CN117687887B - Data security early warning system and method based on neural network - Google Patents
Data security early warning system and method based on neural network Download PDFInfo
- Publication number
- CN117687887B CN117687887B CN202410013737.0A CN202410013737A CN117687887B CN 117687887 B CN117687887 B CN 117687887B CN 202410013737 A CN202410013737 A CN 202410013737A CN 117687887 B CN117687887 B CN 117687887B
- Authority
- CN
- China
- Prior art keywords
- data
- early warning
- neural network
- existing data
- attributes
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000003491 array Methods 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 12
- 238000010276 construction Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims description 9
- 239000013598 vector Substances 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 8
- 238000003860 storage Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 2
- 238000012937 correction Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000007123 defense Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3466—Performance evaluation by tracing or monitoring
- G06F11/3476—Data logging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/50—Monitoring users, programs or devices to maintain the integrity of platforms, e.g. of processors, firmware or operating systems
- G06F21/57—Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
- G06F21/577—Assessing vulnerabilities and evaluating computer system security
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/64—Protecting data integrity, e.g. using checksums, certificates or signatures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioethics (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a data security early warning system and method based on a neural network, and belongs to the technical field of computer data monitoring. The method comprises the steps of building an early warning model and early warning; the construction step of the early warning model comprises the following steps: s10, acquiring attributes of a plurality of groups of existing data and preset quality scores corresponding to the existing data as a training set; s20, setting initial weight arrays for the attributes of the existing data respectively; s30, updating initial weight arrays corresponding to the attributes of the existing data respectively so that the actual quality scores of the existing data are close to the preset quality scores, and outputting new weight arrays to obtain an early warning model. The data security early warning method based on the neural network can monitor the real-time data and send out corresponding early warning when the quality score of the real-time data is too low, so that a series of security problems caused by the too low quality score of the real-time data are avoided.
Description
Technical Field
The invention belongs to the technical field of computer data monitoring, and particularly relates to a data security early warning system and method based on a neural network.
Background
Along with the continuous updating of computer technology, along with the continuous increase of the service life of software, data redundancy has become an increasingly larger problem, and the data redundancy can generate problems of incomplete data, inaccurate screening results, unstable data calling and the like, and seriously affects the accuracy and safety of data screening.
In specific use, with the gradual increase of the service life of software, data itself gradually generates certain defects, and the missing parts can be identified, and on the other hand, the missing data is required to be supplemented by a user, and after the data is supplemented, the data left in front can be identified, so that the data is not deleted, thus generating redundancy of the data and affecting the stability of the whole system. On the other hand, when these missing parts cannot be identified, problems occur in the final data display result when the data is called, and the user experience is affected.
Disclosure of Invention
The invention provides a data security early warning system and method based on a neural network, which can monitor real-time data and send out corresponding early warning when the quality score of the real-time data is too low so as to avoid a series of security problems caused by the too low quality score of the real-time data.
The invention is realized by the following technical scheme:
On one hand, the invention provides a data security early warning method based on a neural network, which comprises a construction step and an early warning step of an early warning model;
The construction step of the early warning model comprises the following steps: s10, acquiring attributes of a plurality of groups of existing data and preset quality scores corresponding to the existing data as a training set; s20, setting initial weight arrays for the attributes of the existing data respectively; s30, updating initial weight arrays corresponding to the attributes of the existing data respectively so that the actual quality scores of the existing data are close to the preset quality scores, and outputting new weight arrays to obtain an early warning model;
The early warning step comprises the following steps: t10, acquiring real-time data and attributes thereof; t20, obtaining quality scores of the real-time data according to the attribute of the real-time data and the new weight array corresponding to the attribute; and T30, when the real-time data quality score exceeds a preset threshold value, sending out early warning.
In some embodiments, the neural network is adopted to update the initial weight arrays corresponding to the attributes of the existing data respectively, the initial weight arrays of the attributes of the existing data and the preset quality scores are used as inputs, and the new weight arrays are used as outputs;
The output layer of the neural network is:
The hidden layer of the neural network is:
The activation function is:
Wherein i is a specific training sample; j is a specific index; k is the number of nodes included in the hidden layer; a j is a specific connection weight vector between the hidden layer and the output layer; b j concealing the layer specific output vector; c ij is the connection weight between the ith node of the input layer and the jth node of the hidden layer; x i is the specific input vector of the input layer; m is the input layer specific node.
In some of these embodiments, the attributes of the data include at least two of error values, duplicate values, data inconsistencies, data integrity, missing values, outliers.
