Transaction information security storage method based on user industry label
By converting user transaction information into three-dimensional spatial points and utilizing clustering and vector operations to construct an infinitely large key space, the problem of limited key space in existing encryption algorithms is solved, achieving efficient data encryption and storage optimization.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- JINAN FARAWAY INFORMATION TECH CO LTD
- Filing Date
- 2024-09-20
- Publication Date
- 2026-07-24
AI Technical Summary
Existing encryption algorithms have limited key space and a large number of keys in the storage of financial transaction information, which limits the encryption effect. In addition, the transaction information is highly complex, making it difficult to effectively guarantee data security and storage efficiency.
User transaction information is represented as information points in a three-dimensional space. The optimal cluster value is obtained through clustering algorithms. Vector operations and translation operations are used to construct an infinitely large key space to enhance the encryption effect. Furthermore, the positions of information points are transformed through vector operations to reduce storage space.
It achieves efficient data encryption, increases data concealment and storage space utilization, reduces the amount of transaction information stored, and improves data security and storage efficiency.
Smart Images
Figure CN119293813B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing, and more specifically to a method for securely storing transaction information based on user industry tags. Background Technology
[0002] With the rapid development of digital technology, the secure digital storage of transaction information across various industries has become a hot topic. For example, financial transaction information from financial institutions is highly complex due to its time sequence and user information. Furthermore, financial institutions often need to perform financial analysis based on the user's industry, requiring their transaction information to be categorized and stored according to user industry tags. Moreover, financial transaction information involves personal privacy; loss or malicious alteration of financial transaction data can lead to economic losses for users and damage the reputation of financial institutions, impacting their development. Therefore, financial institutions need to encrypt transaction information during storage to ensure its security.
[0003] Because financial institutions' transaction data is highly complex due to its time and user characteristics, directly storing the original transaction information results in a significant accumulation of transaction information complexity. Therefore, this invention represents the transaction information of a single user as changes in financial values over time. Simultaneously, it projects different user transaction information from the same industry onto points in three-dimensional space, increasing the relationships between the original information. The positional relationships of these spatial points are then used to encrypt the original information. Existing encryption algorithms suffer from limitations such as limited key space and a large number of keys, resulting in restricted encryption effectiveness and difficulties in key storage. Therefore, this invention utilizes the diversity of vector operations to transform the positions of spatial points and translates different spatial locations to the same relative space, reducing the storage space of the encrypted data and facilitating the storage of the encrypted data, thereby achieving secure storage of user transaction information. Summary of the Invention
[0004] This invention provides a method for securely storing transaction information based on user industry tags to solve existing problems.
[0005] The secure storage method for transaction information based on user industry tags of the present invention adopts the following technical solution:
[0006] One embodiment of the present invention provides a method for securely storing transaction information based on user industry tags, the method comprising the following steps:
[0007] Information points for obtaining transaction information of each user in the same industry;
[0008] Different cluster values are set, and all information point coordinates are clustered according to each cluster value. Each cluster value yields multiple clusters and the cluster center of each cluster. Multiple spherical shell regions of each cluster are obtained based on the distribution range of information points in each cluster corresponding to each cluster value. The mean of the information point density among the spherical shell regions in each cluster is calculated to obtain the information point density of each cluster. The optimization degree of each cluster value is obtained based on the difference in information point density among the clusters corresponding to each cluster value. The cluster value corresponding to the maximum optimization degree is taken as the optimal cluster value.
[0009] Based on the cluster center of each cluster corresponding to the optimal cluster value, as well as the reference vector and transformation coefficient set for each cluster, the vector representation of each information point in each cluster is obtained; the vector representation of each information point in each cluster is used as the first ciphertext data of the transaction information, and the cluster center, reference vector and transformation coefficient of each cluster are used as the first key of the transaction information.
[0010] The label value of each cluster is obtained based on the coordinates of the cluster center of each cluster; the translation vector of each cluster is obtained with the cluster center of each cluster as the starting point and the reference cluster center of each cluster as the ending point; the vector representation of each information point is added to the translation vector of the cluster to which each information point belongs to obtain the vector coordinates of each information point; the vector coordinates of each information point and the label value of the cluster to which each information point belongs are used as the second ciphertext data of the transaction information, and the reference cluster center and the translation vector of each cluster are used as the second key of the transaction information; the second ciphertext data of the transaction information is stored.
