Power data security compliance analysis method and system based on data classification and grading
Through a method based on data classification and grading, combined with natural language processing and traffic mirroring technology, identifying and responding to violations in the power system, the problem of difficult data security and compliance in the power system is solved, and comprehensive monitoring of the data environment is achieved and refined management.
Patent Information
- Application Number
- CN202411021493.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-07-29
AI Technical Summary
Data security and compliance in power systems are difficult to effectively guarantee, and there are risks such as data leakage, unauthorized access and non-compliant use, and an efficient data security compliance analysis method is needed.
Using a method based on data classification and grading, the power text data is divided into word, part-of-speech annotation and entity recognition through natural language processing technology, a data classification and grading matrix is constructed, and the network traffic is monitored using traffic mirroring technology, combining abnormal data behavior recognition model and security combination The rules are linked to the analysis, identify and respond to violations.
It realizes comprehensive monitoring and refined management of the power data environment, promptly detects and responds to security threats and violations, ensures data security and compliance, and ensures flexibility and extensibility of access control and data protection.
Smart Images

Figure CN118981535A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data management, and in particular to a method and system for analyzing power data security and compliance based on data classification and grading. Background Art
[0002] With the digital transformation of the power system, data has become a vital asset in its operation. The power system covers multiple links from power generation, transmission, distribution to power consumption, and there are many types of data, including equipment operation data, user power consumption information, market transaction data, and regulation data. These data are not only huge in quantity, but also highly sensitive and important.
[0003] In the power system, data security and compliance have become a thorny challenge. Data leakage, unauthorized access, non-compliant use and other issues will bring major risks to the safe and stable operation of the power system and user privacy protection. Therefore, an efficient and comprehensive data security compliance analysis method is urgently needed to enhance the protection of power data. Summary of the invention
[0004] In order to solve the above problems, the purpose of the present invention is to provide a power data security and compliance analysis method based on data classification and grading, which can effectively improve the security and compliance of data processing.
[0005] To achieve the above object, the present invention adopts the following technical solutions:
[0006] A power data security compliance analysis method based on data classification and grading, characterized by comprising the following steps:
[0007] S1: Obtain power text data from various data sources of the power system, and use natural language processing technology to perform word segmentation, part-of-speech tagging, and entity recognition on the power text data, and represent the processed data as a feature vector;
[0008] S2: Use the classification model to classify the data, and classify the data according to the classification results and the sensitivity and importance of the data, build a data classification matrix, and store the classified data in a distributed database;
[0009] S3: Build an abnormal data behavior recognition model;
[0010] S4: Use traffic mirroring technology to monitor network traffic, collect data access records, and send mirrored traffic data to an abnormal data behavior recognition model to identify abnormal data behavior;
[0011] S5: Based on the data classification and grading matrix, analyze and judge the association between abnormal data behavior and security compliance rules to identify violations;
[0012] S6: Based on the abnormal data and violation identification results, use the log and reporting system (such as ELK Stack) for visualization and generate periodic reports.
[0013] Furthermore, S1 is specifically:
[0014] Obtain text data from different power system data sources;
[0015] Jieba is used for word segmentation to segment the power text data into word sequences;
[0016] Assume Text is the original text data, the word segmentation process can be expressed as:
[0017] Words=Segmentation(Text)
[0018] Among them, Segmentation is the word segmentation function; Words is the sequence of words after word segmentation;
[0019] Mark each participle with its part of speech, including noun, verb, and adjective;
[0020] Pos_Tags=PosTagging(Words)
[0021] Among them, Pos_Tags represents the part-of-speech tag sequence obtained through part-of-speech tagging; PosTagging represents the part-of-speech tagging function;
[0022] Based on the named entity recognition algorithm, entities in the power field are identified, including equipment name, time, and location; entity recognition can be expressed as:
[0023] Entities = NER(Pos_Tags)
[0024] Among them, Entities represents the entity information identified by entity recognition; NER represents the entity recognition algorithm
[0025] The processed text data is vectorized based on the bag-of-words model BoW.
[0026] Furthermore, the processed text data is vectorized based on the bag-of-words model BoW, as follows:
[0027] Build a vocabulary based on all documents, record each word and its index, let D represent the entire text dataset, and the vocabulary V is represented as:
[0028] V={w1,w2,…,w i ,..,w N}
[0029] Among them, w i is the i-th word in the vocabulary, and there are N different words in total;
[0030] For each document, generate a word frequency vector. The dimension of the word frequency vector is equal to the size of the vocabulary, and each element represents the number of times the corresponding word appears in the document.
