A method for securely uploading financial data for corporate finance

Abstract

The present invention relates to the field of data encryption technology, and more specifically to a method for securely uploading financial data for corporate finance. The method uses the influence of newly added backup financial data on cluster offset and the historical offset of the center point during the iterative clustering of financial data to obtain the center influence index of the newly added backup financial data in the allocation step and the center update index in the update step. The center offset scale is obtained through the iterative center influence index and the center update index to perform the next iterative clustering, until a sensitive clustering cluster is obtained and matched with an encryption algorithm. The encrypted data is then encrypted using a mixed encryption method using the sensitive clustering cluster and the matching encryption algorithm to obtain encrypted data. When the present invention divides financial data based on risk conditions, a supervisory mechanism is added to adjust the offset of the iterative clustering center, thereby improving the selectivity of the clustering results, making the category matching reliability of the multi-encryption algorithm higher and the subsequent data encryption security better.

Description

Technical Field

[0001] The present invention relates to the technical field of data encryption, and in particular to a method for securely uploading financial data for corporate finance. Background Art

[0002] With the development of the information society, various information resources are having an increasingly greater impact on all walks of life. As industries develop informatization, they inevitably generate large amounts of industry data. For example, the financial industry, a common example, typically utilizes enterprise-related financial data for centralized upload and storage, enabling data mining and the development of strategies and paths that better suit the company's development. However, due to the sensitivity of financial data, when uploading financial data, attention must be paid to data security to prevent data theft and the resulting loss of corporate assets. This ensures efficient upload of financial data while also maintaining privacy.

[0003] Existing enterprise financial data uploads typically encrypt the data using a data encryption algorithm, then communicate via protocol authentication to ensure data upload. However, the volume of financial data is typically large, requiring comprehensive multi-layer encryption to ensure secure upload. This multi-layer encryption method is computationally complex and inefficient, and the unified encryption strategy requires strict key management. Failure to properly manage keys can easily lead to data leaks and poor encryption effectiveness. Summary of the Invention

[0004] In order to solve the technical problem of poor encryption effect in the prior art, the purpose of the present invention is to provide a method for securely uploading financial data for corporate finance. The technical solution adopted is as follows:

[0005] The present invention provides a method for securely uploading financial data for corporate finance, the method comprising:

[0006] Obtain the attack risk value of each type of backup financial data in the enterprise financial database and the performance parameters of different encryption algorithms;

[0007] K-means iterative clustering is performed based on the number of encryption algorithms and the attack risk value of the backup financial data. During the current iteration, the newly added backup financial data in each cluster is used as the allocated financial data for each cluster. The central influence index of each allocated financial data is obtained based on the degree of deviation of the attack risk value of the allocated financial data in the cluster. The central update index of each allocated financial data is obtained based on the continuous deviation of the center point of each cluster in the previous iteration and the importance of each allocated financial data in each cluster.

[0008] Based on the offset of the center point of each cluster in the current iteration in the previous iteration, the center influence index and center update index of all distributed financial data in each cluster are adjusted to obtain the center offset scale of each cluster in the current iteration; based on the center offset scale of each cluster in the current iteration, the cluster of the next iteration is obtained; until the iteration stops, the sensitive cluster is obtained;

[0009] According to the distribution of attack risk values in the sensitive clusters and the ranking and matching of performance parameters of different encryption algorithms, the matching encryption algorithm of each sensitive cluster is obtained; according to the matching between the data to be encrypted and the sensitive clusters, the corresponding matching encryption algorithm is used for encryption to obtain the encrypted data.

[0010] Furthermore, the method for obtaining the center influence index includes:

[0011] For any cluster in the current iteration, calculate the difference between the attack risk value of each backup financial data in the cluster and the central point value as the risk deviation of each backup financial data; and take the sum of all risk deviations in the cluster as the distribution deviation index of the cluster;

[0012] The ratio of the risk deviation to the distribution deviation index of each type of distributed financial data in the cluster is normalized to obtain the central influence index of each type of distributed financial data.

[0013] Furthermore, the method for obtaining the center update indicator includes:

[0014] For any cluster, the update influence of each type of allocated financial data in the cluster is obtained according to the deviation between the influence of each type of allocated financial data on the update of the center point and the historical continuous deviation.

[0015] The occurrence frequency of each type of allocated financial data in the cluster in the database is used as the important influence of each type of allocated financial data;

[0016] The central update index of each type of allocated financial data is obtained by combining the important influence and update influence of each type of allocated financial data in the cluster.

