Enterprise management information security data processing system and method
By collecting and processing enterprise management information data, using clustering and convolutional neural networks for feature extraction and security verification, and combining hierarchical analysis method for security evaluation, the problems of insufficient security of enterprise informationization and low efficiency in processing information are solved, and efficient and accurate secure data processing is achieved.
Patent Information
- Application Number
- CN202411819189.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-11
- Publication Date
- 2025-04-04
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing technology has limitations on the problems of insufficient security of enterprise informationization and low efficiency in processing information, and has failed to effectively solve the data security problems.
By collecting enterprise management information data, building training sets and processing data, using technologies such as clustering and convolutional neural networks to extract and verify the data, and combining hierarchical analysis method for security evaluation.
It improves the accuracy and efficiency of enterprise management information security data processing, ensures data reliability and security, and realizes real-time security evaluation.
Smart Images
Figure CN119760438A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and in particular to a system and method for processing enterprise management information security data. Background Art
[0002] With the continuous development of enterprise informatization, the growth rate of enterprise data is getting faster and faster, and the types of data are becoming more and more numerous, which brings huge challenges to the security performance and information processing of enterprise informatization.
[0003] The existing open patent application CN113568810A, the method collects initial security log data, filters and merges the initial security log data, obtains a first security log set after filtering and merging, and then standardizes the data in the first security log set to obtain a second security log set after standardization, and at the same time, performs correlation analysis on the security log data in the second security log set, and outputs the log data results after correlation analysis; however, since it only filters and standardizes the associated data, it ignores the security issues of the data itself, and has certain limitations. Summary of the invention
[0004] 1. Technical issues to be solved
[0005] In view of the deficiencies in the prior art, the present invention provides an enterprise management information security data processing system and method, which has the advantages of security, accuracy, efficiency, and real-time, and solves the problems of insufficient enterprise information security and low information processing efficiency.
[0006] (II) Technical solution
[0007] In order to solve the above-mentioned technical problems of insufficient enterprise information security and low information processing efficiency, the present invention provides the following technical solutions:
[0008] The present invention discloses a method for processing enterprise management information security data, which specifically comprises the following steps:
[0009] S1. Collect enterprise management information data to build a training set, and process the enterprise management information data in the training set by a data processing method to obtain processed enterprise management information data;
[0010] S2. Clustering the processed enterprise management information data by clustering method, and outputting clustering results, the clustering results including: data category, data center and clustered enterprise management information data;
[0011] S3, inputting the clustered enterprise management information data into a convolutional neural network, and extracting a feature vector of the clustered enterprise management information data through the convolutional neural network;
[0012] S4, collect enterprise management information data in real time and extract features through convolutional neural network, and verify the security of the enterprise management information data collected in real time by feature comparison;
[0013] S5. For each category of real-time enterprise management information data after security verification, the enterprise management information data of different categories in the same period are compared through the hierarchical analysis method.
[0014] The present invention constructs a training set by collecting enterprise management information data, processes the enterprise management information data in the training set by a data processing method, clusters the processed enterprise management information data by a clustering method, and extracts features from the clustered enterprise management information data by a convolutional neural network. At the same time, the enterprise management information data is collected in real time and features are extracted by a convolutional neural network. At the same time, the security of the real-time collected enterprise management information data is verified by feature comparison. Then, for each category of real-time enterprise management information data after security verification, different categories of enterprise management information data in the same period are compared by a hierarchical analysis method, thereby improving the accuracy of enterprise management information security data processing.
[0015] Preferably, the collecting of enterprise management information data to construct a training set, and processing the enterprise management information data in the training set by a data processing method to obtain the processed enterprise management information data comprises the following steps:
[0016] S11, filtering the collected enterprise management information data;
[0017] Traverse the collected enterprise management information data to locate the management information data with missing content in the enterprise management information data;
[0018] Determine and process the missing content in the enterprise management information data;
[0019] When the missing content includes a management information object or the missing rate of management information data is greater than 50%, the management information data is directly deleted;
[0020] S12. Summarize the enterprise management information data after the filtering operation to obtain processed enterprise management information data.
[0021] The present invention ensures the reliability of the enterprise management information data by filtering the collected enterprise management information data, removing the missing data in the enterprise management information data and summarizing the data.