[0031] Assume d j Generate a word frequency vector x for the jth document in the text dataset D j for:
[0032] x j =[x j1 ,x j2 ,…,x ji ,...,x jN ]
[0033] Among them, x ji Representation word w i In the document d j The number of times it appears in
[0034] Generate the word frequency matrix X of the entire text dataset:
[0035] Among them, M is the total number of documents, and each row represents the term frequency vector of a document.
[0036]
[0037] Furthermore, the classification model is built based on SVM, as follows:
[0038] For each category k∈{1,2,…,K}, a binary SVM model is trained separately. For each category k, the label transformation is defined as:
[0039]
[0040] in, is the converted label, y i′ is the label before conversion;
[0041] The objective function for training the SVM model for category k is as follows:
[0042]
[0043] Among them, w k ,b k are the weight vector and bias of the SVM model for category k respectively; is the regularization term; C is the regularization parameter; w, b represent weight and bias respectively; T represents transposition;
[0044] When all SVM models are trained, the new test sample x test Perform classification and select the category with the highest score as the final prediction result:
[0045]
[0046] Furthermore, data classification is performed based on the classification results and the sensitivity and importance of the data, as follows:
[0047] Define the data classification function G:
[0048] G(X)=α·C(X)+β·S(X)+γ·I(X);
[0049] Among them, C(X) is the SVM classification result; S(X) is the data sensitivity score; I(X) is the data importance score; α, β, γ are weight parameters;
[0050] Evaluate sensitivity and importance based on entity recognition results and predefined rules:
[0051]
[0052]
[0053] Among them, w i″ For entity e i″ The weight, v j″ Keyword j″ The weight of ; Π is the indicator function;
[0054] Divide the data into different levels according to the value of G(x);
[0055] A data classification and grading matrix is obtained based on the classification results and grading results.
[0056] Furthermore, the abnormal data behavior recognition model is constructed as follows:
[0057] Collect historical behavior data, including user ID, operation type, data type, data volume, timestamp, IP address, and operation results, and pre-process the historical behavior data, including data cleaning and data standardization;
[0058] Based on principal component analysis PCA, the best feature set is obtained;
[0059] An abnormal data behavior recognition model is constructed based on the isolation forest, and is trained based on the optimal feature set to obtain the abnormal data behavior recognition model.
[0060] Furthermore, based on principal component analysis PCA, the best feature set is obtained, as follows:
[0061] Calculate the covariance matrix C′ of the standardized data matrix X′:
[0062]
[0063] Where m is the number of samples;
[0064] Perform eigenvalue decomposition on the covariance matrix C′ to obtain the eigenvalue matrix Λ and eigenvector matrix W′:
[0065] C′W′=W′Λ;
[0066] Arrange the eigenvalues in descending order and calculate the proportion of explained variance corresponding to each eigenvalue:
[0067]
[0068] Where m′ is the total number of features, λ u is the u-th eigenvalue;
[0069] Select the first k principal components whose sum of explained variance proportions reaches the preset threshold, and the corresponding eigenvectors form a dimensionality reduction matrix;
[0070] Based on the dimension reduction matrix, the standardized data matrix X′ is mapped to the new feature space to obtain the dimensionally reduced data matrix.
[0071] Furthermore, based on the data classification and grading matrix, the association between abnormal data behavior and security compliance rules is analyzed and judged to identify violations, as follows:
[0072] According to the data classification and grading matrix, analyze the data categories and levels involved in abnormal data behavior:
[0073] Convert abnormal data behavior into a structured table format, including the following key fields: unique identifier event_id; user IDuser_id who performs the behavior; behavior typeaction; resource IDresource_id of the resource being operated; data classificationresource_classification to which the resource belongs; data security levelresource_level: to which the resource belongs, and timestamp of the time when the behavior occurred;
[0074] According to the data classification and grading matrix, clarify the data classification and security level of each resource, and find the corresponding information from the data classification and grading matrix according to the resource_id in the abnormal data behavior;
[0075] Match abnormal data behavior with the preset security compliance rule base to determine whether the behavior violates relevant regulations;
[0076] Based on the importance and sensitivity of the data, as well as the severity of the violation, a risk assessment is conducted on the violation, and detailed information on the violation is output, including: description of the violation, the data category and level involved, the violated rules, and the risk assessment results.