[0017] Furthermore, the method for obtaining the update impact includes:

[0018] Arrange the center point values of the cluster in time series to obtain the center offset sequence of the cluster; calculate the mean of the difference between each two adjacent center point values in the center offset sequence of the cluster to obtain the total offset update degree of the cluster;

[0019] The ratio of the attack risk value of each type of allocated financial data in the cluster to the total number of types of backup financial data in the cluster is used as the offset impact of each type of allocated financial data in the cluster;

[0020] The ratio of the offset influence degree of each type of allocated financial data in the cluster to the total offset update degree is used as the update influence degree of each type of allocated financial data.

[0021] Furthermore, the method for obtaining the center offset scale includes:

[0022] The difference in the center point value of each cluster between the current iteration and the previous iteration is used as the center migration degree of each cluster; in the current iteration, the control coefficient of each cluster is obtained according to the size of the center migration degree of each cluster;

[0023] Combining the central influence index and central update index of all distributed financial data in each cluster, we can obtain the adjustment degree of each cluster.

[0024] The product of the regulation degree and the regulation coefficient of each cluster is used as the migration regulation weight of each cluster; the product of the center migration degree of each cluster and the migration regulation weight is used as the center offset scale of each cluster.

[0025] Furthermore, obtaining the control coefficient of each cluster according to the size of the center migration degree of each cluster includes:

[0026] For any cluster, when the center migration degree of the cluster is greater than the preset amplitude threshold, the center migration degree is negatively correlated and normalized, and the value is used as the control coefficient of the cluster;

[0027] When the center migration degree of the cluster is less than or equal to the preset amplitude threshold, the control coefficient of the cluster is set to 1.

[0028] Furthermore, the step of combining the central influence index and the central update index of all distributed financial data in each cluster to obtain the adjustment degree of each cluster includes:

[0029] For any cluster, the product of the central influence index and the central update index of each type of distributed financial data in the cluster is negatively correlated to obtain the influence coefficient of each type of distributed financial data;

[0030] The mean of the influence coefficients of all the distributed financial data in the cluster is used as the adjustment degree of the cluster.

[0031] Furthermore, obtaining the clusters of the next iteration based on the center offset scale of each cluster in the current iteration includes:

[0032] For any cluster, the difference between the center point value of the cluster in the current iteration and the previous iteration is used as the iterative update value of the cluster;

[0033] When the iterative update value is positive, the sum of the center point value and the center offset scale of the cluster in the previous iteration is used as the updated center point of the cluster in the current iteration to perform the next iterative clustering and obtain the cluster corresponding to the next iteration;

[0034] When the iterative update value is negative, the difference between the center point value of each cluster in the previous iteration and the center offset scale is used as the updated center point of the cluster in the current iteration to perform clustering in the next iteration and obtain the cluster corresponding to the next iteration.

[0035] Furthermore, the method for obtaining the matching encryption algorithm includes:

[0036] Calculate the mean attack risk value of all backup financial data in each sensitive cluster as the sensitivity of each sensitive cluster; arrange the sensitive clusters in descending order of sensitivity to obtain a sensitive cluster sequence;

[0037] Arrange all encryption algorithms in descending order of performance parameters to obtain an algorithm sequence;

[0038] For any sensitive cluster, the encryption algorithm with the same sequence number as the sensitive cluster in the algorithm sequence is used as the matching encryption algorithm of the sensitive cluster.

[0039] Furthermore, the encryption is performed using a corresponding matching encryption algorithm based on the matching between the data to be encrypted and the sensitive clusters to obtain the encrypted data, including:

[0040] According to the sensitive clusters to which each type of financial data in the data to be encrypted belongs, a matching encryption algorithm for each type of financial data is obtained, and different financial data in the data to be encrypted are encrypted using corresponding matching algorithms to obtain ciphertext and key as encrypted data.