In some embodiments, the method for acquiring the preset quality score corresponding to the existing data includes: and obtaining a preset quality score corresponding to the existing data according to the stability, redundancy and/or defensive demand proportion of the system.
In some embodiments, obtaining the preset quality score corresponding to the existing data according to the stability, redundancy and defensive demand proportion of the system includes: acquiring an application scene of a system; determining the grading duty ratio of the stability, redundancy and defensive performance of the system according to the application scene of the system; and obtaining a preset quality score corresponding to the existing data based on the score proportion of the stability, the redundancy and the defensive property.
In some of these embodiments, the application scenarios of the system include stability scenarios, conciseness scenarios, and security scenarios; the method for determining the grading proportion of the stability, the redundancy and the defenses of the system according to the application scene of the system comprises the following steps: when the system application scene is a stability scene, the stability ratio is more than 50%; when the system application scene is a succinct scene, the redundancy ratio is more than 50%; when the system application scene is a security scene, the defensive ratio is more than 50%.
In some embodiments, after determining the score ratio of stability, redundancy and defenses of the system according to the application scenario of the system, the method further comprises: and respectively storing the preset quality scores under different application scenes.
In some of these embodiments, the pre-warning includes general pre-warning and severe pre-warning.
On the other hand, the embodiment provides a data security early warning system based on a neural network, which comprises a neural network module, a storage module and a processing module, wherein the neural network module comprises an input layer, a hidden layer and an output layer; the storage module is stored with a computer program, and the processing module executes the computer program to realize the neural network-based data security early warning method in any one of the above embodiments.
Compared with the prior art, the invention has the following advantages:
According to the data security early warning method based on the neural network, an early warning model is built through the initial weight array of the existing data type and the preset quality score corresponding to the existing data, then the quality score of the real-time data is obtained according to the real-time data and the multiple attribute conditions of the real-time data based on the built early warning model, and early warning is carried out based on the quality score of the real-time data, so that the stability of a system is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following description will briefly describe the drawings in the embodiments, it being understood that the following drawings only illustrate some embodiments of the present invention and should not be considered as limiting the scope, and that other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart illustrating steps of constructing an early warning model in a data security early warning method based on a neural network according to some embodiments of the present invention;
fig. 2 is a schematic flow chart of early warning in a data security early warning method based on a neural network according to some embodiments of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention.
In the description of the present invention, it should be noted that, as the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., are used to indicate orientations or positional relationships based on those shown in the drawings, or those that are conventionally put in use in the product of the present invention, they are merely used to facilitate description of the present invention and simplify description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention.
Furthermore, the terms "horizontal," "vertical," and the like in the description of the present invention, if any, do not denote absolute levels or overhangs, but rather may be slightly inclined. As "horizontal" merely means that its direction is more horizontal than "vertical", and does not mean that the structure must be perfectly horizontal, but may be slightly inclined.
In the description of the present invention, it should also be noted that, unless explicitly stated and limited otherwise, the terms "disposed," "mounted," "connected," and "connected" should be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
The terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or modules is not limited to only those steps or modules but may include other steps or modules not expressly listed or inherent to such process, method, article, or apparatus.
The sequence of the different steps is not sequential, unless specifically stated.
Example 1:
the embodiment provides a data security early warning method based on a neural network, and particularly the method mainly comprises a construction step and an early warning step of an early warning model.
Specifically, referring to fig. 1, the construction steps of the early warning model include:
S10, acquiring attributes of a plurality of groups of existing data and preset quality scores corresponding to the existing data as a training set. In S10, several sets of existing data are acquired, each set of data including a plurality of attributes, which may be divided differently according to different requirements, and exemplary attributes may include whether there is an error value, whether there is a duplicate value, and so on. Different sets of data may have their preset quality scores based on differences between multiple attributes of the set of data. The preset quality corresponding to the existing data refers to the final total score obtained by final superposition according to the scores of all the attributes of the existing data. The preset quality scoring mode can adopt a manually set computing mode scoring mode, or other modes, such as an exemplary mode, automatic computer identification mode, other machine algorithms mode and the like, scoring the existing data, taking a group of attributes of the existing data and the preset quality scoring corresponding to the group of the existing data as a training sample, and forming a training set by a plurality of training samples.
