CN108076060A - Neutral net Tendency Prediction method based on dynamic k-means clusters - Google Patents
Neutral net Tendency Prediction method based on dynamic k-means clusters Download PDFInfo
- Publication number
- CN108076060A CN108076060A CN201711361681.4A CN201711361681A CN108076060A CN 108076060 A CN108076060 A CN 108076060A CN 201711361681 A CN201711361681 A CN 201711361681A CN 108076060 A CN108076060 A CN 108076060A
- Authority
- CN
- China
- Prior art keywords
- data
- network
- prediction
- rbf neural
- situation
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/20—Network architectures or network communication protocols for network security for managing network security; network security policies in general
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1441—Countermeasures against malicious traffic
- H04L63/1483—Countermeasures against malicious traffic service impersonation, e.g. phishing, pharming or web spoofing
Landscapes
- Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computing Systems (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Biomedical Technology (AREA)
Abstract
Based on the neutral net Tendency Prediction method of dynamic k means clusters, comprise the following steps:1)Certain system network safety basic data is collected, data target infects internet worm host number, is tampered network number, it is implanted the number networks at back door, security incident number of reports, counterfeit page quantity, and security breaches and high-risk loophole quantity, to the processing of basic network security data normalization;2)To the basic network security data after normalization, clustered using dynamic k means clustering algorithms, determine RBF neural central nervous member parameter and number N;3)RBF neural is participated in using the data after normalization to train, and calculates RBF neurons width and determines neuron output;4)In the training process, the output weights of RBF neural are encoded, best initial weights is obtained using PSO algorithms, improve Network Situation precision of prediction;5)Certain month Network Situation is predicted using the RBF neural trained, and is compared with of that month Network Situation assessed value, calculation error;Have the characteristics that precision of prediction is high.
Description
Technical field
The invention belongs to Network Situation electric powder predictions, and in particular to the neutral net based on dynamic k-means clusters
Tendency Prediction method.
Background technology
Increasingly complicated network environment and unpredictable cyberspace vulnerability so that network management work becomes abnormal difficult
It is pungent.In recent years, the rise of Network Situation Predicting Technique (e.g., is leaked by being collected processing to system bottom security element information
The host number etc. of hole information, virus infection), provide facility for network management work.Network Situation Predicting Technique not only can be with
Data analysis is provided to the security postures of future network, network manager is may also help in and makes administrative decision in advance, carried with this
The network resilience of system where high reduces the probability that unsafe incidents occur.Therefore, Network Situation Predicting Technique becomes
Nowadays there is an urgent need for the problems of research.
Bass et al. proposed network situation awareness this concept earliest in 2000, summarised the base of network situation awareness
Plinth concept, target and correlation properties[1].But how to obtain, understand for network safety situation information, applying and to future
The prediction of trend of network development does not specify simultaneously.
Srihari R propose a kind of network situation awareness method based on concept.It is extracted by the element to concept,
Situation awareness information is obtained with this.This method has the extraction of element good effect, and shortcoming is to be only capable of resisting singly invasion to attack
It hits, and data source is single, therefore Tendency Prediction is not furtherd investigate.
Stephen L propose the visualization technique of the network traffic information based on three dimensions.It is sat using X, Y, Z are three-dimensional
It marks to portray network address, source IP and port numbers, and designs a system based on rotation cubic structure, there is preferable state
Gesture evaluation capacity.But this method lays particular emphasis on the visual research of current network situation, is not directed to for Network Situation prediction.
Liu Z etc. combine existing elements recognition technology, and data are merged using the method for data mining, so as to
Network Situation is only assessed and predicted.The use of data mining technology so that Network Situation is assessed and the result of prediction is more smart
Really, but simultaneously because redundant data is excessive, computation complexity is excessive lead to problems such as dimension explosion and algorithm operation slow.
The content of the invention
To overcome above-mentioned the deficiencies in the prior art, the object of the present invention is to provide the nerves clustered based on dynamic k-means
It is more accurate to compare traditional neutral net Tendency Prediction method for Network Situation Forecasting Methodology.