[0041] The present invention has the following beneficial effects:

[0042] The present invention considers reducing local optimality during the iterative clustering of financial data and monitors the newly added backup financial data in each iteration. Based on the impact of the newly added backup financial data on cluster offset and the historical offset of the center point, a center influence index for the newly added backup financial data in the allocation step is obtained, reflecting the degree of influence on the allocation clusters. Furthermore, a center update index for the newly added backup financial data in the update step is obtained, reflecting the impact on the cluster update offset. The center offset scale obtained from the impact of the two steps in the iteration is used to monitor and adjust the next iteration clustering, avoiding the possibility of local optimality and achieving more reasonable classification of the backup financial data. This allows for a more optimal matching encryption algorithm based on risk sensitivity when matching with an encryption algorithm. When encrypting encrypted data using multiple encryption algorithms, more secure encrypted data is obtained. When categorizing historical financial data based on risk, the present invention incorporates a supervisory mechanism to adjust the offset of the iterative cluster centers, improving the selectivity of the clustering results, enhancing the reliability of the category matching of the multiple encryption algorithms, and improving the security of subsequent data encryption. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 A flowchart of a method for securely uploading financial data for corporate finance provided by one embodiment of the present invention;

[0045] Figure 2 A flow chart of a method for obtaining a center update indicator provided by one embodiment of the present invention;

[0046] Figure 3 A flow chart of a method for obtaining a center offset scale provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0047] To further illustrate the technical means and effectiveness of the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of a method for securely uploading financial data for corporate finance proposed in accordance with the present invention. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0048] Unless defined otherwise, 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 belongs.

[0049] The following describes in detail a specific solution of a method for securely uploading financial data for corporate finance provided by the present invention in conjunction with the accompanying drawings.

[0050] See also Figure 1 , which shows a flow chart of a method for securely uploading financial data for corporate finance provided by one embodiment of the present invention, the method comprising the following steps:

[0051] S1: Obtain the attack risk value of each backup financial data in the enterprise financial database and the performance parameters of different encryption algorithms.

[0052] Because there are different types of corporate financial data, and different types of financial data have different sensitivities and corresponding security levels, in order to ensure data transmission efficiency, different encryption algorithms can be used to process financial data through a mixed strategy encryption method, so that there are multiple keys for the data to be uploaded, and it is impossible to decrypt the financial data to be uploaded with a single key. Even if some keys are leaked, all data cannot be leaked, thus ensuring the security of the uploaded financial data.

[0053] In order to more reliably match encryption algorithms for different types of financial data, this embodiment classifies different types of financial data according to risk conditions, and adapts better strategies based on different types of risks and the security performance of encryption algorithms. Therefore, it is necessary to obtain the risk conditions of each type of backup financial data in the enterprise financial database, as well as the security performance of different encryption algorithms.

[0054] In this embodiment of the present invention, to facilitate cryptographic analysis of diverse financial data, historical data from a corporate financial database over a period of time is accessed and backed up to obtain backup financial data for subsequent analysis. The backup financial data in the corporate financial database is then retrieved, and all attacked data is collected to form an attacked database. The frequency of occurrence of each type of backup financial data in the attacked database is recorded as the attack risk value for each backup financial data type. Backup financial data that does not appear in the attacked database is assigned an attack risk value of 0. The collection and acquisition of corporate data in this embodiment of the present invention is authorized and does not violate relevant laws and regulations or public order and morality.

[0055] When analyzing multiple encryption algorithms, we first select multiple encryption algorithms and evaluate the encryption effect of different encryption algorithms on financial data. In an embodiment of the present invention, N encryption algorithms are selected, and any financial data of 1GB in size is selected as test data. The test data is encrypted using N encryption algorithms to obtain the ciphertext of each encryption algorithm. The security parameters of each encryption algorithm are obtained by attacking the ciphertext, which are used as the performance parameters of each encryption algorithm. The larger the performance parameter, the better the anti-attack ability of the encryption algorithm. It should be noted that the number of types of encryption algorithms can be adjusted by the implementer according to the specific implementation scenario, and there is no restriction here. At the same time, the attack test of the encryption algorithm is a technical means well known to those skilled in the art, such as ciphertext-only attack or plaintext attack, which will not be elaborated here.

[0056] At this point, the data preparation is completed.

[0057] S2: Perform K-means iterative clustering based on the number of encryption algorithms and the attack risk value of the backup financial data. In the current iteration, the newly added backup financial data in each cluster is used as the allocated financial data of each cluster. According to the deviation degree of the attack risk value of the allocated financial data in the cluster, the central influence index of each allocated financial data is obtained. According to the continuous offset of the center point of each cluster in the current iteration in the previous iteration and the importance of each allocated financial data in each cluster, the central update index of each allocated financial data is obtained.

[0058] When classifying and dividing the backup financial data into sensitive risk categories, since there are N types of encryption algorithms, the backup financial data also needs to be divided into N categories based on the attack risk value, and different encryption algorithms need to be adapted to each category, so that subsequent financial data can correspond to different encryption algorithms by category.