[0011] Preferably, the method for obtaining transaction information points for each user in the same industry is as follows:
[0012] First, user industry tags are obtained through user registration information, and user transaction information for each transaction is obtained through the transaction portal of financial institutions. The transaction information includes the transaction time and financial value of each user's transaction for each transaction.
[0013] Users are categorized into different industries based on their industry tags. A three-dimensional coordinate system is then established with the user's transaction time as the horizontal axis, the user's financial value at the time of the transaction as the vertical axis, and different users within the same industry as the vertical axis. Each transaction information of each user in the same industry is projected onto the three-dimensional coordinate system to obtain the information points corresponding to each user's transaction information in the same industry.
[0014] Preferably, the method for obtaining multiple spherical shell regions of each cluster based on the distribution range of information points in each cluster corresponding to each cluster value is as follows:
[0015] Set the sphere radius increment; obtain the maximum Euclidean distance from all information points in each cluster to the cluster center of each cluster, the maximum Euclidean distance being the distribution range of information points in each cluster; with the cluster center of each cluster as the center of the sphere, each time the sphere radius increases by one sphere radius increment, each cluster is divided once, each division yields a spherical region, thus obtaining multiple spherical regions, the number of spherical regions being the result obtained by rounding up the ratio of the maximum Euclidean distance to the sphere radius increment;
[0016] Multiple spherical regions are sorted in ascending order of sphere radius. The region obtained by subtracting each spherical region from its adjacent spherical regions is recorded as a spherical shell region, resulting in multiple spherical shell regions for each cluster. The adjacent spherical region refers to the spherical region with the largest radius among all spherical regions whose sphere radius is smaller than that of each of the spherical regions.
[0017] Preferably, the method for calculating the information point density of each spherical shell region in each cluster is as follows:
[0018] Obtain the number of information points contained in each spherical shell region, and the volume of each spherical shell region; use the ratio between the number of information points contained in each spherical shell region and the volume of each spherical shell region as the information point density of each spherical shell region.
[0019] Preferably, the expression for the degree of preference of each cluster value is:
[0020]
[0021] Among them, Q K S represents the degree of preference when the cluster value is K. K Let K be the variance between the density of information points of each cluster when the cluster value is K, and e is the natural constant.
[0022] Preferably, the method for obtaining the vector representation of each information point in each cluster is as follows:
[0023] The vector formed by taking the cluster center of each cluster as the starting point and each information point in each cluster as the ending point is used as the direction vector of each information point. The ratio of the difference between the reference vector and the direction vector of each information point to the conversion coefficient is used as the vector representation of each information point.
[0024] Preferably, the method for obtaining the label value of each cluster based on the coordinates of the cluster center of each cluster is as follows:
[0025] First, sort the cluster centers of each cluster according to their horizontal coordinates from smallest to largest, then sort them according to their vertical coordinates from smallest to largest, and finally sort them according to their vertical axis from smallest to largest. Record the index of each cluster as the label value of each cluster.
[0026] The beneficial effects of this invention are as follows: First, this invention converts the transaction information of all users in the same industry into information points in a three-dimensional space, reducing the tag information of user transaction information and reducing the storage volume of transaction information; then, it clusters the coordinates of the information points and obtains the optimal cluster value by constructing the optimization degree of different cluster values; finally, through the infinite combination of vector operations, it performs an overall translation of each cluster, realizing the position transformation of all information points in space, obtaining an infinitely large key, which greatly increases the encryption effect, performs a scrambling operation on each cluster, reduces the positional correlation of information points in space, increases data concealment, and also increases the utilization rate of storage space, which is beneficial to data storage. Attached Figure Description
[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0028] Figure 1 This is a flowchart illustrating the steps of the secure storage method for transaction information based on user industry tags according to the present invention. Detailed Implementation
[0029] To further illustrate the technical means and effects adopted by the present invention 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 the secure storage method for transaction information based on user industry tags proposed according to the present invention. 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.
[0030] 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 invention pertains.
[0031] The following description, in conjunction with the accompanying drawings, details the specific solution of the transaction information secure storage method based on user industry tags provided by this invention.
[0032] Please see Figure 1The diagram illustrates a flowchart of a method for securely storing transaction information based on user industry tags, according to an embodiment of the present invention. The method includes the following steps:
[0033] Step S001: Obtain transaction information points for each user in the same industry.
[0034] When storing transaction information for financial institutions, the user industries involved are wide-ranging, and data analysis is required for users in different industries. Therefore, it is necessary to store the transaction information of users in different industries separately. In addition, the security of transaction information must be ensured when storing data. Therefore, this invention converts transaction information into points in space and uses the positional relationship of spatial points to encrypt the original transaction information, thereby ensuring the security of transaction information.