To achieve the above object, the technical solution adopted by the present invention is:Neutral net based on dynamic k-means clusters
Tendency Prediction method, comprises the following steps:
Step 1, the basic network security data in certain system half a year are collected, data target infects internet worm host number
Amount, is tampered network number, is implanted the number networks at back door, security incident number of reports, counterfeit page quantity and security breaches
With high-risk loophole quantity, by basic network security data value specification in the range of [0,1], to basic network security data normalizing
Change is handled,;
Step 2, to the basic network security data X=[x after normalization1,x2,...,xn]T, gathered using dynamic k-means
Class algorithm is clustered, and determines RBF neural central nervous member parameter Cji=[Cj1,Cj2,...,Cjn]TWith number N;
Step 3, participate in RBF neural using the data after normalization to train, calculate RBF neuron width Dsj=
[dj1,dj2,...,djn] and determine that neuron exports zj;
Step 4, in the training process, coding W=[w are carried out to the output weights of RBF neural1,w2,...,wk]T,
(k=1,2 ..., n), best initial weights are obtained using PSO algorithms, improve Network Situation precision of prediction;
Step 5, certain month Network Situation is predicted using the RBF neural trained, and with of that month Network Situation
Assessed value compares, according to following formula calculation errors,
MSE can be with the variation degree of evaluating data, and the value of MSE is smaller, illustrates that prediction model describes experimental data and has more
Good accuracy, y in formulaiRepresent Situation Assessment value,Represent Tendency Prediction value.
The beneficial effects of the invention are as follows:
Compared with prior art, the present invention is had found by the predicted value to algorithm and prediction error analysis, dynamic k-means
Application of the clustering algorithm combination population optimization algorithm in neutral net, it is apparent for having on Network Situation precision of prediction
It improves.Therefore, the present invention has application prospect.
The application proposes on the basis of RBF (radial basis function) neutral net to be combined based on dynamic k-means algorithms
The Network Situation Forecasting Methodology of population optimization algorithm (PSO), dynamic k-means clustering algorithms overcome artificial definite initial poly-
The defects of class center, the method that cluster centre is adjusted using dynamic allow mutual distance between initial cluster centre as far as possible
Far, the network center's selection for making RBF is more accurate.Population optimization algorithm (PSO) purpose is most suitable RBF is selected to weigh
Value.
It builds network environment to be emulated, analysis finds to predict Network Situation compared to traditional RBF neural, institute
Algorithm is carried to increase on precision of prediction.Therefore, the algorithm according to the present invention that carries has centainly in Network Situation prediction
Feasibility and validity.
Description of the drawings
Fig. 1 (a) CNVD count the infection of the first half of the year in 2017 internet worm host number figure.
Fig. 1 (b) CNVD count first half of the year cyberspace vulnerability in 2017, high-risk loophole number figure.
Fig. 2 is population optimization algorithm (PSO) flow chart.
Fig. 3 is the structure diagram of RBF neural.
Fig. 4 is experimental situation topology diagram of the present invention.
Fig. 5 inventive algorithms and traditional algorithm predicted value comparison diagram.
Fig. 6 predicts error comparison diagram for inventive algorithm and traditional algorithm.
Specific embodiment
The structural principle and operation principle of the present invention are described in further detail with reference to the accompanying drawings and examples.
The experimental situation that the present embodiment is built is the LAN built of laboratory, including host is several, router, fire wall
And interchanger.Host is configured to Intel (R) Core (TM) i7-4790 CPU 3.60GHz, 8.00GB memories, 64 bit manipulations
System.
Ordinary user and attacker may have access to the host in the LAN.Several hosts in selection LAN are attacked
It hits, attack means are as follows:
1. inject CVE-2016-9732, CVE-2016-2979, CVE-2016-2973 loophole carries out cross site scripting to host
Attack.
2. inject CVE-2016-2299, CVE-2016-4040, CVE-2016-3172 loophole carries out SQL injection to host
Attack.
3. injecting CVE-2017-10804 loopholes obtains host subscriber's privacy information.
Analysis is detected by using burpsuit and WireShark software local area networks, is obtained in LAN 1 year
When under fire infect internet worm host number, be tampered network number, be implanted back door the number networks and security breaches and
The basic data of high-risk loophole quantity, as Research foundation.Experimental situation topological structure is illustrated in figure 4, since space is limited
Topological structure only display portion host.Table 1 is collects the basic network security data built in LAN 1 year monthly.