[0059] This embodiment of the present invention uses the K-means clustering algorithm to cluster backup financial data. During the clustering process, N initial centers are selected based on the attack risk value. The K-means clustering algorithm is essentially an iterative algorithm that iteratively distributes data and updates centers to optimize clusters. Inaccurate selection of initial centers can result in inaccurate clusters or convergence to a local optimum. This can result in some clusters being too sparse or overlapping, affecting the final clustering quality. Therefore, data distribution and center point updates are monitored during each iteration to adjust the updates.

[0060] First, in the data allocation step of the cluster in the current iteration process, the impact of each newly added backup financial data on the change of the cluster center point is analyzed. The newly added backup financial data in each cluster is first used as the allocated financial data of each cluster. The overall deviation of each financial data in the cluster is used to reflect the possible impact on the subsequent update of the cluster center point selection.

[0061] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the center influence index includes:

[0062] First, for any cluster in the current iteration process, the difference between the attack risk value of each backup financial data in the cluster and the center point value is calculated as the risk deviation of each backup financial data. The center point value is the mean of all attack risk values in the cluster. The deviation reflects the discreteness of the value distribution in the cluster.

[0063] Furthermore, the sum of all risk deviations in the cluster is used as the distribution deviation index of the cluster. The deviation reflects the concentration of the attack risk value in the cluster. The smaller the distribution deviation index, the better the concentration of the cluster.

[0064] Finally, the ratio of the risk deviation to the distribution deviation index of each type of distributed financial data in the cluster is normalized to obtain the central influence index of each type of distributed financial data. For each type of distributed financial data, the greater the risk deviation, the more dispersed the distributed financial data is in the cluster. Therefore, the greater the central influence index, the greater the impact of the distributed financial data on the update of the cluster. As an example, the expression of the central influence index is:

[0065] Where M i,n It is represented as the central influence index of the i-th distribution financial data in the n-th cluster, A i Denotes the attack risk value of the i-th distributed financial data; B n Represented as the center point value of the nth cluster; U n It is represented by the total number of types of backup financial data in the nth cluster, A u It is represented as the attack risk value of the u-th backup financial data in the n-th cluster; |A u -B n | represents the risk deviation of the u-th backup financial data in the n-th cluster; |A i -B n | represents the risk deviation of the i-th distribution of financial data in the n-th cluster; It is represented as the distribution deviation index of the nth cluster; Norm() is represented as the normalization function. It should be noted that normalization is a technical means well known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0066] Next, during the current iteration, the cluster center update step analyzes the impact of each newly assigned financial data point on the center shift, reflecting the importance of the assigned financial data point during the center point update. Therefore, by combining the iterative center point change shift with the importance of the assigned financial data point itself, the potential shift impact of each assigned financial data point is quantified.

[0067] Preferably, in the embodiment of the present invention, the method for obtaining the center update indicator can be found in Figure 2 , which shows a flow chart of a method for obtaining a center update indicator provided by an embodiment of the present invention, the method comprising the following steps:

[0068] S211: For any cluster, the update influence of each type of allocated financial data in the cluster is obtained according to the deviation between the influence of each type of allocated financial data on the update of the center point and the historical continuous offset.

[0069] First, when the center point is updated, the update impact brought by each type of allocated financial data is reflected based on the degree to which each type of allocated financial data participates in the center point offset after allocation, combined with the different changes in the center point update in historical iterations.

[0070] Preferably, in an embodiment of the present invention, the method for obtaining the update influence includes:

[0071] Arrange the cluster's center point values in the time series to obtain the cluster's center offset sequence. Calculate the difference between each two adjacent center point values in the cluster's center offset sequence and average them to obtain the cluster's total offset update degree, reflecting the average size of the center point offset changes over historical iterations. It should be noted that even for the first iteration, the center offset sequence contains the initial center point and the current iteration's center point values, so there are always cases where the difference cannot be calculated.

[0072] The ratio of the attack risk value of each type of allocated financial data in the cluster to the total number of types of backup financial data in the cluster is further used as the offset influence of each type of allocated financial data in the cluster. Since the center point is the mean of all attack risk values, the attack risk value of each type of allocated financial data after being evenly divided among all types of the cluster reflects the proportion of influence of a single type on the update of the center point.