S20, respectively setting initial weight arrays for various attributes of the existing data. In S20, the attributes of the data include a plurality of attributes, each attribute corresponds to a weight value, all the weight values form a weight array, the weight value addition of all the attributes is equal to 1, the corresponding maximum scores are assigned to the different attributes according to the weight values of the different attributes, and the final actual quality score is obtained according to the weight values of all the attributes. It should be noted that, the final actual quality score obtained by S20 according to the weight values of all the attributes is not necessarily the same as the preset quality score in step S10.
S30, updating initial weight arrays corresponding to the attributes of the existing data respectively so that the actual quality scores of the existing data are close to the preset quality scores, and outputting new weight arrays to obtain an early warning model. The actual quality score approaching to the preset quality score means that the step of the multiple updating steps is selected, the obtained actual quality score is closest to the preset quality score, and then the weight array in the step is selected as a new weight array. In a specific example, the number of different update steps may be set depending on the need for system accuracy. In S30, by using the preset quality score of the existing data, and aiming at the actual quality score of the existing data approaching to the preset quality score, changing the initial weight values corresponding to the attributes of the existing data respectively, to obtain an early warning model. In most of the current applications, most of the preset quality scores are obtained in a complex and tedious manner, and even some preset quality scores are determined manually without specific calculation schemes, so that the preset quality scores of the existing data are more difficult to explore. By setting in the step S30, the acquisition mode of the preset quality score of the existing data does not need to be explored, a new acquisition mode is adopted, the efficiency is higher, and the accuracy can meet the requirements.
Referring to fig. 2, the early warning step includes the following steps:
And T10, acquiring real-time data and attributes of the real-time data. In T10, the attribute of the real-time data is set as an array, and the attribute of the real-time data is the same as the attribute of the data used in S10.
And T20, inputting the real-time data acquired by the T10 and the attribute of the real-time data into the early warning model obtained in the step S30, and obtaining the quality score of the real-time data.
And T30, when the quality score of the real-time data obtained by the T20 exceeds a preset threshold value, sending out an early warning so as to timely send out the early warning under the condition of low data security and inform maintenance personnel to perform corresponding processing on the data.
In embodiment 1, the pre-warning model is finally obtained by taking the attribute of the existing data and the preset quality score of the data as the basis to obtain the weight array after the corresponding update of the attribute of the existing data. And then, according to the early warning model, the quality score of the real-time data can be accurately obtained, the safety degree of the real-time data is further judged, and early warning is sent out when the safety degree is low. The real-time data may be a set of data of any length.
Example 2:
Embodiment 2 is substantially similar to the solution in embodiment 1, except that:
In step S30, the neural network may be used to update the initial weight data corresponding to the attributes of the existing data. Specifically, the neural network inputs an initial weight array of the existing data attribute and a preset quality score as inputs, and outputs a new weight array of the existing data attribute.
The output layer of the neural network is as follows:
the hidden layer of the neural network is as follows:
The activation function is:
wherein i is a specific training sample; j is a specific attribute; k is the number of nodes included in the hidden layer; a j is a specific connection weight vector between the hidden layer and the output layer; b j is a specific output vector of the hidden layer; c ij is the connection weight between the ith node of the input layer and the jth node of the hidden layer; x i is the specific input vector of the input layer; m is the input layer specific node.
In embodiment 2, the number of input layer specific nodes corresponds to the number of attributes of the training samples. The following describes the updating of the weight array by the neural network, and the specific process of weight modulation of each layer is as follows:
θ1＝(d-o)o(1-o)
θ2＝θ1ajbj(1-bj)
aj(t)＝aj(t-1)+ηθ1bj+μΔaj(t-1)
cij(t)＝cij(t-1)+ηθ2xi+μΔcij(t-1)
Wherein d is the expected output of the output layer; θ 1 is an error signal obtained by comparing the expected output with the actual output of the output layer; θ 2 is the error signal of the hidden layer; η ε (0, 1), is the learning rate; mu E (0, 1) is the dynamic term. Based on the training process, the weight value after one attribute is updated can be obtained. And repeating the steps on the next attribute to obtain a weight value after corresponding updating until the weight values of all the attributes of the existing data are obtained, so as to obtain a new weight array.
Example 3:
example 3 is substantially similar to example 1 except that:
the attributes of the data include at least two of error value, duplicate value, data inconsistency, data integrity, missing value, outlier.