[0022] Preferably, clustering the processed enterprise management information data by clustering method and outputting the clustering results comprises the following steps:
[0023] S21, randomly select a sample point from the training set as the first initial data center, set a sample point to represent a group of processed enterprise management information data, and each data center represents a data category;
[0024] S22, calculating the distance D between each sample in the training set and the current existing data center;
[0025] S23, based on the calculated distance between each sample and the current existing data center, calculate the probability P of each sample being selected as the next data center;
[0026] S24. Repeat steps S22 and S23 until convergence, and output the converged k data centers and clustered enterprise management information data.
[0027] Preferably, the step of calculating the distance D between each sample in the training set and the current existing data center comprises the following steps:
[0028] The distance D formula between each sample and the current data center is as follows:
[0029] D(x i )=|x i -a k |;
[0030] Among them, D(x i ) represents the sample x i To data center a k The distance, x i represents the i-th sample, a k represents the kth data center.
[0031] Preferably, the calculating the probability P of each sample being selected as the next data center based on the calculated distance between each sample and the currently existing data center comprises the following steps:
[0032] The probability P formula for each sample being selected as the next data center is as follows:
[0033]
[0034] Among them, P(x i ) represents the sample x i The probability of being selected as the next data center, X represents the training set.
[0035] The present invention randomly selects a sample point as the first initial data center, calculates the distance between each sample and the current existing data center, and calculates the probability of each sample being selected as the next data center through the calculated distance, and continuously iterates the calculation to achieve clustering of enterprise management information data, thereby improving the efficiency of enterprise management information data processing.
[0036] Preferably, the step of inputting the clustered enterprise management information data into a convolutional neural network and extracting a feature vector of the clustered enterprise management information data through the convolutional neural network comprises the following steps:
[0037] S31, taking the clustered enterprise management information data as input and inputting it into the convolutional neural network;
[0038] S32, extracting features from the clustered enterprise management information data input by a convolutional neural network;
[0039] S33, after receiving the input clustered enterprise management information data, the convolution layer moves on the input clustered enterprise management information data through the set convolution kernel according to the set step size, and performs convolution calculation on the corresponding area of each step with the feature value of the area, so as to realize feature extraction of each clustered enterprise management information data; the convolution calculation formula is as follows:
[0040] F = I × ω + b;
[0041] Among them, I represents the input clustered enterprise management information data, ω represents the weight of the corresponding convolution kernel, b represents the bias value, and F represents the output feature data;
[0042] S34. In the process of continuous convolution of the convolutional neural network, the output of the previous convolutional neural network will be used as the input of the next convolutional neural network, and the data will be processed by the activation function in the process of inputting into the next convolutional neural network;
[0043] S35, extracting features of the clustered enterprise management information data in the process of continuously stacking convolution and pooling;
[0044] S36, passing the extracted and clustered enterprise management information data features to the fully connected layer;
[0045] S37. Expand and combine the data through the fully connected layer to obtain a set of feature vectors and save them.
[0046] Preferably, the processing of the activation function comprises the following steps:
[0047] Set the input value to μ λ (λ=1,2,……,L), each output value μ λThe corresponding input weight is ω λ , the output result after inputting the input value into the convolutional neural network is:
[0048]
[0049] Among them, g is the corresponding activation function, and Y is the output result of the activation function.
[0050] The present invention uses a convolutional neural network to perform feature extraction on enterprise management information data, and completes the extraction of feature vectors of enterprise management information data by means of convolution calculation and activation function processing, thereby ensuring the accuracy of the extraction of feature vectors of enterprise management information data.
[0051] Preferably, the real-time collection of enterprise management information data and feature extraction through a convolutional neural network, and the verification of the security of the real-time collected enterprise management information data by feature comparison, include the following steps:
[0052] Calculate the similarity between the feature vector of the real-time collected enterprise management information data and the feature vector of the stored enterprise management information data, and set a similarity threshold;
[0053] The similarity calculation formula is as follows:
[0054]
[0055] in, Represents the feature vector and the eigenvector The similarity value between Represents the feature vector of real-time collection of enterprise management information data, A feature vector representing the stored enterprise management information data;
[0056] A similarity threshold is set. When the calculated similarity is less than the set threshold, it means that the feature vectors of the real-time collected enterprise management information data and the stored enterprise management information data are not similar. It is determined that there is an abnormality in the real-time collected enterprise management information data currently collected in real time. When the calculated similarity is greater than or equal to the set threshold, it means that the feature vectors of the real-time collected enterprise management information data and the stored enterprise management information data are similar, and it is determined that the real-time collected enterprise management information data currently collected in real time is safe.