1 basic network security data sample of table
Based on the neutral net Tendency Prediction method of dynamic k-means clusters, comprise the following steps:
Step 1, the basic network security data for building half a year in interior LAN in laboratory are collected, data target is felt
Internet worm host number is contaminated, is tampered network number, is implanted the number networks at back door and security breaches and high-risk loophole number
Amount, and basic network security data are normalized;Since each index feature parameter usually has different dimension and physics
Meaning.If data above directly participates in Network Situation prediction computing, the different data of physical unit can cause uncertain
Error causes Tendency Prediction to fail, therefore basic network security data normalization need to be handled, and basic network security data are taken
It is worth specification in the range of [0,1];
Step 2, to the basic network security data X=[x after normalization1,x2,...,xn]T, gathered using dynamic k-means
Class algorithm is clustered, and determines RBF neural central nervous member parameter Cji=[Cj1,Cj2,...,Cjn]TWith number N;
Step 3, participate in RBF neural using the data after normalization to train, calculate RBF neuron width Dsj=
[dj1,dj2,...,djn] and determine that neuron exports zj;
Step 4, in the training process, coding W=[w are carried out to the output weights of RBF neural1,w2,...,wk]T,
(k=1,2 ..., n), best initial weights are obtained using PSO algorithms, improve Network Situation precision of prediction;
Step 5, certain month Network Situation is predicted using the RBF neural trained, and with of that month Network Situation
Assessed value compares, and in order to ensure the accuracy of assessed value, with reference to CVSS standards and randomly chooses 10 Internet security experts couple
Each index carries out assessment marking and takes its average value, obtains of that month Network Situation assessed value, calculation error;Table 2 show CVSS marks
Network Situation rank under accurate, i.e., the situation value in any time section correspond to a risk class, are provided more for network manager
For intuitively Network Situation situation.
The substandard Network Situation rank tables of table 2CVSS
The infection of the first half of the year in 2017 internet worm host number figure is counted for CNVD referring to Fig. 1 (a), Fig. 1 (b) is respectively CNVD
Count first half of the year cyberspace vulnerability in 2017, high-risk loophole number figure.
National information security breaches shared platform (China National Vulnerable Database, abbreviation CNVD)
It is by important information system unit, basic telecommunication operation in the United Nations of national computer network emergence technology processing Consultation Center
The information security vulnerability information that business, network security manufacturer etc. establish jointly shares knowledge base.
CNVD refers to CVSS (general loophole points-scoring system) evaluation criterion for the setting of network safety grade.With reference to net
The safety status classification of network safety situation is 5 grades by the elements features such as network threat, loophole, is respectively safety, slight dangerous,
It is general dangerous, poor risk, highly dangerous, and carry out equivalent description with the numerical value in [0,1] section.It is marked as shown in table 1 for CVSS
Network Situation rank under accurate[7]。
The basic information that CNVD is mainly collected includes infection internet worm host number, network number is tampered, after being implanted
The number networks of door, counterfeit page quantity, security incident number of reports and security breaches and high-risk loophole quantity.It emulates herein
The basic data of system where experiment is collected is above seven classes.
Since each index feature parameter usually has different dimension and physical significance.If data above directly participates in net
Network Tendency Prediction computing, the different data of physical unit can cause uncertain error, Tendency Prediction caused to fail.Therefore root
According to actual needs, Situation Assessment index is normalized herein, the interference of Data Physical unit is removed, by data value
Specification is in the range of [0,1][8].Index initialization formula is as follows:
Wherein:xiRepresent data value, xmaxRepresent numerical upper limits, xminRepresent numerical lower limits.
The basic information of half a year in the LAN (referring to Fig. 4) built by collecting the present embodiment uses this calculation with this
PSO-RBF neutral nets in method are trained, and the situation of final local area network is predicted.
Experimental analysis is to be compared using traditional RBF neural prediction algorithm with this algorithm.Comparison includes prediction
Value, prediction error.Wherein prediction error is evaluated using mean square error (MSE).Table 3 is predicted for this algorithm and traditional algorithm
Output and actual value comparison.Fig. 5 and Fig. 6 is respectively this algorithm and traditional algorithm predicted value comparison diagram and prediction error comparison diagram.
Mean square error refers to the desired value of the difference square of estimates of parameters and parameter true value.MSE can be with the change of evaluating data
Change degree, the value of MSE is smaller, illustrates that prediction model describes experimental data with better accuracy.Expression formula is as follows:
Y in formulaiRepresent situation actual value,Represent Tendency Prediction value.
3 algorithms of table are compared with traditional algorithm prediction output and actual value
Find that dynamic k-means clustering algorithm combinations population is most by the predicted value to algorithm and prediction error analysis
Application of the optimization algorithm in neutral net, for being significantly improved on Network Situation precision of prediction.Therefore, it is contemplated that this hair
Bright algorithm has certain application prospect.
Referring to Fig. 2, Fig. 2 is population optimization algorithm (PSO) flow chart.