In embodiment 3, in S20, obtaining the corresponding score according to the weight of the attribute of the existing data may include the steps of respectively obtaining the weights corresponding to the error value, the repetition value, the data inconsistency, the data integrity, the missing value and the abnormal value of the data, obtaining the score including the proportion under the weight according to the actual condition of one attribute, taking the error value of the data as an example, the error value of the data may be a group of bytes in error in the data, taking the total score of 100 as an example, the error value is 20 at most, and when the error value of the data reaches a certain degree, that is, when a certain number of bytes exist, obtaining the score of a certain proportion in the 20.
In a specific example, the final score of any attribute of the data can be determined according to the degree of achievement of the attribute, and in a specific example, the total score of the quality score is still 100 points, the error value is maximally divided into 20 points, when an error exists in the error value, a certain score can be deducted until the position is completely deducted from the 20 points. Duplicate values, data inconsistencies, data integrity, missing values, and outliers may also be scored in this manner.
In other examples, the final score for any attribute of the data may also be determined based on whether the attribute is satisfied, and in particular examples, the total score is still 100 points, the error value is up to 20 points, when the error value exists in the data, the 20 points are deducted, and if the error value does not exist, the total 20 points are obtained. Duplicate values, data inconsistencies, data integrity, missing values, and outliers may also be scored in this manner.
Example 4:
example 4 is substantially similar to example 1 except that:
In S10, the method for obtaining the preset quality score corresponding to the existing data may include the following steps:
And obtaining a preset quality score corresponding to the existing data according to the stability, redundancy and/or defensive demand proportion of the system. When the preset quality score is carried out on the existing data, different preset quality scores of the existing data can be obtained according to the stability, redundancy and/or defensive emphasis of the system, and then a weight array of data attributes with different emphasis can be obtained according to the difference of the preset quality scores.
In embodiment 4, the preset quality scores of the system will be different according to the different requirements of the system, and for example, when the system pursues stability, the error value attribute of the data, the attribute of the data inconsistency, and the attribute of the data integrity are more important. In a specific example, the algorithm for manually determining the preset quality score or using the existing computer model may be different according to the requirements of the system.
Example 5:
example 5 is substantially similar to example 4 except that:
And obtaining a preset quality score corresponding to the existing data according to the stability, redundancy and defensive demand proportion of the system. Specifically, the method comprises the following steps:
And K10, acquiring an application scene of the system. In K10, multiple application scenarios may be derived according to different requirements of the system, and an exemplary application scenario may include a stability scenario, where the scenario tends to have better stability of the system, and the requirement for security is relatively low, so as to achieve the purpose of stably using the system, such as industrial production. The application scenario may also include a brevity scenario, where the brevity scenario mainly favors the operation speed of the system, and the use of the server database, for example, requires that the system have better brevity in a system with relatively low security and certain stability, so as to reduce unnecessary redundancy, so as to increase the operation speed of the system. The application scenarios may also include defensive scenarios that tend to give the system better security, such as application scenarios with certain privacy requirements.
K20, determining the grading proportion of the stability, redundancy and defensive performance of the system according to different application scenes of the system.
And K30, obtaining a preset quality score corresponding to the existing data.
In embodiment 5, different preset quality scores are correspondingly obtained according to different requirements of the system, so that a weight array of a subsequent real-time data type is changed, further, the subsequent quality scores can be obtained more accurately according to the real-time data, and finally, a better early warning effect is obtained.
Example 6:
example 6 is substantially similar to example 5 except that:
the application scene of the system is mainly divided into a stability scene, a conciseness scene and a safety scene.
K20 may specifically comprise the steps of:
when the system application scene is a stability scene, the stability ratio of the system is more than 50%, and the sum of the redundancy and defensive performance of the system is less than 50%.
When the application scene of the system is a succinct scene, the redundancy ratio of the system is more than 50%, and the sum of the stability and defensive performance of the system is less than 50%;
when the application scene of the system is a security scene, the defensive ratio of the system is more than 50%, and the sum of the stability and the redundancy of the system is less than 50%.
Example 7:
Example 7 is substantially similar to example 6 except that:
k20 further comprises the following steps:
and K40, respectively storing the preset quality scores under different application scenes.
In embodiment 7, the preset quality scores under different application scenes are stored, and the corresponding preset quality scores are selected according to the specific application scene of the system to obtain the weight array of the corresponding data type, so that a system with wider application can be formed to adapt to the requirements of different users on the system.
Example 8:
example 8 is substantially similar to example 1 except that:
the early warning includes general early warning and serious early warning.