[0057] The present invention improves the reliability of security verification of enterprise management information data by calculating the similarity between the feature vector of real-time collected enterprise management information data and the feature vector of stored enterprise management information data, and setting a similarity threshold to complete the comparison between real-time collected enterprise management information data and historically collected enterprise management information data.
[0058] Preferably, for each category of real-time enterprise management information data after security verification, comparing different categories of enterprise management information data in the same period by using the hierarchical analysis method includes the following steps:
[0059] S51, setting a hierarchical structure diagram;
[0060] The hierarchy diagram is divided into three layers;
[0061] The highest level is the target level, which is used to judge the target or ideal result; the middle level is the criterion level, which includes various factors in the intermediate links involved in achieving the target, and each factor is set as a criterion. The criterion value ranges from 1 to 9, and the proportion scale of each criterion is set; the lowest level is the program level, which is used to provide various measures to achieve the target;
[0062] S52, construct a comparison matrix;
[0063] Select the same type of data existing in different categories of enterprise management information data within the same period, and set each type of data as a factor;
[0064] Set the number of factors to n, determine the weight of each factor in the highest level, and construct a judgment matrix;
[0065] For two factors C h and C e , use q he Represents C h and C e The ratio of the degree of influence on the top layer, measured on a scale of 1 to 9. he (h,e=1,2,...,n);
[0066] Determination of the ratio scale: q he Take 9 levels from 1 to 9, the ratio scale q eh Get q he The reciprocal of;
[0067] Summarize the comparison matrix Q of each factor pairwise comparison = (q he ) n×n ;
[0068] S53, relative weight vector is determined;
[0069] The geometric mean of each row and column vector of the comparison matrix Q is normalized, and the normalized vector value is used as the weight;
[0070] The relative weight vector determination formula is as follows:
[0071]
[0072] Among them, w j represents the relative weight vector of the jth factor;
[0073] Based on the calculated weight of each factor, different categories of enterprise management information data in the same period are compared, and the weights are summarized, and the enterprise management information data with the highest weight is set to have the highest security evaluation.
[0074] The present invention compares different categories of enterprise management information data in the same period by using the hierarchical analysis method, and quantifies the security evaluation by constructing a comparison matrix and calculating weights, thereby improving the real-time performance of the security evaluation of the enterprise management information data.
[0075] The present invention also discloses an enterprise management information security data processing system, which is used to implement an enterprise management information security data processing method. The system includes: a data acquisition module, a clustering module, a feature extraction module, a security verification module and a security evaluation module;
[0076] The data acquisition module is used to collect enterprise management information data in real time;
[0077] The clustering module is used to cluster the enterprise management information data collected in real time;
[0078] The feature extraction module is used to extract features from the clustered enterprise management information data;
[0079] The security verification module is used to verify the security of the enterprise management information data collected in real time by means of feature comparison;
[0080] The security evaluation module is used to compare and evaluate enterprise management information data of different categories within the same period through the hierarchical analysis method.
[0081] (III) Beneficial effects
[0082] Compared with the prior art, the present invention provides an enterprise management information security data processing system and method, which has the following beneficial effects:
[0083] 1. The invention constructs a training set by collecting enterprise management information data, processes the enterprise management information data in the training set by a data processing method, clusters the processed enterprise management information data by a clustering method, and extracts features from the clustered enterprise management information data by a convolutional neural network. At the same time, the enterprise management information data is collected in real time and features are extracted by a convolutional neural network. At the same time, the security of the real-time collected enterprise management information data is verified by feature comparison. Then, for each category of real-time enterprise management information data after security verification, the enterprise management information data of different categories in the same period are compared by a hierarchical analysis method, thereby improving the accuracy of enterprise management information security data processing.