[0073] Finally, the ratio of the offset influence of each type of allocated financial data in the cluster to the total offset update is used as the update influence of each type of allocated financial data. By comparing the offset influence of a single type with the historical overall update offset, the degree of influence of the change in the center offset of the current iteration is reflected. The closer the offset influence is to the total offset update, that is, the greater the update influence, the greater the degree of center offset caused by the allocated financial data in the current iteration.

[0074] S212: The occurrence frequency of each type of allocated financial data in the cluster in the database is used as the important influence of each type of allocated financial data.

[0075] The frequency of occurrence of each type of distributed financial data in the database can reflect the proportion of this type of financial data in the original database. The greater the proportion of distributed financial data in the database, the more important the data performance is, and it is more frequent in actual application. It is easy to be cracked by regularity when it is not considered properly. The proportion of its influence on cluster changes should also be high, so as to improve cluster impact analysis and avoid falling into the local optimal situation of cluster division through numerical analysis alone.

[0076] S213: Combining the importance influence and update influence of each type of allocated financial data in the cluster, obtaining a central update index for each type of allocated financial data.

[0077] In this embodiment of the present invention, the product of the importance and update influence of each type of distributed financial data is normalized to obtain the central update index of each type of distributed financial data. A higher importance and update influence indicate a higher degree of influence of the corresponding distributed financial data on the update of the central point, and therefore a higher central update index is required. As an example, the central update index is expressed as:

[0078] R n,i =Norm(P i ×T n,i );where R n,i It is represented as the central update index of the i-th distribution financial data in the n-th cluster, P i It represents the important influence of the i-th type of financial data in the n-th cluster, T n,i It represents the update influence of the i-th distributed financial data in the n-th cluster, and Norm() represents the normalization function.

[0079] At this point, the supervision analysis of the two steps in each iteration process is completed, and the impact indicators of each type of allocated financial data are obtained.

[0080] S3: Based on the degree of center point offset of each cluster in the current iteration, the center influence index and center update index of all distributed financial data in each cluster are adjusted to obtain the center offset scale of each cluster in the current iteration; based on the center offset scale of each cluster in the current iteration, the cluster of the next iteration is obtained; until the iteration stops, the sensitive cluster is obtained.

[0081] In the K-means iterative clustering process, after cluster assignment, all data in the cluster are re-acquired to obtain the center point for the next iterative cluster assignment. However, considering that the re-acquired center point is prone to local optimality or extreme cluster dispersion, the scale is adjusted according to the allocation data and update impact of each iteration.

[0082] Preferably, in the embodiment of the present invention, the method for obtaining the center offset scale can be found in Figure 3 , which shows a flow chart of a method for obtaining a center offset scale provided by an embodiment of the present invention, the method comprising the following steps:

[0083] S301: The difference in the center point value of each cluster between the current iteration and the previous iteration is used as the center migration degree of each cluster; in the current iteration, the control coefficient of each cluster is obtained according to the size of the center migration degree of each cluster.

[0084] The difference in the center point values between the current iteration and the previous iteration is used to reflect the offset of the original center point, and then the scale is adjusted according to the offset. By allocating influencing indicators of financial data, the discreteness of some data is avoided, which leads to a small offset of the iterative center, resulting in poor clustering convergence or falling into a local optimal solution.

[0085] Preferably, in an embodiment of the present invention, the method for obtaining the control coefficient includes:

[0086] For any cluster, when the center migration degree of the cluster is greater than the preset amplitude threshold, it means that the offset is high and the control intensity needs to be reduced. The center migration degree is negatively correlated and normalized to the value used as the control coefficient of the cluster. The greater the degree of deviation, the smaller the degree of control required. It should be noted that negative correlation mapping is a technical means well known to technicians in this field, and can be in the form of inverse proportional values or negative exponential powers, etc., and is not limited here.

[0087] When the center migration degree of the cluster is less than or equal to the preset amplitude threshold, it indicates that the offset is small and the control coefficient needs to be larger, and the control coefficient of the cluster is set to 1. In a specific implementation of the embodiment of the present invention, the preset amplitude threshold is set to 0.53. The specific value can be adjusted by the implementer and is not limited here.

[0088] S302: Combining the central influence index and the central update index of all distributed financial data in each cluster to obtain the adjustment degree of each cluster.