[0077] A power data security compliance analysis system based on data classification and grading includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps in the power data security compliance analysis method based on data classification and grading as described above.
[0078] A computer storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the above method steps.
[0079] The present invention has the following beneficial effects:
[0080] 1. The present invention realizes comprehensive monitoring and refined management of the data environment, timely discovers and responds to various potential security threats and violations, and ensures data security and compliance;
[0081] 2. The present invention constructs a data classification and grading matrix to describe in detail the security level and protection requirements of various types of data, and stores the classified data in a distributed manner in a database to ensure the flexibility and scalability of access control and data protection. It also constructs a model for identifying abnormal data behavior based on machine learning to identify abnormal data access, data leakage and other abnormal behaviors;
[0082] 3. The present invention adopts traffic mirroring technology to monitor data traffic in the network in real time and collect data access records. By mirroring traffic, various user operations on data are captured, and these data are sent to the abnormal data behavior recognition model for analysis, and abnormal behaviors are recognized and marked in real time. Based on the data classification and grading matrix, the identified abnormal data behaviors are further analyzed, and combined with pre-defined security compliance rules, compliance is determined to achieve comprehensive monitoring and refined management of the data environment. BRIEF DESCRIPTION OF THE DRAWINGS
[0083] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION
[0084] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:
[0085] refer to Figure 1 In this embodiment, a power data security compliance analysis method based on data classification and grading includes the following steps:
[0086] S1: Obtain power text data from various data sources of the power system, and use natural language processing technology to perform word segmentation, part-of-speech tagging, and entity recognition on the power text data, and represent the processed data as a feature vector;
[0087] S2: Use the classification model to classify the data, and classify the data according to the classification results and the sensitivity and importance of the data, build a data classification matrix, and store the classified data in a distributed database;
[0088] S3: Build an abnormal data behavior recognition model;
[0089] S4: Use traffic mirroring technology to monitor network traffic, collect data access records, and send mirrored traffic data to an abnormal data behavior recognition model to identify abnormal data behavior;
[0090] S5: Based on the data classification and grading matrix, analyze and judge the association between abnormal data behavior and security compliance rules to identify violations;
[0091] S6: Based on the abnormal data and violation identification results, use the log and reporting system (such as ELK Stack) for visualization and generate periodic reports.
[0092] In this embodiment, S1 is specifically:
[0093] Acquiring text data from different power system data sources, including sensors, monitoring systems, and user reports;
[0094] Jieba is used for word segmentation to segment the power text data into word sequences;
[0095] Assume Text is the original text data, the word segmentation process can be expressed as:
[0096] Words=Segmentation(Text)
[0097] Among them, Segmentation is the word segmentation function; Words is the sequence of words after word segmentation;
[0098] For example, the text "In July 2023, the transformer failed." is segmented to obtain the word sequence Words = {"2023", "July", ",","Transformer", "Occurrence", "Failure", "."};
[0099] Mark each participle with its part of speech, including noun, verb, and adjective;
[0100] Pos_Tags=PosTagging(Words)
[0101] Among them, Pos_Tags represents the part-of-speech tag sequence obtained through part-of-speech tagging; PosTagging represents the part-of-speech tagging function;
[0102] For example, the part-of-speech tagging result: Pos_Tags = {("2023","n"),("July","n"),(",","x"),("transformer","n"),("appeared","v"),("fault","n"),("."","x")}.
[0103] Based on the named entity recognition algorithm, entities in the power field are identified, including equipment name, time, and location; entity recognition can be expressed as:
[0104] Entities = NER(Pos_Tags)
[0105] Among them, Entities represents the entity information identified by entity recognition; NER represents the entity recognition algorithm
[0106] For example, the entity recognition result is: Entities = {("2023","DATE"),("July","DATE"),("Transformer","DEVICE"),("Fault","EVENT")}
[0107] The processed text data is vectorized based on the bag-of-words model BoW.