[0084] 2. The invention ensures the reliability of the enterprise management information data by filtering the collected enterprise management information data, removing the missing data in the enterprise management information data and summarizing the data.
[0085] 3. The invention randomly selects a sample point as the first initial data center, and calculates the distance between each sample and the current existing data center. At the same time, the probability of each sample being selected as the next data center is calculated through the calculated distance, and the clustering of enterprise management information data is realized by continuous iterative calculation, thereby improving the efficiency of enterprise management information data processing.
[0086] 4. The present invention uses a convolutional neural network to perform feature extraction on enterprise management information data, and completes the extraction of feature vectors of enterprise management information data through convolution calculation and activation function processing, thereby ensuring the efficiency of feature vector extraction of enterprise management information data.
[0087] 5. The present invention improves the reliability of security verification of enterprise management information data by calculating the similarity between the feature vector of real-time collected enterprise management information data and the feature vector of stored enterprise management information data, and setting a similarity threshold to complete the comparison between real-time collected enterprise management information data and historically collected enterprise management information data.
[0088] 6. The invention compares different categories of enterprise management information data in the same period by using the hierarchical analysis method, and quantifies the security evaluation by constructing a comparison matrix and calculating weights, thereby improving the real-time performance of the security evaluation of enterprise management information data. BRIEF DESCRIPTION OF THE DRAWINGS
[0089] Figure 1 It is a structural schematic diagram of the enterprise management information security data processing flow of the present invention. DETAILED DESCRIPTION
[0090] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0091] Example 1
[0092] See also Figure 1 This embodiment discloses a method for processing enterprise management information security data, which specifically includes the following steps:
[0093] S1. Collect enterprise management information data to build a training set, and process the enterprise management information data in the training set by a data processing method to obtain processed enterprise management information data;
[0094] S2. Clustering the processed enterprise management information data by clustering method, and outputting clustering results, the clustering results including: data category, data center and clustered enterprise management information data;
[0095] S3, inputting the clustered enterprise management information data into a convolutional neural network, and extracting a feature vector of the clustered enterprise management information data through the convolutional neural network;
[0096] S4, collect enterprise management information data in real time and extract features through convolutional neural network, and verify the security of the enterprise management information data collected in real time by feature comparison;
[0097] S5. For each category of real-time enterprise management information data after security verification, the enterprise management information data of different categories in the same period are compared by using the hierarchical analysis method;
[0098] For further information, see Figure 1 , collecting enterprise management information data to build a training set, and processing the enterprise management information data in the training set by a data processing method, and obtaining the processed enterprise management information data includes the following steps:
[0099] S11, filtering the collected enterprise management information data;
[0100] Traverse the collected enterprise management information data to locate the management information data with missing content in the enterprise management information data;
[0101] Determine and process the missing content in the enterprise management information data;
[0102] When the missing content includes a management information object or the missing rate of management information data is greater than 50%, the management information data is directly deleted;
[0103] S12, summarizing the enterprise management information data after the filtering operation to obtain processed enterprise management information data;
[0104] For further information, see Figure 1 , clustering the processed enterprise management information data by clustering method, and outputting the clustering results includes the following steps:
[0105] S21, randomly select a sample point from the training set as the first initial data center, set a sample point to represent a group of processed enterprise management information data, and each data center represents a data category;
[0106] S22, calculating the distance D between each sample in the training set and the current existing data center;
[0107] The distance D formula between each sample and the current data center is as follows:
[0108] D(x i )=|x i -a k |;
[0109] Among them, D(x i ) represents the sample x i To data center a k The distance, x i represents the i-th sample, a k represents the kth data center;
[0110] S23, based on the calculated distance between each sample and the current existing data center, calculate the probability P of each sample being selected as the next data center;
[0111] The probability P formula for each sample being selected as the next data center is as follows:
[0112]
[0113] Among them, P(x i ) represents the sample x i The probability of being selected as the next data center, X represents the training set;
[0114] S24, repeating steps S22 and S23 until convergence, and outputting the k data centers after convergence and the clustered enterprise management information data;
[0115] For further information, see Figure 1, inputting the clustered enterprise management information data into the convolutional neural network, and extracting the feature vector of the clustered enterprise management information data through the convolutional neural network includes the following steps:
[0116] S31, taking the clustered enterprise management information data as input and inputting it into the convolutional neural network;
[0117] S32, extracting features from the clustered enterprise management information data input by a convolutional neural network;