[0089] The influence of the allocated financial data is then combined to further restrict the adjustment scale of the clusters. The greater the influence of the allocated financial data on the clusters after allocation, the more changes the clusters have undergone, and the smaller the adjustment force required. Therefore, in an embodiment of the present invention, the method for obtaining the adjustment negative includes:

[0090] For any cluster, the product of the central influence index and the central update index of each type of distributed financial data in the cluster is negatively correlated to obtain the influence coefficient of each type of distributed financial data. The greater the impact, the smaller the adjustment force brought by the impact during adjustment.

[0091] Further combined with the impact analysis results of all the allocated financial data, the mean of the impact coefficients of all the allocated financial data in the cluster is used as the adjustment degree of the cluster.

[0092] S303: The product of the adjustment degree and the control coefficient of each cluster is used as the migration adjustment weight of each cluster; the product of the center migration degree of each cluster and the migration adjustment weight is used as the offset adjustment degree of each cluster; the sum of the center migration degree and the offset adjustment degree of each cluster is calculated to obtain the center offset scale of each cluster.

[0093] By analyzing the degree of cluster center offset and the impact of distributed financial data, the adjustment weight of each cluster offset is comprehensively obtained, and the migration adjustment weight is obtained by combining them in the form of product.

[0094] Furthermore, on the basis of improving the degree of offset adjustment, when the migration adjustment weight is higher, it means that the adjustment strength is greater. The greater the offset adjustment degree is obtained by multiplying the center migration degree and the migration adjustment weight. Then, the center offset scale after supervision adjustment is obtained by the sum of the center migration degree and the offset adjustment degree, avoiding the local optimization situation where the center point migration is smaller.

[0095] At this point, the supervision and adjustment process of each iteration is completed, and the adaptive update strength of the cluster center point of the current iteration is obtained.

[0096] Therefore, preferably, in an embodiment of the present invention, obtaining the clusters of the next iteration based on the center offset scale of each cluster in the current iteration includes:

[0097] For any cluster, the difference between the center point values of the cluster in the current iteration and the previous iteration is used as the iterative update value of the cluster, and the sign of the iterative update value reflects the update direction of the center point after the original iteration.

[0098] When the iterative update value is positive, it means that the update direction of the center point value is positive. The sum of the center point value of the cluster cluster in the previous iteration and the center offset scale is used as the updated center point of the cluster cluster in the current iteration for the next iterative clustering. The cluster cluster corresponding to the next iteration is obtained, and the center point of the cluster cluster is updated by the adjusted offset scale.

[0099] When the iterative update value is negative, it means that the update direction of the center point value is negative. The difference between the center point value of each cluster in the previous iteration and the center offset scale is used as the updated center point of the cluster in the current iteration to perform the next iterative clustering and obtain the cluster corresponding to the next iteration.

[0100] By continuously monitoring and adjusting the scale of the cluster center points, the iterative clustering division is made more accurate until the iteration stops, and the final clustering result is a sensitive cluster. It should be noted that the iterative termination condition of K-means iterative clustering is cluster convergence or reaching the number of iterations. In this embodiment of the present invention, the maximum number of iterations can be set to 12. The implementer can adjust the iterative termination condition according to the specific implementation situation, and this is not limited here.

[0101] S4: According to the distribution of attack risk values in the sensitive clusters and the ranking and matching of performance parameters of different encryption algorithms, the matching encryption algorithm of each sensitive cluster is obtained; according to the matching between the data to be encrypted and the sensitive clusters, the corresponding matching encryption algorithm is used for encryption to obtain encrypted data.

[0102] After obtaining the sensitive clusters, a corresponding encryption algorithm may be selected for each sensitive cluster. In this embodiment, the matching is performed by ranking and matching the degree of attack risk with the performance of the encryption algorithm.

[0103] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the matching encryption algorithm includes:

[0104] First, the mean attack risk value of all backup financial data in each sensitive cluster is calculated as the sensitivity of each sensitive cluster, reflecting the possibility of the sensitive cluster as a whole being attacked. The greater the sensitivity, the higher the risk of attack.

[0105] Arrange the sensitive clusters in descending order of sensitivity to obtain a sensitive cluster sequence, and then arrange the sensitive clusters in descending order of attack risk. Arrange all encryption algorithms in descending order of performance parameters to obtain an algorithm sequence, and then arrange the encryption algorithms in descending order of encryption security.

