CN102915423A - System and method for filtering electric power business data on basis of rough sets and gene expressions - Google Patents

Abstract

The present invention proposes a new system and method for power business data filtering to solve the problem of sensitive data filtering in the power business data protection system. The system mainly includes data preprocessor, data attribute reduction controller, data filtering control device and other components. The rough set method is used to simplify the power business data set to reduce the complexity of the data. The gene expression method is used to establish the power business data classification model. Based on the classification model, the sensitivity of the power business data is actively identified, and the policy knowledge base is used to prevent leakage.

Description

A power service data filtering system and method based on rough sets and gene expression

technical field

The invention relates to sensitive data filtering in a power business data security protection system, mainly used to solve the problem of sensitive data filtering in the power business data protection system, and belongs to the field of information security.

Background technique

From the "Eleventh Five-Year Plan" to the present, the State Grid Corporation has passed the construction of the "SG186" project, and its information system has basically covered its main business areas. For the informatization construction of the "Twelfth Five-Year Plan", the State Grid Corporation of China proposed the "SG-ERP" plan to provide information support for the smart grid around the smart grid. On the basis of the above, deepen the data collection, transmission, storage and utilization of all links of electric power, realize the digitalization of data collection, automation of production process, interactive business processing, informatization of operation and management, and scientific strategic decision-making; help the construction of smart grid and comprehensively improve The level of production, operation, management and decision-making of the company. It can be seen that SG-ERP, as an informatization construction project, has taken business integration as the primary starting point, has focused on breaking the previous situation where power automation and power informatization were relatively independent and developed independently, and has extended informatization construction to the core business of power production.

With the implementation and continuous advancement of SG-ERP information construction, more and more power business application systems (safety production, marketing management, material management, etc.) will widely use mobile smart terminal access methods and power information intranet Carry out real-time and non-real-time data communication and data exchange. At the same time, with the construction of a strong smart grid, the use of a large number of intelligent collection and intelligent terminal equipment, and the widespread use of wireless communication technologies such as 3G/WIFI, the ways for various power service application data to be destroyed and leaked have greatly increased. Power business application data protection is the basis for the safe and stable operation of business systems in all links of smart grid generation, transmission, transformation, distribution, and use. At the same time, with the construction of the three major data centers of the State Grid Corporation of China, various business system data are increasingly stored in a centralized manner, and the reliable storage and protection of sensitive data has become increasingly important.

For the protection of sensitive data of power business applications, the core technology is that power business application data can be effectively identified during storage and transmission, and filtered according to corresponding policies, so as to prevent its leakage. There are also many methods for data identification and filtering, including policy-based matching and BP network-based methods. Different data identification and filtering methods have different effects, which are mainly reflected in performance, automation, daily management, and scalability. The ultimate purpose of identifying and filtering power business application data is to effectively protect it through certain strategies and methods during its storage and transmission, and prevent the leakage of business data involving State Grid Corporation's confidentiality. Therefore, it is of great significance to study an effective data identification and filtering scheme in the power business application sensitive data protection system to improve the protection ability of sensitive data, reduce its leakage, and ensure the safe and stable operation of the power business system.

With the continuous and in-depth development of the information technology of the State Grid Corporation of China, various information technologies have gradually matured and been applied to various power business applications. The data used are located on different storage nodes. The existing security mechanism cannot guarantee that it will not be leaked during storage and transmission. In order to ensure the security of these power business application data during storage and transmission and prevent the leakage of sensitive data, methods such as full-text encryption, policy matching, and artificial intelligence can be used. Full-text encryption can solve the security of data during storage and transmission, but it is difficult to effectively guarantee its performance and accuracy for massive power business data. The strategy matching method can not only solve the security problem of power business data, but also ensure its performance and accuracy. However, for different power business systems, the strategies that need to be formulated are different, so a set of more complex policy library support is needed to achieve power. The purpose of business application data leakage prevention. Various artificial intelligence methods can not only meet the needs of data security protection, but also can make full use of the intelligence and self-learning ability of the method to improve the performance of data filtering.

The data identification and filtering method is mainly considered from the following aspects: (1) For the collected various types of power business data, in order to facilitate the later data filtering and processing, perform preprocessing such as data cleaning and noise data elimination; (2) For the preprocessed power business data, combined with the rough set method to reduce its attributes to reduce the complexity of the data; (3) For the power business data after attribute reduction, the gene expression method is used to establish a classification model of power business data, based on During the storage and transmission of power business data, the classification model actively identifies its sensitivity, and cooperates with the policy knowledge base to prevent its leakage.