[0118] S33, after receiving the input clustered enterprise management information data, the convolution layer moves on the input clustered enterprise management information data through the set convolution kernel according to the set step size, and performs convolution calculation on the corresponding area of each step with the feature value of the area, so as to realize feature extraction of each clustered enterprise management information data; the convolution calculation formula is as follows:
[0119] F = I × ω + b;
[0120] Among them, I represents the input clustered enterprise management information data, ω represents the weight of the corresponding convolution kernel, b represents the bias value, and F represents the output feature data;
[0121] S34. In the process of continuous convolution of the convolutional neural network, the output of the previous convolutional neural network will be used as the input of the next convolutional neural network, and the data will be processed by the activation function in the process of inputting into the next convolutional neural network;
[0122] The activation function processing includes the following steps:
[0123] Set the input value to μ λ (λ=1,2,……,L), each output value μ λ The corresponding input weight is ω λ , the output result after inputting the input value into the convolutional neural network is:
[0124]
[0125] Among them, g is the corresponding activation function, and Y is the output result of the activation function;
[0126] S35, extracting features of the clustered enterprise management information data in the process of continuously stacking convolution and pooling;
[0127] S36, passing the extracted and clustered enterprise management information data features to the fully connected layer;
[0128] S37, expand and combine the data through the fully connected layer to obtain a set of feature vectors and save them;
[0129] For further information, see Figure 1 , real-time collection of enterprise management information data and feature extraction through convolutional neural networks, and at the same time, feature comparison to verify the security of real-time collected enterprise management information data, including the following steps:
[0130] Calculate the similarity between the feature vector of the real-time collected enterprise management information data and the feature vector of the stored enterprise management information data, and set a similarity threshold;
[0131] The similarity calculation formula is as follows:
[0132]
[0133] in, Represents the feature vector and the eigenvector The similarity value between Represents the feature vector of real-time collection of enterprise management information data, A feature vector representing the stored enterprise management information data;
[0134] A similarity threshold is set. When the calculated similarity is less than the set threshold, it indicates that the feature vectors of the real-time collected enterprise management information data and the stored enterprise management information data are not similar. It is determined that the real-time collected enterprise management information data currently collected in real time has an abnormality. When the calculated similarity is greater than or equal to the set threshold, it indicates that the feature vectors of the real-time collected enterprise management information data and the stored enterprise management information data are similar, and it is determined that the real-time collected enterprise management information data currently collected in real time is safe.
[0135] For further information, see Figure 1 , for each category of real-time enterprise management information data after security verification, the enterprise management information data of different categories in the same period are compared by using the hierarchical analysis method, including the following steps:
[0136] S51, setting a hierarchical structure diagram;
[0137] The hierarchy diagram is divided into three layers;
[0138] The highest level is the target level, which is used to judge the target or ideal result; the middle level is the criterion level, which includes various factors in the intermediate links involved in achieving the target, and each factor is set as a criterion. The criterion value ranges from 1 to 9, and the proportion scale of each criterion is set; the lowest level is the program level, which is used to provide various measures to achieve the target;
[0139] S52, construct a comparison matrix;
[0140] Select the same type of data existing in different categories of enterprise management information data within the same period, and set each type of data as a factor;
[0141] Set the number of factors to n, determine the weight of each factor in the highest level, and construct a judgment matrix;
[0142] For two factors C h and C e , use q he Represents C h and C e The ratio of the degree of influence on the top layer, measured on a scale of 1 to 9. he (h,e=1,2,...,n);
[0143] Determination of the ratio scale: q he Take 9 levels from 1 to 9, the ratio scale q eh Get q he The reciprocal of;
[0144] Summarize the comparison matrix Q of each factor pairwise comparison = (q he ) n×n ;
[0145] S53, relative weight vector is determined;
[0146] The geometric mean of each row and column vector of the comparison matrix Q is normalized, and the normalized vector value is used as the weight;
[0147] The relative weight vector determination formula is as follows:
[0148]
[0149] Among them, w j represents the relative weight vector of the jth factor;
[0150] Furthermore, based on the calculated weight of each factor, the enterprise management information data of different categories in the same period are compared, and the weights are summarized, and the enterprise management information data with the highest weight is set to have the highest security evaluation;
[0151] Example 2
[0152] See also Figure 1 ,This embodiment also discloses an enterprise management information security data processing system, which is used to implement an enterprise management information security data processing method. The system includes: a data acquisition module, a clustering module, a feature extraction module, a security verification module, and a security evaluation module;
[0153] The data acquisition module is used to collect enterprise management information data in real time;
[0154] The clustering module is used to cluster the enterprise management information data collected in real time;
[0155] The feature extraction module is used to extract features from the clustered enterprise management information data;
[0156] The security verification module is used to verify the security of the enterprise management information data collected in real time by means of feature comparison;
[0157] The security evaluation module is used to compare and evaluate enterprise management information data of different categories within the same period through the hierarchical analysis method.