[0035] Financial institutions generally require real-name registration for transactions. Therefore, the first step is to obtain the industry tags registered by different users during the registration process. Then, based on the financial institution's transaction portal, the transaction information of each user at each transaction is obtained, including the user's transaction time and transaction amount. In other words, the financial transaction information simultaneously obtains the user's industry tag information and each transaction information for different users.
[0036] First, users are categorized into different industries based on their industry tags. Then, the transaction information of users in each industry is processed separately: a three-dimensional coordinate system is established with the user's transaction time as the horizontal axis, the user's financial value at the time of the transaction as the vertical axis, and different users in the same industry as the vertical axis. Each user's transaction information is projected onto the three-dimensional coordinate system. At this point, each industry corresponds to a three-dimensional coordinate system, and each user's transaction information in the same industry corresponds to an information point in the three-dimensional coordinate system.
[0037] By using the above method, transaction information of different users in the same industry is projected into a three-dimensional space. All transaction information of the same user is on the same plane. Therefore, by utilizing the relationship between different planes in the three-dimensional space, the transaction information of individual users can be reflected at the same time as the transaction information of all users at different times, which is convenient for statistical analysis of transaction information. Representing the transaction information of all users in the same industry as points in space reduces redundant data and facilitates data transmission and storage. At the same time, representing the original and relatively complex information as points in space unifies the format of the original information, and using the positional relationship of spatial points to represent the correlation between the original information facilitates the encryption operation of the original information.
[0038] Step S002: Set different cluster values, and cluster the coordinates of all information points according to each cluster value; calculate the optimization degree of each cluster value to obtain the optimal cluster value.
[0039] To ensure the security of financial transaction information of financial institutions, it is necessary to encrypt user transaction information, that is, to encrypt all information points obtained in the above steps.
[0040] The encryption process of this invention is accomplished through the transformation of the spatial position of points. First, clustering is performed on the points in the current space to obtain the cluster center of each cluster. Then, the vector representation of the information points in each cluster relative to the cluster center is obtained. Next, each cluster center is used as the reference point for the position transformation of the information points in each cluster. The cluster center of each cluster is translated to the same position, and all information points in each cluster move with the cluster center, that is, the overall translation of each cluster is achieved. This enables the scrambling operation on multiple clusters, thereby improving the encryption effect.
[0041] This invention first uses the K-means clustering algorithm to cluster all information points. Since the K-means algorithm requires a preset number of clusters to be obtained after clustering, the cluster values of the K-means algorithm must be obtained first.
[0042] However, for a fixed number of information points, the smaller the cluster value, the fewer the number of clusters obtained from clustering, and the more information points each cluster contains. This results in more combinations of vector representations of all information points, greater encryption and scrambling disorder, and better encryption performance. Conversely, the larger the cluster value, the more clusters are obtained, resulting in fewer translation relationships when performing a global translation on each cluster. Furthermore, the positional difference of the same cluster before and after translation of the cluster center may be small, leading to poorer translation encryption performance. Additionally, even within the same industry, different users... Transactions involving different financial values occur at different times, resulting in a massive amount of transaction information. If the cluster value is small, the information points within a single cluster tend to be sparsely distributed. After translating each cluster, more space is needed to store these sparsely distributed information points, leading to unnecessary space waste. Conversely, if the number of clusters is large, the number of information points within a single cluster is small, and the range of coordinate changes for these information points is also smaller. After translating each cluster, the distribution of information points is relatively concentrated, which is more conducive to information storage. Therefore, for the encryption process of transaction information, a smaller cluster value is better; for the storage of transaction information, a larger cluster value is better. This invention requires first obtaining the optimal cluster value:
[0043] The initial cluster value K = 2 is set, and the K-means clustering algorithm is used to cluster all information points in the three-dimensional coordinate system according to the set cluster value to obtain K clusters;
[0044] In this context, the mean of the x-coordinate, the mean of the y-coordinate, and the mean of the vertical coordinate of all information points in the i-th cluster are taken as the x-coordinate, y-coordinate, and vertical coordinate of the cluster center of the i-th cluster, respectively. The cluster center of the i-th cluster is denoted as C. i If i∈[1,K], then calculate the maximum Euclidean distance from the information points contained in the i-th cluster to the cluster center. Set the sphere radius increment ΔR = 5, and take the cluster center of the i-th cluster as the sphere center. For each increment of ΔR in the sphere radius, the cluster is divided once. Each division results in a spherical region. That is, the first division of the i-th cluster results in a region centered at C. i The second division of a spherical region, centered at C, with radius ΔR, yields a region with radius C. i Let $\mathbf{i}$ be the center of a sphere and $\mathbf{2\D}$ be the radius of the sphere. This process continues until the radius of the sphere is greater than or equal to the maximum Euclidean distance, at which point the partitioning ends. This yields multiple spherical regions that can be partitioned within a cluster. The number of spherical regions is the integer part of the result of dividing the maximum Euclidean distance by $\mathbf{2\D}$. Let $n$ be the number of spherical regions that can be partitioned into the $i$-th cluster. i .