[0108] In this embodiment, the processed text data is vectorized based on the bag-of-words model BoW, as follows:
[0109] Build a vocabulary based on all documents, record each word and its index, let D represent the entire text dataset (containing multiple documents), and the vocabulary V is represented as:
[0110] V={w1,w2,…,w i ,..,w N}
[0111] Among them, w i is the i-th word in the vocabulary, and there are N different words in total;
[0112] For each document, generate a word frequency vector. The dimension of the word frequency vector is equal to the size of the vocabulary, and each element represents the number of times the corresponding word appears in the document.
[0113] Assume d j Generate a word frequency vector x for the jth document in the text dataset D j for:
[0114] x j =[x j1 ,x j2 ,…,x ji ,...,x jN ]
[0115] Among them, x ji Representation word w i In the document d j The number of times it appears in
[0116] Generate the word frequency matrix X of the entire text dataset:
[0117] Among them, M is the total number of documents, and each row represents the term frequency vector of a document.
[0118]
[0119] In this embodiment, the classification model is constructed based on SVM, as follows:
[0120] For each category k∈{1,2,…,K}, a binary SVM model is trained separately. For each category k, the label transformation is defined as:
[0121]
[0122] in, is the converted label, y i′ is the label before conversion;
[0123] The objective function for training the SVM model for category k is as follows:
[0124]
[0125] Among them, w k ,b k are the weight vector and bias of the SVM model for category k respectively; is the regularization term; C is the regularization parameter; w, b represent weight and bias respectively; T represents transposition;
[0126] When all SVM models are trained, the new test sample x test Perform classification and select the category with the highest score as the final prediction result:
[0127]
[0128] In this embodiment, data classification is performed according to the classification results and the sensitivity and importance of the data, as follows:
[0129] Define the data classification function G:
[0130] G(X)=α·C(X)+β·S(X)+γ·I(X);
[0131] Among them, C(X) is the SVM classification result; S(X) is the data sensitivity score; I(X) is the data importance score; α, β, γ are weight parameters;
[0132] Evaluate sensitivity and importance based on entity recognition results and predefined rules:
[0133]
[0134]
[0135] Among them, w i″ For entity e i″ The weight, v j″ Keyword j″ The weight of ; Π is the indicator function;
[0136] Divide the data into different levels according to the value of G(x);
[0137] A data classification and grading matrix is obtained based on the classification results and grading results.
[0138] In this embodiment, the abnormal data behavior recognition model is constructed as follows:
[0139] Collect historical behavior data, including user ID, operation type, data type, data volume, timestamp, IP address, and operation results, and pre-process the historical behavior data, including data cleaning and data standardization;
[0140] Based on principal component analysis PCA, the best feature set is obtained;
[0141] An abnormal data behavior recognition model is constructed based on the isolation forest, and is trained based on the optimal feature set to obtain the abnormal data behavior recognition model.
[0142] In this embodiment, based on principal component analysis PCA, the optimal feature set is obtained, as follows:
[0143] Calculate the covariance matrix C′ of the standardized data matrix X′:
[0144]
[0145] Where m is the number of samples;
[0146] Perform eigenvalue decomposition on the covariance matrix C′ to obtain the eigenvalue matrix Λ and eigenvector matrix W′:
[0147] C′W′=W′Λ;
[0148] Arrange the eigenvalues in descending order and calculate the proportion of explained variance corresponding to each eigenvalue:
[0149]
[0150] λ u
[0151] Among them, m′ is the total number of features, and is the uth eigenvalue;
[0152] Select the first k principal components whose sum of explained variance proportions reaches the preset threshold, and the corresponding eigenvectors form a dimensionality reduction matrix;
[0153] Based on the dimension reduction matrix, the standardized data matrix X′ is mapped to the new feature space to obtain the dimensionally reduced data matrix.
[0154] In this embodiment, based on the data classification and grading matrix, the association between abnormal data behavior and security compliance rules is analyzed and judged to identify illegal behaviors, as follows:
[0155] According to the data classification and grading matrix, analyze the data categories and levels involved in abnormal data behavior:
[0156] Convert abnormal data behavior into a structured table format, including the following key fields: unique identifier event_id; user ID user_id who performed the behavior; behavior type action, such as read, write, delete, etc.; resource ID resource_id of the resource being operated; data classification resource_classification: such as personal sensitive information, commercial secrets, etc.; data security level resource_level: such as public, secret, confidential, top secret, etc. Behavior occurrence time timestamp;
[0157] According to the data classification and grading matrix, clarify the data classification and security level of each resource, and find the corresponding information from the data classification and grading matrix according to the resource_id in the abnormal data behavior;
[0158] Match abnormal data behavior with the preset security compliance rule base to determine whether the behavior violates relevant regulations;
[0159] In this embodiment, a formal language (e.g., ABAC policy, logical expression) is used to express security compliance rules. For example:
[0160] Rule 1: Non-administrator users are prohibited from reading personal sensitive information - top secret data.