[0158] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A method for processing enterprise management information security data, characterized in that: The following steps are involved: S1. Collect enterprise management information data to build a training set, and process the enterprise management information data in the training set by a data processing method to obtain processed enterprise management information data; S2. Clustering the processed enterprise management information data by clustering method, and outputting clustering results, the clustering results including: data category, data center and clustered enterprise management information data; S3, inputting the clustered enterprise management information data into a convolutional neural network, and extracting a feature vector of the clustered enterprise management information data through the convolutional neural network; S4, collect enterprise management information data in real time and extract features through convolutional neural network, and verify the security of the enterprise management information data collected in real time by feature comparison; S5. For each category of real-time enterprise management information data after security verification, the enterprise management information data of different categories in the same period are compared through the hierarchical analysis method.
2. The enterprise management information security data processing method according to claim 1, characterized in that: The step of collecting enterprise management information data to construct a training set, and processing the enterprise management information data in the training set by a data processing method to obtain the processed enterprise management information data comprises the following steps: S11, filtering the collected enterprise management information data; Traverse the collected enterprise management information data to locate the management information data with missing content in the enterprise management information data; Determine and process the missing content in the enterprise management information data; When the missing content includes a management information object or the missing rate of management information data is greater than 50%, the management information data is directly deleted; S12. Summarize the enterprise management information data after the filtering operation to obtain processed enterprise management information data.
3. The enterprise management information security data processing method according to claim 1, characterized in that: The step of clustering the processed enterprise management information data by clustering method and outputting the clustering results comprises the following steps: S21, randomly select a sample point from the training set as the first initial data center, set a sample point to represent a group of processed enterprise management information data, and each data center represents a data category; S22, calculating the distance D between each sample in the training set and the current existing data center; S23, based on the calculated distance between each sample and the current existing data center, calculate the probability P of each sample being selected as the next data center; S24. Repeat steps S22 and S23 until convergence, and output the converged k data centers and clustered enterprise management information data.
4. The enterprise management information security data processing method according to claim 3 is characterized in that: The calculation of the distance D between each sample in the training set and the current existing data center includes the following steps: The distance D formula between each sample and the current data center is as follows: D(x i )=|x i -a k |; Among them, D(x i ) represents the sample x i To data center a k The distance, x i represents the i-th sample, a k represents the kth data center.
5. The enterprise management information security data processing system and method according to claim 3, characterized in that: The method of calculating the probability P of each sample being selected as the next data center based on the calculated distance between each sample and the currently existing data center includes the following steps: The probability P formula for each sample being selected as the next data center is as follows: Among them, P(x i ) represents the sample x i The probability of being selected as the next data center, X represents the training set.
6. The enterprise management information security data processing method according to claim 1, characterized in that: The step of inputting the clustered enterprise management information data into a convolutional neural network and extracting the feature vector of the clustered enterprise management information data through the convolutional neural network comprises the following steps: S31, taking the clustered enterprise management information data as input and inputting it into the convolutional neural network; S32, extracting features from the clustered enterprise management information data input by a convolutional neural network; S33, after receiving the input clustered enterprise management information data, the convolution layer moves on the input clustered enterprise management information data through the set convolution kernel according to the set step size, and performs convolution calculation on the corresponding area of each step with the feature value of the area, so as to realize feature extraction of each clustered enterprise management information data; The convolution calculation formula is as follows: F = I × ω + b; Among them, I represents the input clustered enterprise management information data, ω represents the weight of the corresponding convolution kernel, b represents the bias value, and F represents the output feature data; S34. In the process of continuous convolution of the convolutional neural network, the output of the previous convolutional neural network will be used as the input of the next convolutional neural network, and the data will be processed by the activation function in the process of inputting into the next convolutional neural network; S35, extracting features of the clustered enterprise management information data in the process of continuously stacking convolution and pooling; S36, passing the extracted and clustered enterprise management information data features to the fully connected layer; S37. Expand and combine the data through the fully connected layer to obtain a set of feature vectors and save them.