[0045] Sort all spherical regions by their radii from smallest to largest. Analyze the spherical shell regions formed by each sphere and its adjacent spherical regions. The region obtained by subtracting the (i-1)th spherical region from the i-th spherical region is the i-th spherical shell region, and we denote the first spherical shell region as the first spherical region. The ratio of the number of information points contained in the i-th spherical shell region to its volume is denoted as the information point density of that spherical shell region. The number of spherical shell regions is equal to the number of spherical regions.
[0046] Since a smaller number of clusters generally results in better encryption, but a smaller cluster value can lead to unnecessary space waste, this invention aims to improve storage space utilization while maintaining optimal encryption performance. Therefore, this invention uses the variance between the average information point density of each cluster to characterize the clustering effect. A larger variance indicates a sparse distribution of information points within a single cluster, resulting in poor clustering and unnecessary waste of storage space. Conversely, a smaller variance indicates a denser distribution of information points within each cluster, reducing the probability of sparse distribution and improving clustering performance. Since fewer clusters are generally better for encryption, meaning the optimal cluster value is negatively correlated with its size, this invention uses e... -xThis invention establishes a negative correlation between the degree of preference and the size of the cluster value. Regarding storage space utilization, higher utilization is better, and correspondingly, lower variance is better. Therefore, this invention establishes a negative correlation between the degree of preference and variance using the reciprocal of the obtained variance. That is, when the cluster value of the K-means clustering algorithm is K, the degree of preference Q of that cluster value is... K It can be represented as:
[0047]
[0048] Where K represents the currently set cluster value; S K ρ is the variance between the information point densities of each cluster when the cluster value is K; j Let n be the information point density of the j-th spherical shell region in the i-th cluster when the cluster value is K; e is the natural constant; n is the information point density of the j-th spherical shell region in the i-th cluster. i Let be the number of spherical shell regions in the i-th cluster; The average information point density of all spherical regions in the i-th cluster is used to represent the information point density of the i-th cluster; VAR(.) is the variance function used to calculate the variance between the information point densities of all clusters when the cluster value is K.
[0049] The values of each cluster are continuously adjusted, and the cluster value corresponding to the maximum value among the optimization degrees of different cluster values is taken as the optimal cluster value.
[0050] Step S003: Obtain the cluster center of each cluster corresponding to the optimal cluster value; based on the cluster center of each cluster and the set reference vector and transformation coefficient, obtain the vector representation of each information point in each cluster, and then obtain the first ciphertext data and the first key of the transaction information.
[0051] Obtain the clusters corresponding to the optimal cluster value. For the t-th information point in the u-th cluster, the vector formed by taking the cluster center of the u-th cluster as the starting point and the information point as the ending point is denoted as f. ut Choose any vector from the given vectors as the reference vector f. u0 Set the conversion coefficient W for the u-th cluster. u When the conversion coefficient W u The larger the value, the smaller the representation range of each pixel after conversion, and the more concentrated the distribution of information points, thereby further improving the utilization of storage space. The vector representation of the t-th information point in the u-th cluster is denoted as F. ut ,but:
[0052]
[0053] The vector representation of each information point is denoted as the first ciphertext data of the transaction information; the base vector of the cluster to which each information point belongs, the transformation coefficient, and the cluster center of that cluster constitute the first key of the transaction information.
[0054] In the above steps, vector operations are first used to change the position of the original information points. Since vectors have both direction and magnitude, there are countless ways to combine a vector difference. That is, the current encryption key space is infinitely large, so the encryption effect of hiding the position of the original spatial points is better, that is, the encryption effect is better. Then, through the transformation coefficient, the information points after the position transformation are made more dense in space, which is more conducive to the storage of ciphertext data.