Contents of the invention

In order to solve the above problems, the purpose of the present invention is to provide a new system and method for filtering power business data to solve the problem of sensitive data filtering in the power business data protection system. The mechanism adopted is a strategic method By using the present invention, the leakage of sensitive data of various electric power services can be prevented to the greatest extent when the sensitive data of electric power service is transmitted between the terminal and the network, thereby protecting the safe and stable operation of various electric service systems.

According to one aspect of the present invention, a power service data filtering system based on rough sets and gene expressions is proposed, and the power service data filtering system includes:

A data preprocessor is used to perform data preprocessing on various types of power business data to be processed, and the data preprocessing may include data cleaning and data transformation, etc.;

The data attribute reduction controller is used to reduce the power business data set and simplify the power business data set;

The data filtering controller is used to intelligently filter sensitive power business data to ensure the security of power business data transmission.

According to one aspect of the present invention, in the power service data security protection system, the data attribute reduction controller adopts a rough set method to reduce the power service data set to reduce the complexity of the data.

According to one aspect of the present invention, in the power business data security protection system, the data filtering controller uses the gene expression method to establish a power business data classification model, and actively identifies the sensitivity of the power business data based on the classification model, and cooperates with the policy knowledge base to prevent leakage.

According to one aspect of the present invention, a method for filtering power business data based on rough sets and gene expressions is proposed. The power business data method is used for power business data security protection, and the steps included are:

Perform data preprocessing on various types of power business data to be processed through the data preprocessor, and the data preprocessing may include data cleaning and data transformation, etc.;

Reduce the power business data set through the data attribute reduction controller to simplify the power business data set;

Intelligently filter sensitive power business data through the data filtering controller to ensure the security of power business data transmission.

According to one aspect of the present invention, in the data filtering method, the data attribute reduction controller adopts the rough set method to reduce the power business data set, reducing the complexity of the data.

According to one aspect of the present invention, in the data filtering method, the data filtering controller uses the gene expression method to establish a power service data classification model, actively identifies the sensitivity of the power service data based on the classification model, and cooperates with the policy knowledge base to prevent leakage.

Description of drawings

The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

Fig. 1 is a structure diagram of data filtering according to an embodiment of the present invention, which mainly includes: a data preprocessor, a data attribute reduction controller, a data filtering controller and a data filtering operation core.

Fig. 2 is a schematic diagram of a reference architecture according to an embodiment of the present invention, showing components included in the present invention.

Fig. 3 is a schematic flowchart of a method according to an embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

The method of the present invention is a strategic method. By preprocessing the power business data in various storage or transmission processes, the power business sample data set that conforms to rough set processing is obtained, and then the preprocessed power business data set is processed by rough set theory. The effective attribute reduction of the power business sample data set greatly reduces the time complexity of power business data identification and filtering. Finally, based on the gene expression algorithm, the reduced data set is classified and processed to realize the filtering of various power business data.

Figure 1 shows the structure diagram of the power business data filtering system based on rough sets and gene expressions, which mainly includes five parts: data preprocessor, data attribute reduction controller, data filtering controller and data filtering operation core . The core of the data filtering operation in the figure includes the specific operations required to filter the data when the data is preprocessed and classified. The present invention adds three other parts to ensure that data filtering can be carried out more smoothly and effectively, ensure the ability of identifying and filtering data to the greatest extent, and reduce the risk of leakage of sensitive data of electric power business.

The specific introduction is given below:

Data preprocessor: In order to ensure the safe and stable operation of the power business system, the most important thing is the security of various power business data during storage and transmission, among which preventing data leakage is the top priority. Before identifying and protecting power business data, it is necessary to perform data preprocessing such as data cleaning and data transformation for various power business data to be processed. In order to improve the quality of filtering power business data and reduce its time complexity, firstly, for all kinds of power business data, the data is cleaned by filling in the missing values of corresponding attributes, smoothing noise data, etc.; secondly, the power business data is converted into applicable in the form of recognition filtering. In this patent, there is no limitation on the specific implementation of data preprocessing.

Data attribute reduction controller: For the preprocessed power business data set, due to the characteristics of the power business itself, the data set has many attributes. If there is no corresponding attribute reduction method, it will lead to identification and filtering of the power business data set The complexity of the process is greatly increased, and the efficiency of its processing will also be greatly reduced. The data attribute reduction controller adopts the rough set method to reduce the power business data set and deletes the attribute redundant information in the power business data set without changing the inherent classification ability of the preprocessed power business data set , to simplify the power business dataset.