[0106] Therefore, for any sensitive cluster, the encryption algorithm with the same sequence number as the sensitive cluster in the algorithm sequence is used as the matching encryption algorithm for the sensitive cluster. Two sequences are matched by the same sequence number. The smaller the sequence number, the higher the attack risk of the sensitive cluster, and the higher the encryption security of the matching encryption algorithm. Because the clustering process is based on the number of encryption algorithms as the number of clusters, each sensitive cluster corresponds to an encryption algorithm with the same sequence number.

[0107] By backing up financial data, the correspondence between encryption algorithms for different financial data is completed. As a matching encryption strategy, the corresponding matching encryption algorithm can be used to encrypt the data to be encrypted and the sensitive clusters to obtain encrypted data. In an embodiment of the present invention, the matching encryption algorithm for each financial data in the data to be encrypted is obtained according to the sensitive clusters to which each financial data in the data to be encrypted is located. The corresponding matching algorithm is used to encrypt different financial data in the data to be encrypted to obtain the ciphertext and key as the encrypted data.

[0108] The matching encryption strategy is used to reduce the computational complexity of the overall encryption, improve efficiency while ensuring the security of the encrypted data, and increase the difficulty of data cracking. In an embodiment of the present invention, the ciphertext can be converted into the same format and securely uploaded to the data storage terminal, and can be subsequently decrypted based on the key and the matching encryption strategy.

[0109] In summary, the present invention considers reducing local optimality during the iterative clustering of financial data and monitors the newly added backup financial data in each iteration. Based on the impact of the newly added backup financial data on cluster offset and the historical offset of the center point, a center influence index for the newly added backup financial data in the allocation step is obtained, reflecting the degree of influence on the allocation clusters. Furthermore, a center update index for the newly added backup financial data in the update step is obtained, reflecting the impact of the cluster update offset. The center offset scale obtained from the impact of the two steps in the iteration is used to supervise and adjust the next iteration clustering, avoiding the possibility of local optimality and achieving more reasonable classification of the backup financial data. This allows for a more optimal matching encryption algorithm based on risk sensitivity when matching with encryption algorithms, resulting in more secure encrypted data when encrypted data is encrypted using multiple encryption algorithms. When historical financial data is classified based on risk, the present invention incorporates a supervisory mechanism to adjust the offset of the iterative cluster centers, improving the selectivity of the clustering results, enhancing the reliability of the classification matching of the multiple encryption algorithms, and improving the security of subsequent data encryption.

[0110] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0111] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A method for securely uploading financial data for corporate finance, characterized in that: The method comprises: Obtaining an attack risk value for each type of backup financial data in the enterprise financial database and performance parameters of different encryption algorithms, wherein the performance parameters are used to characterize the encryption algorithm's ability to resist attacks, and the attack risk value is the frequency of occurrence of the backup financial data in an attacked database, where the attacked database is a database consisting of all attacked data; K-means iterative clustering is performed based on the number of encryption algorithms and the attack risk value of the backup financial data. During the current iteration, the newly added backup financial data in each cluster is used as the allocated financial data for each cluster. Based on the deviation of the attack risk value of the allocated financial data in the cluster, a central influence index for each allocated financial data is obtained. The central influence index is used to reflect the degree of influence on the allocated cluster. The central influence index is obtained by normalizing the ratio of the risk deviation of each allocated financial data in the cluster to the distribution deviation index. The risk deviation is the difference between the attack risk value of each backup financial data in the cluster and the center point value. The risk deviation is used to reflect the discreteness of the value distribution in the cluster. The distribution deviation index is the sum of all risk deviations in the cluster. The distribution deviation index is used to reflect the concentration of the attack risk value in the cluster. Based on the continuous deviation of the center point of each cluster in the current iteration in the previous iteration and the importance of each allocated financial data in each cluster, a central update index for each allocated financial data is obtained. The central update index is used to reflect the influence on the cluster update deviation. Based on the offset of the center point of each cluster in the current iteration in the previous iteration, the center influence index and center update index of all distributed financial data in each cluster are adjusted to obtain the center offset scale of each cluster in the current iteration; based on the center offset scale of each cluster in the current iteration, the cluster of the next iteration is obtained; until the iteration stops, the sensitive cluster is obtained; According to the distribution of attack risk values in the sensitive clusters and the ranking and matching of performance parameters of different encryption algorithms, the matching encryption algorithm of each sensitive cluster is obtained; according to the matching between the data to be encrypted and the sensitive clusters, the corresponding matching encryption algorithm is used for encryption to obtain the encrypted data.