[0055] Step S004: Obtain the label value and translation vector of each cluster; obtain the vector coordinates of each information point based on the vector representation of each information point in each cluster and the translation vector of its respective cluster, and then obtain the final stored ciphertext data and its key.
[0056] First, arbitrarily select one cluster center from all cluster centers as the baseline cluster center C0. Then, translate the cluster centers of all remaining clusters to the position of the baseline cluster center C0. During this process, each information point in each cluster moves simultaneously with the cluster center. Therefore, for the u-th cluster, the translation vector Y is... u Let be the vector formed by taking the u-th cluster center as the starting point and the reference cluster center as the ending point.
[0057] To distinguish between different clusters, the clusters are sorted according to the coordinates of their centers. First, the clusters are sorted from smallest to largest according to the horizontal coordinates of their centers. Then, they are sorted from smallest to largest according to the vertical coordinates of their centers. Finally, they are sorted from smallest to largest according to the vertical coordinates of their centers. The index of each cluster is then the label value of that cluster.
[0058] Add the vector representation of each information point in each cluster before translation to the translation vector of the cluster to which the information point belongs to obtain the vector coordinates of the information point after translation. Then, for the t-th information point in the u-th cluster, the vector coordinates s of the point after translation are... ut It can be represented as:
[0059] s ut =F ut +Y u
[0060] The vector coordinates of each information point after translation and the label value of the cluster to which the information point belongs are used as the final ciphertext data of the information point. Therefore, the second ciphertext data of the t-th information point in the u-th cluster is (s utThe second ciphertext data of all information points constitutes the second ciphertext data of all transaction information of all users in the same industry. The benchmark cluster center and the translation vector of each cluster constitute the second key of all transaction information.
[0061] Thus, the encrypted data finally stored on the data platform in this invention is obtained, namely the second encrypted data, and the key corresponding to the encrypted data is the second key.
[0062] By using existing compression techniques, such as Huffman coding, to compress the second ciphertext data, and then transmitting the compressed data to a unified data platform for data storage, this invention completes the secure storage of transaction information of users in various industries.
[0063] By placing different clusters into the same relative space through the above steps, the relative positional relationship between different clusters is avoided, which would increase the confusion of the ciphertext and help hide the original information. At the same time, the relative positions of all spatial points are consistent, which means that the space required to store all points is reduced, thereby reducing the storage space of the ciphertext and facilitating the transmission and storage of the ciphertext.
[0064] When extracting transaction information data from a data platform, it is first necessary to obtain the relevant extraction permissions, i.e., obtain the key for the encrypted data, and then decrypt the stored transaction information. The specific decryption process is as follows:
[0065] Obtain the second encrypted data of the transaction information and the second key of the transaction information. Obtaining the second encrypted data of the transaction information means obtaining the vector coordinates and cluster label values of the information points corresponding to each transaction information in the transaction information. Obtaining the second key of the transaction information means obtaining the reference cluster center and translation vector of each cluster.
[0066] First, all points are divided according to the cluster label value to obtain different clusters. Then, each cluster is translated based on the baseline cluster center. That is, the vector coordinates of the information points in each cluster are subtracted from the translation vector of the cluster to preliminarily restore the spatial position of each information point and obtain the first ciphertext data of the transaction information.
[0067] Then, the first key of the transaction information is obtained, namely, the base vector, transformation coefficient, and cluster center of each cluster in the transaction information. Since the position transformation results of each information point in each cluster are obtained based on the base vector and transformation coefficient, the original position of each information point can be obtained by reverse deduction, thus obtaining the original transaction information.
[0068] This completes the decryption process of the transaction information stored in the data platform.
[0069] By following the steps above, the secure storage of transaction information from users in different industries can be achieved.
[0070] This invention first acquires the information points of each user's transaction information; then, it clusters the coordinates of all information points using different cluster values; it calculates the optimization degree of each cluster value to obtain the optimal cluster value; it acquires each cluster corresponding to the optimal cluster value and its cluster center; based on the cluster center of each cluster and the set reference vector and transformation coefficient, it obtains the first encrypted data and the first key of the transaction information; it acquires the label value of each cluster; it obtains the translation vector of each cluster based on the reference cluster center; it obtains the vector coordinates of each information point based on its vector representation and the label value of its cluster, thereby obtaining the second encrypted data and the second key of the transaction information, and then stores the second encrypted data, thus achieving secure storage of user transaction information. This invention can obtain an infinitely large key, improving data storage security while increasing storage space utilization.