In embodiment 8, by setting different preset thresholds a and b, where a > b, the system is normal when the real-time data quality score obtained according to the real-time data type is equal to or greater than a. When the real-time data quality score is more than or equal to b and less than a, the system is generally early-warning, and the system can be corrected according to the redundancy of the system at the moment without real-time correction, and centralized processing is performed after routine correction is waited. When the real-time data quality score is less than b, serious damage may occur to the system, and timely correction is needed. Through the arrangement, the corresponding correction scheme can be obtained based on early warning, and maintenance work of maintenance personnel is reduced while the stability of the system is ensured.
Example 9:
The embodiment provides a data security early warning system based on a neural network, which comprises a neural network module, a storage module and a processing module, wherein the neural network module comprises an input layer, a hidden layer and an output layer; the storage module stores a computer program thereon, and the processing module executes the computer program to implement the neural network-based data security early warning method of any one of embodiments 1 to 8.
Example 10:
The present embodiment provides a computer storage medium having a computer program stored thereon, the computer program being loaded by a processing module to implement the neural network-based data security early warning method of any one of embodiments 1 to 8.
The foregoing description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and any simple modification, equivalent variation, etc. of the above embodiment according to the technical matter of the present invention fall within the scope of the present invention.
Claims (4)
1. The data security early warning method based on the neural network is characterized by comprising a construction step and an early warning step of an early warning model;
the construction step of the early warning model comprises the following steps:
s10, acquiring attributes of a plurality of groups of existing data and preset quality scores corresponding to the existing data as a training set;
S20, setting initial weight arrays for the attributes of the existing data respectively;
S30, updating initial weight arrays corresponding to the attributes of the existing data respectively so that the actual quality scores of the existing data are close to preset quality scores, and outputting a new weight array to obtain an early warning model;
the early warning step comprises the following steps:
T10, acquiring real-time data and attributes thereof;
t20, obtaining quality scores of the real-time data according to the attribute of the real-time data and the new weight array corresponding to the attribute;
t30, when the real-time data quality score exceeds a preset threshold value, sending out early warning;
Updating initial weight arrays respectively corresponding to the attributes of the existing data by using a neural network, taking the initial weight arrays and preset quality scores of the attributes of the existing data as inputs, and taking new weight arrays as outputs;
The output layer of the neural network is as follows:
the hidden layer of the neural network is as follows:
The activation function is:
Wherein, Is a specific training sample; Is a specific index; the number of nodes included for the hidden layer; A specific connection weight vector between the hidden layer and the output layer; hiding a layer specific output vector; To input layer no From the personal node to the hidden layerThe connection weight among the individual nodes; specific input vectors are input layers; is a specific node of the input layer;
the neural network updates the weight array, and the specific process of weight modulation of each layer is as follows:
Wherein, Expects an output for the output layer; the error signal obtained by comparing the actual output with the expected output of the output layer; an error signal for the hidden layer; Is the learning rate; Is a dynamic term;
the attribute of the data comprises at least two of an error value, a repetition value, data inconsistency, data integrity, a missing value and an abnormal value;
The method for acquiring the preset quality score corresponding to the existing data comprises the following steps: obtaining a preset quality score corresponding to the existing data according to the stability, redundancy and/or defensive demand proportion of the system;
According to the stability, redundancy and defensive demand proportion of the system, obtaining the preset quality score corresponding to the existing data comprises the following steps:
Acquiring an application scene of a system;
Determining the grading duty ratio of the stability, redundancy and defensive performance of the system according to the application scene of the system;
Obtaining a preset quality score corresponding to the existing data based on the score proportion of the stability, the redundancy and the defensive property;
The application scenes of the system comprise a stability scene, a conciseness scene and a safety scene;
The determining the grading ratio of the stability, the redundancy and the defensive performance of the system according to the application scene of the system comprises the following steps:
when the system application scene is a stability scene, the stability ratio is more than 50%;
when the system application scene is a succinct scene, the redundancy ratio is more than 50%;
when the system application scene is a security scene, the defensive ratio is more than 50%.
2. The neural network-based data security pre-warning method according to claim 1, further comprising, after determining the score ratio of stability, redundancy and defensive performance of the system according to the application scenario of the system:
and respectively storing the preset quality scores under different application scenes.
3. The neural network-based data security pre-warning method of claim 1, wherein the pre-warning includes general pre-warning and serious pre-warning.