Population optimization algorithm
Population optimization algorithm (PSO) belongs to one kind of swarm intelligence algorithm, it is set by simulating flock of birds predation
Meter.
Particle cluster algorithm simulates the bird in flock of birds by designing a kind of particle, and particle only has two kind of speed and position
Property.Each particle individually searches optimal solution in search space, obtains current individual extreme value Pbest, and by this extreme value and entire grain
Other particles in subgroup are shared, and find current globally optimal solution G of the optimal individual extreme value as entire populationbest.Grain
All particles in subgroup are according to PbestAnd GbestAdjust itself speed and position.Algorithm steps and expression formula are as follows[12-14]:
1. population initializes
It needs to set maximum speed section, location information is entire search space, random initializtion speed and position.If
Put population size m.
2. individual extreme value and globally optimal solution
Individual extreme value is the location information optimal in history that each particle is found, and from these individual history optimal solutions
A globally optimal solution is found, and compared with history optimal solution, selected optimal as current history optimal solution.
3. renewal speed and location formula
A. speed more new formula:
Vid=ω Vid+C1random(0,1)(Pid-Xid)+C2random(0,1)(Pgd-Xid)(2-3)
Wherein, ω is known as inertial factor, C1And C2Referred to as aceleration pulse generally takes C1=C2∈ [0,4]。random(0,
1) random number of the section on [0,1] is represented.PidRepresent the individual extreme value of i-th of variable in d dimensions.PgdRepresent complete in d dimensions
Office's optimal solution.
B. location update formula:
Xid=Xid+Vid (2-4)
4. end condition
End condition is divided into two classes, first, greatest iteration number:Gmax;Second is that the error between adjacent generations is specified at one
In the range of stop.Used in this application is second of end condition.
Referring to Fig. 3, Fig. 3 is the structure diagram of RBF neural, wherein, in order to be suitable for Tendency Prediction, the present invention makes
RBF neural contains m input node, n concealed nodes and l output node, and total regression function is Gaussian function
Number.The expression formula and computational methods of parameters are as follows[16]:
1. determine input vector X:
X=[x1,x2,...,xn]T, n is input layer unit number.From first segment, input vector number is 7.
2. hidden layer is initialized to the connection weight of output layer:
W=[w1,w2,...,wk]T, (k=1,2 ..., n), wherein k is hidden layer unit number.The initialization of connection weight
And selection is determined by population optimization algorithm (PSO).
3. determine the neuronal center parameter of hidden layer:
Cji=[Cj1,Cj2,...,Cjn]T, n is neuronal center number.It the center of different hidden layer neurons should be different
Value, and can be adjusted with the corresponding width at center so that different input information characteristics can be by different hidden layer god
It is reflected through first maximum.Herein neutral net center C is determined using dynamic k-means clustering algorithmsji。
4. initialize width vector Dj=[dj1,dj2,...,djn] (n is neuronal center number), definition is as follows:
Wherein, dfFor width adjusting coefficient, value is less than 1, and effect is that each neuron is easier to realize to local letter
The feeling ability of breath is conducive to improve local acknowledgement's ability of RBF neural.
Width vector affects sphere of action of the neuron to input information:Width is smaller, and corresponding hidden layer neuron is made
Narrower with the shape of function, then response of the information near other neuronal centers at the neuron is with regard to smaller
5. calculate the output valve z of j-th of neuron of hidden layerj, definition is as follows:
Wherein CjIt is the center vector of j-th of neuron of hidden layer, DjIt is vectorial for j-th of neuron width of hidden layer, with Cj
It is corresponding, | | | | it is European norm.
6. the output of output layer neuron:
Wherein, k is the number of weight coefficient.