[0161] Rule 2: If a user accesses commercial confidential data more than 100 times within 10 minutes, an alarm is triggered.
[0162] Match each abnormal behavior record with the rule base and record the ID, description and other information of the triggered rule.
[0163] Based on the importance and sensitivity of the data, as well as the severity of the violation, a risk assessment is conducted on the violation, and detailed information on the violation is output, including: description of the violation, the data category and level involved, the violated rules, and the risk assessment results.
[0164] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0165] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0166] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0167] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0168] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.
Claims
1. A power data security compliance analysis method based on data classification and grading, characterized in that: The following steps are involved: S1: Obtain power text data from various data sources of the power system, and use natural language processing technology to perform word segmentation, part-of-speech tagging, and entity recognition on the power text data, and represent the processed data as a feature vector; S2: Use the classification model to classify the data, and classify the data according to the classification results and the sensitivity and importance of the data, build a data classification matrix, and store the classified data in a distributed database; S3: Build an abnormal data behavior recognition model; S4: Use traffic mirroring technology to monitor network traffic, collect data access records, and send mirrored traffic data to an abnormal data behavior recognition model to identify abnormal data behavior; S5: Based on the data classification and grading matrix, analyze and judge the association between abnormal data behavior and security compliance rules to identify violations; S6: Based on the abnormal data and violation identification results, use the log and reporting system for visualization and generate periodic reports.
2. According to claim 1, a power data security compliance analysis method based on data classification and grading is characterized in that: The S1 is specifically: Obtain text data from different power system data sources; Jieba is used for word segmentation to segment the power text data into word sequences; Assume Text is the original text data, the word segmentation process can be expressed as: Words=Segmentation(Text) Among them, Segmentation is the word segmentation function; Words is the sequence of words after word segmentation; Mark each participle with its part of speech, including noun, verb, and adjective; Pos_Tags=PosTagging(Words) Among them, Pos_Tags represents the part-of-speech tag sequence obtained through part-of-speech tagging; PosTagging represents the part-of-speech tagging function; Based on the named entity recognition algorithm, entities in the power field are identified, including equipment name, time, and location; entity recognition can be expressed as: Entities = NER(Pos_Tags) Among them, Entities represents the entity information identified by entity recognition; NER represents the entity recognition algorithm The processed text data is vectorized based on the bag-of-words model BoW.
3. According to claim 2, a method for analyzing power data security compliance based on data classification and grading is characterized in that: The processed text data is vectorized based on the bag-of-words model BoW, as follows: Build a vocabulary based on all documents, record each word and its index, let D represent the entire text dataset, and the vocabulary V is represented as: V={w1,w2,…,w i ,..,In N } Among them, w i is the i-th word in the vocabulary, and there are N different words in total; For each document, generate a word frequency vector. The dimension of the word frequency vector is equal to the size of the vocabulary, and each element represents the number of times the corresponding word appears in the document. Assume d j Generate a word frequency vector x for the jth document in the text dataset D j for: x j =[x j1 ,x j2 ,…,x ji ,...,x jN ] Among them, x ji Representation word w i In the document d j The number of times it appears in Generate the word frequency matrix X of the entire text dataset: Among them, M is the total number of documents, and each row represents the term frequency vector of a document.