Data filtering controller: In order to ensure that when the power business data is transmitted between the terminal and the network, it can quickly and effectively identify whether the power business data contains sensitive information. In the data protection system, only based on keyword matching is far from enough, because the key Word matching cannot effectively check behaviors such as changing data attributes by malicious leakers, and cannot guarantee the confidentiality of various sensitive power business data during transmission. If a malicious leaker changes the file name, file attributes, etc. of the sensitive data to be transferred, then the data protection system will not be able to keep it safe. Therefore, an intelligent and automated data identification and filtering model must be established. In this patent, gene expression is used to establish a power service data identification and filtering model.

1. Data preprocessor

Whether the power business data set needs to be preprocessed depends on whether the format of the data set meets the requirements of the identification and filtering method. In order to judge in time whether corresponding preprocessing operations are required for the power business data set, in this method, a data preprocessing rule base is established. When the user is performing data filtering, firstly, by querying the data preprocessing rule base, it is judged that the current data Whether the attribute value in the set is missing, if a certain attribute value is missing, it will be filled by using the mean value of the attribute value of the current data set; secondly, the clustering method is used to judge whether the current data set contains noise data, and if so, then Delete the corresponding noise data; finally determine whether the current data set contains character data, and if so, convert it into numerical data through a normalization method, and finally form a power business data set that meets the requirements of the data filtering method. Here we take the original power business dataset ODataSet as an example, the data structure of the entire ODataSet is shown in Table 1.

Table 1 Data structure of ODataSet

serial number Voltage (V) Current (A) Base wind speed (m/s)  … wind direction 1 220 25 25  … southeast 2 220 20  … northwest 3 220 25 20  … East 4 2000 25 20  … southeast  …  …  …  …  …

It can be seen from Table 1 that the current attribute value of the second piece of data in the original power business data set ODataSet is missing, and the attribute value is filled by calculating the mean value (25A) of the attribute value. After the clustering method, it can be found that the abnormal voltage value in the fourth data is noise data. In order not to affect the performance and accuracy of the final data filtering, this data is deleted. The wind direction attribute values passed through ODataSet are all character type, which does not meet the requirement of the data filtering method proposed in this patent for the attribute value to be numerical. Therefore, according to the combination of wind direction attribute values, it is numerically processed. The data structure form of the final data set UDataSet after the whole processing is shown in Table 2.

Table 2 Data structure of UDataSet

serial number Voltage (V) Current (A) Base wind speed (m/s)  … wind direction 1 220 25 25  … 5 2 220 25 20  … 8 3 220 25 20  … 1  …  …  …  …  …

2. Data attribute reduction controller

When data filtering is performed on the electric power business data set after the above-mentioned data preprocessing, the performance of data filtering is greatly degraded because the attributes of the data set itself are not reduced. In order to improve the efficiency of identifying and filtering sensitive power business data sets, the data attribute reduction controller uses the rough set method to reduce the preprocessed power business data sets without changing its inherent classification ability, greatly reducing the data set filtering. of complexity.

In order to clearly describe the workflow of data attribute reduction, first set the sample decision table T=<U, C∪D, V, f>, where U is the research object set of sample data, C∪D=R is the attribute of sample data Set, C={c ₁ , c ₂ ,...,c _n } is the set of condition attributes of the sample data, D={d ₁ , d ₂ ,...,d _m } is the set of decision attributes of the sample data, V=∪v _r , r∈R is a set of sample data attribute values, v _r represents the attribute value range of a certain attribute r∈R, f: U×R→V defines an information function, which specifies each object in U The attribute value of x, that is, for There is f(x, r) ∈ v _r .

The main workflow is as follows:

(1) First judge whether the sample decision table T is coordinated, if not, divide the sample decision table T into a coordinated sample decision table T′ and an uncoordinated sample decision table T″, and divide the sample decision table T″ All condition attributes in are added to the final attribute reduction set reductionSet;

(2) Then, for each condition attribute c in the coordinated sample decision table T′ condition attribute set, judge whether the positive domain POS _C (D) of the condition attribute in the sample decision table T′ relative to the decision attribute is equal to the sample decision table The conditional attribute set in T′ after removing c is relative to the positive field POS _C-{c} (D) of the decision attribute. If they are equal, it means that the conditional attribute c can be reduced, and the conditional attribute c can be added to the final attribute reduction Jane collection reductionSet.