In addition, RBF networks can by fitness function control optimal solution, optimization aim be network desired output Y' and
The error function MSE of reality output Y is minimum, then error function E is:
Claims (1)
1. the neutral net Tendency Prediction method based on dynamic k-means clusters, which is characterized in that comprise the following steps:
Step 1, the basic network security data in certain system half a year are collected, data target infects internet worm host number,
Be tampered network number, be implanted the number networks at back door, security incident number of reports, counterfeit page quantity and security breaches and
High-risk loophole quantity, by basic network security data value specification in the range of [0,1], to basic network security data normalization
Processing,;
Step 2, to the basic network security data X=[x after normalization1,x2,...,xn]T, clustered and calculated using dynamic k-means
Method is clustered, and determines RBF neural central nervous member parameter Cji=[Cj1,Cj2,...,Cjn]TWith number N;
Step 3, participate in RBF neural using the data after normalization to train, calculate RBF neuron width Dsj=[dj1,
dj2,...,djn] and determine that neuron exports zj;
Step 4, in the training process, coding W=[w are carried out to the output weights of RBF neural1,w2,...,wk]T, (k=1,
2 ..., n), best initial weights are obtained using PSO algorithms, improve Network Situation precision of prediction;
Step 5, certain month Network Situation is predicted using the RBF neural trained, and is assessed with of that month Network Situation
Value compares, according to following formula calculation errors,
<mrow>
<mi>M</mi>
<mi>S</mi>
<mi>E</mi>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mover>
<mi>y</mi>
<mo>^</mo>
</mover>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
MSE can be with the variation degree of evaluating data, and the value of MSE is smaller, illustrates that prediction model describes experimental data and has preferably
Accuracy, y in formulaiRepresent Situation Assessment value,Represent Tendency Prediction value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711361681.4A CN108076060B (en) | 2017-12-18 | 2017-12-18 | Neural network situation prediction method based on dynamic k-means clustering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711361681.4A CN108076060B (en) | 2017-12-18 | 2017-12-18 | Neural network situation prediction method based on dynamic k-means clustering |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108076060A true CN108076060A (en) | 2018-05-25 |
CN108076060B CN108076060B (en) | 2020-12-08 |
Family
ID=62158864
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711361681.4A Active CN108076060B (en) | 2017-12-18 | 2017-12-18 | Neural network situation prediction method based on dynamic k-means clustering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108076060B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109856530A (en) * | 2018-12-25 | 2019-06-07 | 国网江苏省电力有限公司南京供电分公司 | A kind of load ratio bridging switch on-line monitoring method for diagnosing faults |
CN110276476A (en) * | 2019-05-27 | 2019-09-24 | 南京信息工程大学 | A kind of automobile engine smoke intensity prediction technique based on PSO-RBF neural network |
CN110796237A (en) * | 2019-10-28 | 2020-02-14 | 宁夏吉虎科技有限公司 | Method and device for detecting attack resistance of deep neural network |
CN110855467A (en) * | 2019-08-19 | 2020-02-28 | 中国电子科技集团公司第三十研究所 | Network comprehensive situation prediction method based on computer vision technology |
CN111242291A (en) * | 2020-04-24 | 2020-06-05 | 支付宝(杭州)信息技术有限公司 | Neural network backdoor attack detection method and device and electronic equipment |
CN111474490A (en) * | 2020-04-09 | 2020-07-31 | 北方工业大学 | Rapid screening method for batteries used in echelon |
CN112291098A (en) * | 2020-10-30 | 2021-01-29 | 北京源堡科技有限公司 | Network security risk prediction method and related device thereof |
CN113364742A (en) * | 2021-05-17 | 2021-09-07 | 北京邮电大学 | Quantitative elastic calculation method and device for network security threat |
CN117254981A (en) * | 2023-11-17 | 2023-12-19 | 长扬科技(北京)股份有限公司 | Industrial control network security situation prediction method and device |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1367534A3 (en) * | 2002-05-30 | 2008-07-02 | L'oreal | Method involving artificial intelligence |
CN106339357A (en) * | 2016-08-12 | 2017-01-18 | 北京同余科技有限公司 | Optimizing parameterized network information security detection system and method |
-
2017
- 2017-12-18 CN CN201711361681.4A patent/CN108076060B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1367534A3 (en) * | 2002-05-30 | 2008-07-02 | L'oreal | Method involving artificial intelligence |
CN106339357A (en) * | 2016-08-12 | 2017-01-18 | 北京同余科技有限公司 | Optimizing parameterized network information security detection system and method |
Non-Patent Citations (2)
Title |
---|
刘建波: "基于动态聚类算法的IRC 僵尸网络检测", 《哈尔滨商业大学学报( 自然科学版)》 * |