7. The enterprise management information security data processing system and method according to claim 6, characterized in that: The processing of the activation function comprises the following steps: Set the input value to μ λ (λ=1,2,……,L), each output value μ λ The corresponding input weight is ω λ , the output result after inputting the input value into the convolutional neural network is: Among them, g is the corresponding activation function, and Y is the output result of the activation function.
8. The enterprise management information security data processing method according to claim 1, characterized in that: The real-time collection of enterprise management information data and feature extraction through a convolutional neural network, and the verification of the security of the real-time collected enterprise management information data by feature comparison, include the following steps: Calculate the similarity between the feature vector of the real-time collected enterprise management information data and the feature vector of the stored enterprise management information data, and set a similarity threshold; The similarity calculation formula is as follows: in, Represents the feature vector and the eigenvector The similarity value between Represents the feature vector of real-time collection of enterprise management information data, A feature vector representing the stored enterprise management information data; A similarity threshold is set. When the calculated similarity is less than the set threshold, it means that the feature vectors of the real-time collected enterprise management information data and the stored enterprise management information data are not similar. It is determined that there is an abnormality in the real-time collected enterprise management information data currently collected in real time. When the calculated similarity is greater than or equal to the set threshold, it means that the feature vectors of the real-time collected enterprise management information data and the stored enterprise management information data are similar, and it is determined that the real-time collected enterprise management information data currently collected in real time is safe.
9. The enterprise management information security data processing method according to claim 1, characterized in that: The real-time enterprise management information data of each category after security verification is compared with the enterprise management information data of different categories in the same period by using the hierarchical analysis method, including the following steps: S51, setting a hierarchical structure diagram; The hierarchy diagram is divided into three layers; The highest level is the target level, which is used to judge the target or ideal result; the middle level is the criterion level, which includes various factors in the intermediate links involved in achieving the target, and each factor is set as a criterion. The criterion value ranges from 1 to 9, and the proportion scale of each criterion is set; the lowest level is the program level, which is used to provide various measures to achieve the target; S52, construct a comparison matrix; Select the same type of data existing in different categories of enterprise management information data within the same period, and set each type of data as a factor; Set the number of factors to n, determine the weight of each factor in the highest level, and construct a judgment matrix; For two factors C h and C e , use q he Represents C h and C e The ratio of the degree of influence on the top layer, measured on a scale of 1 to 9. he (h,e=1,2,...,n); Determination of the ratio scale: q he Take 9 levels from 1 to 9, the ratio scale q eh Get q he The reciprocal of; Summarize the comparison matrix Q of each factor pairwise comparison = (q he ) n×n ; S53, relative weight vector is determined; The geometric mean of each row and column vector of the comparison matrix Q is normalized, and the normalized vector value is used as the weight; The relative weight vector determination formula is as follows: Among them, w j represents the relative weight vector of the jth factor; Based on the calculated weight of each factor, different categories of enterprise management information data in the same period are compared, and the weights are summarized, and the enterprise management information data with the highest weight is set to have the highest security evaluation.
10. A system for implementing the enterprise management information security data processing method according to any one of claims 1 to 9, characterized in that: include: Data collection module, clustering module, feature extraction module, safety verification module and safety evaluation module; The data acquisition module is used to collect enterprise management information data in real time; The clustering module is used to cluster the enterprise management information data collected in real time; The feature extraction module is used to extract features from the clustered enterprise management information data; The security verification module is used to verify the security of the enterprise management information data collected in real time by means of feature comparison; The security evaluation module is used to compare and evaluate enterprise management information data of different categories within the same period through the hierarchical analysis method.