(3) Finally, remove the attribute reduction set reductionSet from the conditional attribute set C in the sample decision table T=<U, C∪D, V, f> to obtain the reduced sample decision table T=<U, C′∪ D, V, f>.

For the coordinated sample decision table, by calculating the positive domain of its condition attribute relative to the decision attribute, it is judged whether each condition attribute in the condition attribute set can be reduced, so as to ensure that the inherent classification decision-making ability of the original sample data is not changed. To achieve the purpose of reducing the conditional attributes in the sample data.

3. Data filtering controller

In order to quickly and effectively ensure that sensitive power business data is not leaked during the transmission of various power business data, it is necessary to effectively filter sensitive power business data intelligently during the data transmission to ensure the security of power business data transmission. How to design an effective intelligent data filtering method that can not only ensure the security of data transmission, but also maximize the filtering performance of sensitive data, the present invention proposes a power service data filtering method based on gene expression.

Firstly, according to the expert knowledge base, the power business training sample data with filter attributes is constructed, as shown in Table 3, and then the gene expression method is used to mine the filter attributes of the power business training sample data after preprocessing and attribute reduction The functional relationship F=f(x ₁ ,x ₂ ,...,x _n ) between F and the corresponding condition attributes {x ₁ ,x ₂ ,...,x _n }, and then the power service test sample to be filtered Put each piece of data in the data into F=f(x ₁ ,x ₂ ,...,x _n ) to get the value F', and calculate the error between the target value F' and the actual value F, satisfying the given in advance The threshold is to judge the piece of data as sensitive data, and prevent its transmission according to the expert knowledge base.

Table 3 Example of power business training sample data with filtering attributes

Among them, a, b, c, e, f, and g represent the attribute values corresponding to the conditional attributes x ₁ , x ₂ , and x _n respectively. The value of the filter attribute F is 1, indicating that the piece of data is sensitive data, and 0, indicating that the piece of data is sensitive. The data is normal data.

The main workflow of the entire data filtering controller is as follows:

(1) Construct electric power business data with and without filter attributes, and preprocess and attribute reduce the two types of data at the same time to form the training sample data set to be mined and the test sample data set to be filtered;

(2) According to the characteristics of the training sample data set, determine the parameters of the gene expression method and initialize the population;

(3) Evaluate the fitness function value of each individual in the population;

(4) Judging whether the termination condition is satisfied, if so, go to step (7), otherwise continue;

(5) Perform various genetic operations according to probability;

(6) Create a new population and go to step (3).

(7) Return the functional relationship F=f(x ₁ ,x ₂ ,...,x _n ), and bring in the test sample data to be filtered for calculation. The data whose function value F is closest to 1 is judged as sensitive data .

According to one aspect of the present invention, a new data filtering method in the power service data security protection system can be implemented by the following steps:

Step 1: Construct electric power business datasets A and B with and without filter attributes respectively. Users can query the data preprocessing rule library according to the data preprocessor to determine whether data preprocessing is required. If preprocessing is required, enter Go to the next step, otherwise go to step 3;

Step 2: first judge whether there are missing attribute values in the current power business data sets A and B to be processed, and if so, fill them in by using the mean values of the attribute values in the current data sets A and B to be processed; Secondly, through the clustering method, it is judged whether the current power business data sets A and B contain noise data, and if so, the corresponding noise data is deleted; finally, it is judged whether the current power business data sets A and B contain character data, if If there is, transform it into numerical data through a normalization method, and finally form a power business training sample data set and a power business test sample data set that meet the requirements of the data filtering method;

Step 3: According to the power business training sample data and power business test sample data obtained by preprocessing, respectively construct the corresponding sample decision tables T _train and T _test , and then judge whether the sample decision tables T _train and T _test are coordinated, and if not , the sample decision tables T _train and T _test are divided into coordinated sample decision tables T′ _train and T′ _test and uncoordinated sample decision tables T″ _train and T″ _test respectively, and the sample decision tables T″ _train and T ″ All conditional attributes in _test are added to the final attribute reduction sets reductionSet _train and reductionSet _test respectively;

Step 4: For each condition attribute c in the coordinated sample decision table T′ _train and T′ _test condition attribute set, respectively judge the positive value of the condition attribute in the sample decision table T′ _train and T′ _test relative to the decision attribute Whether the domain is equal to the positive domain of the decision attribute after removing c from the conditional attribute set in the corresponding sample decision table. If they are equal, it means that the conditional attribute c in the sample decision table can be reduced, and the conditional attribute c is added to In the corresponding attribute reduction set;