4. The method for analyzing power data security compliance based on data classification and grading according to claim 1 is characterized in that: The classification model is constructed based on SVM, as follows: For each category k∈{1,2,…,K}, a binary SVM model is trained separately. For each category k, the label transformation is defined as: in, is the converted label, y i′ is the label before conversion; The objective function for training the SVM model for category k is as follows: Among them, w k ,b k are the weight vector and bias of the SVM model for category k respectively; is the regularization term; C is the regularization parameter; w, b represent weight and bias respectively; T represents transposition; When all SVM models are trained, the new test sample x test Perform classification and select the category with the highest score as the final prediction result:
5. The method for analyzing power data security compliance based on data classification and grading according to claim 4 is characterized in that: The data is classified according to the classification results and the sensitivity and importance of the data, as follows: Define the data classification function G: G(X)=α·C(X)+β·S(X)+γ·I(X); Among them, C(X) is the SVM classification result; S(X) is the data sensitivity score; I(X) is the data importance score; α, β, γ are weight parameters; Evaluate sensitivity and importance based on entity recognition results and predefined rules: Among them, w i″ For entity e i″ The weight, v j″ Keyword j″ The weight of ; Π is the indicator function; Divide the data into different levels according to the value of G(x); A data classification and grading matrix is obtained based on the classification results and grading results.
6. The method for analyzing power data security compliance based on data classification and grading according to claim 1 is characterized in that: The abnormal data behavior recognition model is constructed as follows: Collect historical behavior data, including user ID, operation type, data type, data volume, timestamp, IP address, and operation results, and pre-process the historical behavior data, including data cleaning and data standardization; Based on principal component analysis PCA, the best feature set is obtained; An abnormal data behavior recognition model is constructed based on the isolation forest, and is trained based on the optimal feature set to obtain the abnormal data behavior recognition model.
7. A power data security compliance analysis method based on data classification and grading according to claim 6, characterized in that: The method of obtaining the best feature set based on principal component analysis PCA is as follows: Calculate the covariance matrix C′ of the standardized data matrix X′: Where m is the number of samples; Perform eigenvalue decomposition on the covariance matrix C′ to obtain the eigenvalue matrix Λ and eigenvector matrix W′: C′W′=W′Λ; Arrange the eigenvalues in descending order and calculate the proportion of explained variance corresponding to each eigenvalue: l u Among them, m′ is the total number of features, and is the uth eigenvalue; Select the first k principal components whose sum of explained variance proportions reaches the preset threshold, and the corresponding eigenvectors form a dimensionality reduction matrix; Based on the dimension reduction matrix, the standardized data matrix X′ is mapped to the new feature space to obtain the dimensionally reduced data matrix.
8. The method for analyzing power data security compliance based on data classification and grading according to claim 5 is characterized in that: The data classification and grading matrix is used to analyze and judge the association between abnormal data behavior and security compliance rules, and identify violations, as follows: According to the data classification and grading matrix, analyze the data categories and levels involved in abnormal data behavior: Convert abnormal data behavior into a structured table format, including the following key fields: unique identifier event_id; user IDuser_id who performs the behavior; behavior typeaction; resource IDresource_id of the resource being operated; data classificationresource_classification to which the resource belongs; data security levelresource_level: to which the resource belongs, and timestamp of the time when the behavior occurred; According to the data classification and grading matrix, clarify the data classification and security level of each resource, and find the corresponding information from the data classification and grading matrix according to the resource_id in the abnormal data behavior; Match abnormal data behavior with the preset security compliance rule base to determine whether the behavior violates relevant regulations; Based on the importance and sensitivity of the data, as well as the severity of the violation, a risk assessment is conducted on the violation, and detailed information on the violation is output, including: description of the violation, the data category and level involved, the violated rules, and the risk assessment results.
9. A power data security compliance analysis system based on data classification and grading, characterized in that: It includes a processor, a memory and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps in the power data security compliance analysis method based on data classification and grading as described in any one of claims 1 to 8.
10. A computer storage medium, characterized in that: The computer storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the method steps according to any one of claims 1 to 8.
Citation Information
Patent Citations
Vehicle maintenance reimbursement behavior anomaly detection method and device and storage medium
CN110852860A
Electric power data analysis method and system based on data grading model
CN112257425A
Data anomaly detection method in power system
CN114358160A
Pipe network data security management method and device
CN114911908A
Sensitive information discovery and automatic classification and grading method based on multi-modal fusion
CN116049397A
Cited By
Data classification and grading method, system and equipment based on big data and medium
CN119441564A
Abnormal behavior state recognition and early warning method based on video image processing
CN119963606A
Behavior Abnormality State Recognition and Early Warning Method Based on Video Image Processing
CN119963606B
Power sensitive database lightweight supervision method and system based on metadata analysis
CN121029706A
Power-sensitive database lightweight supervision method and system based on metadata analysis
CN121029706B