[0071] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for securely storing transaction information based on user industry tags, characterized in that, The method includes the following steps: Information points for obtaining transaction information of each user in the same industry; Different cluster values are set, and all information point coordinates are clustered according to each cluster value. Each cluster value yields multiple clusters and the cluster center of each cluster. Multiple spherical shell regions of each cluster are obtained based on the distribution range of information points in each cluster corresponding to each cluster value. The mean of the information point density among the spherical shell regions in each cluster is calculated to obtain the information point density of each cluster. The optimization degree of each cluster value is obtained based on the difference in information point density among the clusters corresponding to each cluster value. The cluster value corresponding to the maximum optimization degree is taken as the optimal cluster value. Based on the cluster center of each cluster corresponding to the optimal cluster value, as well as the reference vector and transformation coefficient set for each cluster, the vector representation of each information point in each cluster is obtained; the vector representation of each information point in each cluster is used as the first ciphertext data of the transaction information, and the cluster center, reference vector and transformation coefficient of each cluster are used as the first key of the transaction information. The label value of each cluster is obtained based on the coordinates of the cluster center of each cluster; the translation vector of each cluster is obtained with the cluster center of each cluster as the starting point and the reference cluster center of each cluster as the ending point; the vector representation of each information point is added to the translation vector of the cluster to which each information point belongs to obtain the vector coordinates of each information point; the vector coordinates of each information point and the label value of the cluster to which each information point belongs are used as the second ciphertext data of the transaction information, and the reference cluster center and the translation vector of each cluster are used as the second key of the transaction information; the second ciphertext data of the transaction information is stored. The method for obtaining multiple spherical shell regions of each cluster based on the distribution range of information points in each cluster corresponding to each cluster value is as follows: Set the sphere radius increment; obtain the maximum Euclidean distance from all information points in each cluster to the cluster center of each cluster, the maximum Euclidean distance being the distribution range of information points in each cluster; with the cluster center of each cluster as the center of the sphere, each time the sphere radius increases by one sphere radius increment, each cluster is divided once, each division yields a spherical region, thus obtaining multiple spherical regions, the number of spherical regions being the result obtained by rounding up the ratio of the maximum Euclidean distance to the sphere radius increment; Multiple spherical regions are sorted in ascending order of sphere radius. The region obtained by subtracting each spherical region from its adjacent spherical regions is recorded as a spherical shell region, resulting in multiple spherical shell regions for each cluster. The adjacent spherical region refers to the spherical region with the largest radius among all spherical regions whose sphere radius is smaller than that of each spherical region. The expression for the preference level of each cluster value is: Among them, Q K S represents the degree of preference when the cluster value is K. K Let K be the variance between the density of information points of each cluster when the cluster value is K, and e is the natural constant.
2. The method for securely storing transaction information based on user industry tags according to claim 1, characterized in that, The method for obtaining transaction information points for each user in the same industry is as follows: First, user industry tags are obtained through user registration information, and user transaction information for each transaction is obtained through the transaction portal of financial institutions. The transaction information includes the transaction time and financial value of each user's transaction for each transaction. Users are categorized into different industries based on their industry tags, and then a three-dimensional coordinate system is established with the user's transaction time as the horizontal axis, the user's financial value at the time of transaction as the vertical axis, and different users in the same industry as the vertical axis. By projecting each user's transaction information into a three-dimensional coordinate system, we can obtain the information points corresponding to each user's transaction information in the same industry.
3. The method for securely storing transaction information based on user industry tags according to claim 1, characterized in that, The method for calculating the information point density of each spherical region in each cluster is as follows: Obtain the number of information points contained in each spherical shell region, and the volume of each spherical shell region; use the ratio between the number of information points contained in each spherical shell region and the volume of each spherical shell region as the information point density of each spherical shell region.
4. The method for securely storing transaction information based on user industry tags according to claim 1, characterized in that, The method for obtaining the vector representation of each information point in each cluster is as follows: The vector formed by taking the cluster center of each cluster as the starting point and each information point in each cluster as the ending point is used as the direction vector of each information point. The ratio of the difference between the reference vector and the direction vector of each information point to the conversion coefficient is used as the vector representation of each information point.
5. The method for securely storing transaction information based on user industry tags according to claim 1, characterized in that, The method for obtaining the label value of each cluster based on the coordinates of the cluster center of each cluster is as follows: First, sort the cluster centers of each cluster according to their horizontal coordinates from smallest to largest, then sort them according to their vertical coordinates from smallest to largest, and finally sort them according to their vertical axis from smallest to largest. Record the index of each cluster as the label value of each cluster.