Step 5: Remove the attribute reduction sets reductionSet _train and reductionSet _test from the condition attribute set C in the sample decision table T′ _train and T′ _test respectively, and obtain the reduced training and test sample data sets RT′ _train and RT′ _test respectively ;

Step 6: According to the characteristics of the training sample data set, determine the parameters of the gene expression method, and initialize the population;

Step 7: Run the gene expression method to mine the functional relationship F=f(x ₁ ,x ₂ ,...,x _n ) between the filter attribute and the condition attribute of the training sample data;

Step 8: According to the functional relational expression, bring in the test sample data to be filtered for calculation, if the absolute value of the error between the obtained function value F and 1 is less than 0.001, then it is judged that the data is sensitive data, and according to the filtering rule base, implement Blocking does not allow its transmission;

Step 9: The power service data filtering ends.

According to one aspect of the present invention, a power service data filtering method based on rough sets and gene expressions is adopted, which is mainly used to solve the leakage prevention problem of sensitive data of power services in the process of terminal and network transmission. The proposed method can effectively realize the identification and filtering of power business data, prevent its leakage during terminal and network transmission, and improve the security of sensitive power business data.

A specific description is given below.

The data attribute reduction controller first preprocesses the power service training and test data through the data preprocessor, and if there are many condition attributes in the power service training data, if the condition attributes cannot be effectively reduced, it will inevitably affect To the performance of data filtering, thereby increasing the risk of data leakage of sensitive power business. Therefore, an attribute reduction method based on rough sets is introduced in the data attribute reduction controller. First, the corresponding sample decision table is constructed for the preprocessed power service training and test data sets, and then the training and test sample decision tables are respectively judged whether they are Coordination, if not coordinated, the training and testing sample decision tables are divided into corresponding coordinated sample decision tables and uncoordinated sample decision tables, and all conditional attributes in the uncoordinated sample decision tables are added to the final attribute In the reduction set; secondly, for each condition attribute c in the coordinated sample decision table condition attribute set, respectively, by judging whether the positive domain of the condition attribute relative to the decision attribute is equal to the condition attribute set in the corresponding sample decision table to remove If c is equal to the positive domain of the decision attribute, it means that the conditional attribute c in the sample decision table can be reduced, and the conditional attribute c is added to the corresponding attribute reduction set. Finally, the corresponding attribute reduction sets are removed from the conditional attribute sets in the coordinated sample decision table, so as to finally obtain the reduced training and test sample data sets respectively.

The data filtering controller increases the intelligence of power business data filtering by using the gene expression method. When power business data is transmitted between the terminal and the network, it is necessary to be able to judge whether the power business data is sensitive data in a timely manner. In the data security protection system, only based on keyword matching cannot meet the actual needs, because keyword matching cannot The behavior of malicious leakers changing the attributes of the data itself is effectively checked. If the malicious leaker changes the file name and file attributes of the sensitive data to be transmitted, the data protection system will not be able to guarantee the security of the transmission process. Therefore, the gene expression method is introduced into the data filtering controller. After preprocessing and attribute reduction of the constructed power service training and test data sets, according to the characteristics of the training sample set to be mined, run the gene expression programming Algorithm, construct the functional relationship F=f(x ₁ ,x ₂ ,...,x _n ) between the filter function F and the conditional attributes (x ₁ ,x ₂ ,...,x _n ) of the training sample data to be mined, Through this functional relational expression, users can easily obtain the filter function value according to the attributes of the reduced test data set, so as to judge whether the data is sensitive data, and finally implement the protection of sensitive data with the help of the filter rule library.

In practical applications, a power company has a power business training data X with filter attributes and test data Y without filter attributes. In order to establish a data security protection system, it is necessary to base on the characteristics of the power business training data The method is used to establish the functional relationship between the filter attribute and the condition attribute of the electric power business training data X, and through this functional relationship to judge whether the electric power business test data Y without the filter attribute is sensitive data, and then use the filter rule base to determine Whether to allow the data Y to be transmitted and exchanged between the network and the terminal.