王亚等: "基于混沌PSO算法优化RBF网络入侵检测模型", 《计算机工程与应用》 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109856530B (en) * | 2018-12-25 | 2021-11-02 | 国网江苏省电力有限公司南京供电分公司 | On-load tap-changer on-line monitoring fault diagnosis method |
CN109856530A (en) * | 2018-12-25 | 2019-06-07 | 国网江苏省电力有限公司南京供电分公司 | A kind of load ratio bridging switch on-line monitoring method for diagnosing faults |
CN110276476A (en) * | 2019-05-27 | 2019-09-24 | 南京信息工程大学 | A kind of automobile engine smoke intensity prediction technique based on PSO-RBF neural network |
CN110855467A (en) * | 2019-08-19 | 2020-02-28 | 中国电子科技集团公司第三十研究所 | Network comprehensive situation prediction method based on computer vision technology |
CN110855467B (en) * | 2019-08-19 | 2022-02-08 | 中国电子科技集团公司第三十研究所 | Network comprehensive situation prediction method based on computer vision technology |
CN110796237B (en) * | 2019-10-28 | 2023-04-07 | 宁夏吉虎科技有限公司 | Method and device for detecting attack resistance of deep neural network |
CN110796237A (en) * | 2019-10-28 | 2020-02-14 | 宁夏吉虎科技有限公司 | Method and device for detecting attack resistance of deep neural network |
CN111474490A (en) * | 2020-04-09 | 2020-07-31 | 北方工业大学 | Rapid screening method for batteries used in echelon |
CN111474490B (en) * | 2020-04-09 | 2022-06-24 | 北方工业大学 | Rapid screening method for batteries used in echelon |
CN111242291A (en) * | 2020-04-24 | 2020-06-05 | 支付宝(杭州)信息技术有限公司 | Neural network backdoor attack detection method and device and electronic equipment |
CN112291098A (en) * | 2020-10-30 | 2021-01-29 | 北京源堡科技有限公司 | Network security risk prediction method and related device thereof |
CN112291098B (en) * | 2020-10-30 | 2023-05-02 | 北京源堡科技有限公司 | Network security risk prediction method and related device thereof |
CN113364742A (en) * | 2021-05-17 | 2021-09-07 | 北京邮电大学 | Quantitative elastic calculation method and device for network security threat |
CN113364742B (en) * | 2021-05-17 | 2022-10-11 | 北京邮电大学 | Quantitative elastic calculation method and device for network security threat |
CN117254981A (en) * | 2023-11-17 | 2023-12-19 | 长扬科技(北京)股份有限公司 | Industrial control network security situation prediction method and device |
CN117254981B (en) * | 2023-11-17 | 2024-02-02 | 长扬科技(北京)股份有限公司 | Industrial control network security situation prediction method and device |
Also Published As
Publication number | Publication date |
---|---|
CN108076060B (en) | 2020-12-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108076060A (en) | Neutral net Tendency Prediction method based on dynamic k-means clusters | |
Ullah et al. | Design and development of RNN anomaly detection model for IoT networks | |
CN106453293B (en) | A kind of network security situation prediction method based on improved BPNN | |
CN108718310A (en) | Multi-level attack signatures generation based on deep learning and malicious act recognition methods | |
Qu et al. | An intrusion detection model based on deep belief network | |
CN112766343B (en) | Network security situation assessment method based on improved WOA-SVM | |
Kumar et al. | Increasing performance of intrusion detection system using neural network | |
Yin et al. | Towards accurate intrusion detection based on improved clonal selection algorithm | |
CN108900513B (en) | DDOS effect evaluation method based on BP neural network | |
CN109670302B (en) | SVM-based classification method for false data injection attacks | |
CN110290120A (en) | A kind of timing evolved network safe early warning method of cloud platform | |
Yi et al. | Network security risk assessment model based on fuzzy theory | |
CN108632266A (en) | A kind of power distribution communication network security situational awareness method | |
Yang et al. | Network security situation assessment with network attack behavior classification | |
CN109977118A (en) | A kind of abnormal domain name detection method of word-based embedded technology and LSTM | |
Xiao et al. | Network security situation prediction method based on MEA-BP | |
Song et al. | A method of intrusion detection based on WOA‐XGBoost algorithm | |
CN109657147A (en) | Microblogging abnormal user detection method based on firefly and weighting extreme learning machine | |
Chen et al. | Multi-level adaptive coupled method for industrial control networks safety based on machine learning | |
CN116846565A (en) | SAA-SSA-BPNN-based network security situation assessment method | |
Yang et al. | Spam transaction attack detection model based on GRU and WGAN-div | |
Jagadeesan et al. | An efficient botnet detection with the enhanced support vector neural network | |
CN115037553A (en) | Information security monitoring model construction method and device, information security monitoring model application method and device, and storage medium | |
Borisenko et al. | Intrusion detection using multilayer perceptron and neural networks with long short-term memory | |
Ravipati et al. | A survey on different machine learning algorithms and weak classifiers based on KDD and NSL-KDD datasets |
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 |