2. A method for securely uploading financial data for corporate finance according to claim 1, characterized in that: The method for obtaining the center update indicator includes: For any cluster, the update influence of each type of allocated financial data in the cluster is obtained according to the deviation between the influence of each type of allocated financial data on the update of the center point and the historical continuous deviation. The occurrence frequency of each type of allocated financial data in the cluster in the database is used as the important influence of each type of allocated financial data; The central update index of each type of allocated financial data is obtained by combining the important influence and update influence of each type of allocated financial data in the cluster.

3. A method for securely uploading financial data for corporate finance according to claim 2, characterized in that: The method for obtaining the update impact includes: Arrange the center point values of the cluster in time series to obtain the center offset sequence of the cluster; calculate the mean of the difference between each two adjacent center point values in the center offset sequence of the cluster to obtain the total offset update degree of the cluster; The ratio of the attack risk value of each type of allocated financial data in the cluster to the total number of types of backup financial data in the cluster is used as the offset impact of each type of allocated financial data in the cluster; The ratio of the offset influence degree of each type of allocated financial data in the cluster to the total offset update degree is used as the update influence degree of each type of allocated financial data.

4. A method for securely uploading financial data for corporate finance according to claim 1, characterized in that: The method for obtaining the center offset scale includes: The difference in the center point value of each cluster between the current iteration and the previous iteration is used as the center migration degree of each cluster; in the current iteration, the control coefficient of each cluster is obtained according to the size of the center migration degree of each cluster; Combining the central influence index and central update index of all distributed financial data in each cluster, we can obtain the adjustment degree of each cluster. The product of the regulation degree and the regulation coefficient of each cluster is used as the migration regulation weight of each cluster; the product of the center migration degree of each cluster and the migration regulation weight is used as the center offset scale of each cluster.

5. A method for securely uploading financial data for corporate finance according to claim 4, characterized in that: The control coefficient of each cluster is obtained according to the size of the center migration degree of each cluster, including: For any cluster, when the center migration degree of the cluster is greater than the preset amplitude threshold, the center migration degree is negatively correlated and normalized, and the value is used as the control coefficient of the cluster; When the center migration degree of the cluster is less than or equal to the preset amplitude threshold, the control coefficient of the cluster is set to 1.

6. A method for securely uploading financial data for corporate finance according to claim 4, characterized in that: The method combines the central influence index and the central update index of all distributed financial data in each cluster to obtain the adjustment degree of each cluster, including: For any cluster, the product of the central influence index and the central update index of each type of distributed financial data in the cluster is negatively correlated to obtain the influence coefficient of each type of distributed financial data; The mean of the influence coefficients of all the distributed financial data in the cluster is used as the adjustment degree of the cluster.

7. A method for securely uploading financial data for corporate finance according to claim 1, characterized in that: The step of obtaining the next iteration cluster based on the center offset scale of each cluster in the current iteration includes: For any cluster, the difference between the center point value of the cluster in the current iteration and the previous iteration is used as the iterative update value of the cluster; When the iterative update value is positive, the sum of the center point value and the center offset scale of the cluster in the previous iteration is used as the updated center point of the cluster in the current iteration to perform the next iterative clustering and obtain the cluster corresponding to the next iteration; When the iterative update value is negative, the difference between the center point value of each cluster in the previous iteration and the center offset scale is used as the updated center point of the cluster in the current iteration to perform clustering in the next iteration and obtain the cluster corresponding to the next iteration.

8. A method for securely uploading financial data for corporate finance according to claim 1, characterized in that: The method for obtaining the matching encryption algorithm includes: Calculate the mean attack risk value of all backup financial data in each sensitive cluster as the sensitivity of each sensitive cluster; arrange the sensitive clusters in descending order of sensitivity to obtain a sensitive cluster sequence; Arrange all encryption algorithms in descending order of performance parameters to obtain an algorithm sequence; For any sensitive cluster, the encryption algorithm with the same sequence number as the sensitive cluster in the algorithm sequence is used as the matching encryption algorithm of the sensitive cluster.

9. A method for securely uploading financial data for corporate finance according to claim 1, characterized in that: The method of encrypting the data to be encrypted using a corresponding matching encryption algorithm based on the matching between the data to be encrypted and the sensitive clusters to obtain the encrypted data includes: According to the sensitive clusters to which each type of financial data in the data to be encrypted belongs, a matching encryption algorithm for each type of financial data is obtained, and different financial data in the data to be encrypted are encrypted using corresponding matching algorithms to obtain ciphertext and key as encrypted data.

Images