Its specific implementation method is:

(1) Judge whether training data set X and test data set Y need to be preprocessed according to the data preprocessing rule base. Data preprocessing request, and then the data preprocessor performs operations such as attribute value supplementation, noise data elimination, and character data conversion into numerical data on the training data X and test data Y after receiving the request, and finally form data that meet the requirements of the data filtering method. The power business training sample data set and the power business test sample data set;

(2) According to the pre-processed power business training sample data and power business test sample data, respectively construct the corresponding sample decision table, and then judge whether it is coordinated, if not, decompose the sample decision table into a coordinated sample decision table and an incongruous sample decision table;

(3) For each conditional attribute in the coordinated sample decision table, by calculating whether the positive domain of its conditional attribute set relative to the decision attribute set is equal to the conditional attribute set in the corresponding sample decision table after removing a conditional attribute c relative to If the positive domain of the decision attribute set is equal, it means that the conditional attribute c in the sample decision table can be reduced, and the conditional attribute c is added to the corresponding attribute reduction set;

(4) Remove the corresponding attribute reduction sets from the condition attribute sets in the training and test sample decision tables, respectively, and obtain the reduced training and test sample data sets RT′ _train and RT′ _test respectively;

(5) According to the characteristics of the training sample data set RT′ _train , determine the parameters of the gene expression method and initialize the population;

(6) Run the gene expression method to mine the functional relationship between the filter attribute and the condition attribute of the training sample data;

(7) According to the function relational expression, bring in the test sample data to be filtered for calculation, if the absolute value of the error between the obtained function value F and 1 is less than 0.001, then it is judged that the data is sensitive data, and according to the filtering rule base, implement Blocking does not allow its transmission. The entire data filtering process is over.

Although the embodiments of the present invention and their various functional components have been described in specific embodiments, it should be understood that the embodiments of the present invention can be implemented in hardware, software, firmware, middleware or a combination thereof, and the present invention Embodiments of the can be used in a variety of systems, subsystems, components, or subassemblies thereof. When implemented in software or firmware, the elements of the invention are the instructions/code segments for performing the necessary tasks. The program or code segments can be stored on a machine-readable medium (e.g., a processor-readable medium or a computer program product), or in a transmission medium or a communication link by a computer embodied in a carrier wave or a signal modulated by a carrier wave. data signal to transmit. A machine-readable medium may include any medium that can store or transmit information in a form readable and executable by a machine (eg, processor, computer, etc.). Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable programmable ROM (EPROM), floppy disks, compact disks (CD-ROMs), optical disks, hard disks, fiber optic media, radio frequency (RF) links Road and so on. A computer data signal may include any signal that can travel over a transmission medium, such as an electronic network channel, fiber optics, air, electromagnetic media, radio frequency (RF) links, bar codes, and the like. The code segments can be downloaded via a network such as the Internet, an intranet, or the like.

Although the present invention has been shown and described in detail with reference to a specific embodiment, it should be understood by those skilled in the art that various changes in form and details may be made without departing from the spirit and scope of the present invention. kind of change. These changes will all fall within the scope of protection required by the claims of the present invention.

Claims (6)

1. power business data filtering system based on rough set and gene expression is used for ensureing it is characterized in that the safety of power business data, and described power business data filtering system comprises:

Data pre-processor is used for pending all kinds of power business data are carried out the data pre-service, and described data pre-service can comprise data scrubbing and data transformation;

Data attribute yojan controller is used for yojan power business data acquisition, simplifies the power business data set;

The data filtering controller is used for responsive power business data are carried out intelligently filters, guarantees the security of power business data transmission.

2. the power business data security protecting system in according to claim 1 is characterized in that:

Data attribute yojan controller adopts rough set method yojan power business data acquisition, reduces the complexity of data.

3. the power business data security protecting system in according to claim 2 is characterized in that:

The data filtering controller utilizes the gene expression method to set up the power business data classification model, based on the sensitivity of these disaggregated model initiative recognition power business data, Cooperation Strategy knowledge base Leakage prevention.

4. power business data filtering method based on rough set and gene expression is characterized in that the step that the method comprises is:

By data pre-processor pending all kinds of power business data are carried out the data pre-service, described data pre-service can comprise data scrubbing and data transformation;

By data attribute yojan controller yojan power business data acquisition, simplify the power business data set;

By the data filtering controller responsive power business data are carried out intelligently filters, guarantee the security of power business data transmission.

5. the data filtering method in according to claim 4 is characterized in that:

Data attribute yojan controller adopts rough set method yojan power business data acquisition, reduces the complexity of data.

6. the data filtering method in according to claim 5 is characterized in that:

The data filtering controller utilizes the gene expression method to set up the power business data classification model, based on the sensitivity of these disaggregated model initiative recognition power business data, Cooperation Strategy knowledge base Leakage prevention.

Images