4. The data security early warning system based on the neural network is characterized by comprising a neural network module, a storage module and a processing module, wherein the neural network module comprises an input layer, a hidden layer and an output layer; the storage module stores a computer program, and the processing module executes the computer program to implement the neural network-based data security early warning method according to any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410013737.0A CN117687887B (en) | 2024-01-04 | 2024-01-04 | Data security early warning system and method based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410013737.0A CN117687887B (en) | 2024-01-04 | 2024-01-04 | Data security early warning system and method based on neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117687887A CN117687887A (en) | 2024-03-12 |
| CN117687887B true CN117687887B (en) | 2024-08-23 |
Family
ID=90126561
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410013737.0A Active CN117687887B (en) | 2024-01-04 | 2024-01-04 | Data security early warning system and method based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117687887B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222753A (en) * | 2019-12-17 | 2020-06-02 | 合肥工业大学 | An e-government performance evaluation system and method |
| CN115543975A (en) * | 2022-09-22 | 2022-12-30 | 清华大学 | A data set multi-scale evaluation method, system, device and storage medium |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10373047B2 (en) * | 2014-02-28 | 2019-08-06 | Educational Testing Service | Deep convolutional neural networks for automated scoring of constructed responses |
| CN107506941A (en) * | 2017-09-09 | 2017-12-22 | 杭州数立方征信有限公司 | A kind of enterprise in charge of construction's credit assessment method and system based on big data technology |
| CN109472345A (en) * | 2018-09-28 | 2019-03-15 | 深圳百诺名医汇网络技术有限公司 | A kind of weight update method, device, computer equipment and storage medium |
| CN112037011A (en) * | 2020-08-14 | 2020-12-04 | 百维金科(上海)信息科技有限公司 | Credit scoring method based on FOA-RBF neural network |
| CN112435651B (en) * | 2020-11-20 | 2023-05-02 | 昆明学院 | Quality assessment method for automatic annotation of voice data |
| CN114091842A (en) * | 2021-10-29 | 2022-02-25 | 上海聚音信息科技有限公司 | Commodity data quality evaluation method, commodity data replenishment method, commodity data quality evaluation apparatus, and storage medium |
-
2024
- 2024-01-04 CN CN202410013737.0A patent/CN117687887B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222753A (en) * | 2019-12-17 | 2020-06-02 | 合肥工业大学 | An e-government performance evaluation system and method |
| CN115543975A (en) * | 2022-09-22 | 2022-12-30 | 清华大学 | A data set multi-scale evaluation method, system, device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117687887A (en) | 2024-03-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108776683A (en) | A kind of electric power operation/maintenance data cleaning method based on isolated forest algorithm and neural network | |
| CN114707686B (en) | A joint optimization method for multi-process mixed production | |
| CN118246837B (en) | Intelligent management method and system in digital logistics transportation process | |
| CN114511112A (en) | Intelligent operation and maintenance method and system based on Internet of things and readable storage medium | |
| CN113459867A (en) | Electric vehicle charging process fault early warning method based on adaptive deep confidence network | |
| CN116227786A (en) | Unmanned aerial vehicle comprehensive efficiency evaluation system | |
| CN112199805B (en) | Power transmission line hidden danger identification model evaluation method and device | |
| CN117687887B (en) | Data security early warning system and method based on neural network | |
| CN117236809A (en) | Automobile part production management method and system based on artificial intelligence | |
| CN116483602A (en) | Abnormality detection method, abnormality detection device and computer storage medium | |
| CN120143855A (en) | A short vertical aircraft multi-mode intelligent propulsion control method, device, equipment and storage medium | |
| CN114743060B (en) | A method and system for predicting rare earth element component content | |
| CN114742172B (en) | A steel defect tracing method based on Bayesian network | |
| CN115758265A (en) | Complex electromechanical equipment fault prediction method, electronic equipment and storage medium | |
| CN118694615B (en) | A network intrusion detection method and system based on big data | |
| JPH0481988A (en) | Clustering system | |
| CN109308496A (en) | An optimal solution selection method based on fuzzy clustering iteration and projection pursuit | |
| CN110414667B (en) | Training method and device for network detection model, computer equipment and storage medium | |
| CN112766410A (en) | Rotary kiln firing state identification method based on graph neural network feature fusion | |
| CN113988185B (en) | Data processing method and related device | |
| CN110457906B (en) | Intelligent alarm method for network security event | |
| CN116632859A (en) | Multi-type flexible resource cluster parameter identification method for power distribution network | |
| CN111860630B (en) | Model building method and system based on feature importance | |
| CN115700629A (en) | Equipment energy consumption prediction method, device, electronic equipment and storage medium | |
| CN113419496A (en) | Waterproof material production intelligent management and control method and system based on big data statistics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |