Dimensional data, It is the first The key is obtained by encrypting the privacy data of all users within a cluster.

[0061] Among them, for the first The key obtained by encrypting the privacy data of all users within a cluster can be obtained through the symmetric encryption algorithm AES. The specific encryption algorithm can be selected by the implementer.

[0062] It should be noted that the blinding coefficient vector The dimension is equal to After the blinding coefficient vector is generated, it is sent to the user along with the membership vector. The privacy data types in this embodiment include basic user identity information, user transaction records, and sensitive user identity information, namely... .

[0063] Understandably, by using the mapping function constructed above, users can blind the original privacy data based on the membership vector. During subsequent decryption, in addition to the unified encryption key, the user's own membership vector is also required. This improves the fine-grained access control capabilities of the supply chain platform on top of traditional layered encryption. This ensures that only user mapping data within the same cluster can be obtained during subsequent decryption. The membership vector guarantees that only data of the corresponding user in the cluster can be obtained, thus achieving fine-grained access control. The blinding coefficient vector further enhances the fine-grained access control capability and can control the specific data type obtained.

[0064] Based on the mapping function constructed above, each type of privacy data of each user in the supply chain platform is blinded to obtain the blinded user privacy data.

[0065] S5 encrypts all user privacy data after blinding in each cluster.

[0066] After obtaining the blinded user privacy data, based on the results of fuzzy clustering, all the blinded user privacy data in each cluster is uniformly encrypted.

[0067] It should be noted that encryption targets the data after mapping by the mapping function, that is, the user's privacy data after blinding. In different implementations, encryption can use any cryptographic algorithm.

[0068] In addition, when a user performs decryption, they need to input the decryption key, membership vector, and blinding coefficient vector used for unified encryption of the corresponding cluster. First, the decryption key is used to decrypt and obtain the blinded user privacy data. Then, the mapping function is used to restore the blinded user privacy data to the original user privacy data, thereby realizing the acquisition of user privacy data.

[0069] In addition, users can delegate their permissions to others by simply passing the membership vector and the blinding coefficient vector.

[0070] Since the membership vector and blinding coefficient vector are constantly updated as user privacy data changes in the supply chain platform, and no storage management updates are required from the supply chain platform, there is no need to consider subsequent permission revocation when delegating permissions.

[0071] The above technical features constitute the preferred embodiment of this application, which has strong adaptability and the best implementation effect. Unnecessary technical features can be added or removed according to actual needs to meet the needs of different situations.

Claims

1. A method for encrypting and protecting user privacy data on a commodity supply chain platform, characterized in that, The method includes the following steps: S1, Obtain user privacy data from the commodity supply chain platform. The user privacy data includes three categories of privacy data: basic user identity information, user transaction records, and sensitive user identity information. S2, based on the feature vectors of all users' transaction records in the user privacy data, a fuzzy clustering algorithm is used to cluster the users and output the membership vector of each user. The membership vector is used to measure the membership degree of the user privacy data to each cluster and to indicate the degree of correlation between the user privacy data and the elements in each cluster. S3. Based on the membership degree between the user and each cluster, and the difference between each user and other users in each cluster in each type of privacy data, determine the blinding coefficient vector for each user in each cluster for each type of privacy data. The blinding coefficient vector is used to control the difference between user privacy data before and after blinding. S4, combining the key obtained by encrypting all user privacy data in each cluster, the membership vector of each user, and the blinding coefficient vector, construct a mapping function for each type of privacy data of each user in each cluster, and perform blinding mapping on each type of privacy data of the user; S5 encrypts all user privacy data after blinding in each cluster; The method for determining the blinding coefficient vector is as follows: The first Users within a cluster The first in the privacy data The blinding coefficient vector of privacy-preserving data is denoted as , ; in, For users Belonging to the The membership degree of each cluster. For users The sum of the elements in the membership vector. For the first Users in each cluster The first with the other users A vector of differences between privacy-preserving data.

2. The method for encrypting and protecting user privacy data on a commodity supply chain platform as described in claim 1, characterized in that, The user's basic identity information includes name, gender, age, and contact information; the user's sensitive identity information includes personal ID number and bank card number; the user's transaction record includes transaction amount, category of goods traded, and identities of both parties to the transaction.

3. The method for encrypting and protecting user privacy data on a commodity supply chain platform as described in claim 2, characterized in that, The feature vector of each user's transaction record consists of each user's total historical transaction amount, the category of the traded goods, and the identities of both parties in the transaction.

4. The method for encrypting and protecting user privacy data on a commodity supply chain platform as described in claim 1, characterized in that, The clustering distance when using the fuzzy clustering algorithm is the distance between feature vectors.

5. The method for encrypting and protecting user privacy data on a commodity supply chain platform as described in claim 1, characterized in that, ,in For the first The number of attributes in privacy-preserving data User With the The remaining users in the cluster The mean of the first attribute data difference measure in privacy-preserving data. , Corresponding to the 2nd and 3rd respectively The mean of the difference measure of attribute data.

6. The method for encrypting and protecting user privacy data on a commodity supply chain platform as described in claim 1, characterized in that, The method for constructing a mapping function for each class of privacy data for each user in each cluster is as follows: The first Users within a cluster The first in the privacy data The mapping function for privacy-preserving data is denoted as , ; in, For the first Users within a cluster The first in the privacy data Blinding coefficient vector of privacy-preserving data User The A column vector composed of privacy-preserving data. User The transpose of the membership vector. User The The dimension of the column vector composed of privacy-preserving data. User The The first column vector composed of privacy-preserving data Dimensional data, It is the first The key is obtained by encrypting the privacy data of all users within a cluster.

7. The method for encrypting and protecting user privacy data on a commodity supply chain platform as described in claim 1, characterized in that, The membership vector and blinding coefficient vector are continuously updated as user privacy data changes in the supply chain platform.

8. The method for encrypting and protecting user privacy data on a commodity supply chain platform as described in claim 1, characterized in that, When decrypting user privacy data, it is necessary to input the decryption key, membership vector, and blinding coefficient vector of the corresponding cluster unified encryption. First, the decryption key is used to decrypt and obtain the blinded user privacy data. Then, the mapping function is used to restore the blinded user privacy data to the original user privacy data.

Images