CN106411896A - APDE-RBF neural network based network security situation prediction method - Google Patents

APDE-RBF neural network based network security situation prediction method Download PDF

Info

Publication number
CN106411896A
CN106411896A CN201610871705.XA CN201610871705A CN106411896A CN 106411896 A CN106411896 A CN 106411896A CN 201610871705 A CN201610871705 A CN 201610871705A CN 106411896 A CN106411896 A CN 106411896A
Authority
CN
China
Prior art keywords
degree
rbf
individual
population
represent
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
Application number
CN201610871705.XA
Other languages
Chinese (zh)
Other versions
CN106411896B (en
Inventor
李方伟
李骐
李俊瑶
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post and Telecommunications
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN201610871705.XA priority Critical patent/CN106411896B/en
Publication of CN106411896A publication Critical patent/CN106411896A/en
Application granted granted Critical
Publication of CN106411896B publication Critical patent/CN106411896B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • 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
    • 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
    • G06N3/0418—Architecture, e.g. interconnection topology using chaos or fractal principles
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Theoretical Computer Science (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention belongs to the technical field of network security and particularly relates to an APDE-RBF (Affinity Propagation Differential Evolution-Radial Basic Function) neural network based network security situation prediction method. The APDE-RBF neural network based network security situation prediction method comprises the steps of dividing and clustering sample data by utilizing an AP clustering algorithm to obtain the number of nodes of hidden layers of the center and network of the RBF; obtaining population diversity by using AP clustering, changing a zoom factor and a crossover probability of a DE algorithm adaptively and optimizing the width and connection weight of the RBF; meanwhile, performing chaotic search on elite individuals and population diversity center of each generation of population in order to avoid falling into local optimization and jumping out of a local extreme point; and determining a final RBF network model, inputting a test dataset and outputting a situation prediction value. The APDE-RBF neural network based network security situation prediction method aims at improving the prediction precision for the network security situation while enhancing the generalization ability.

Description

Network security situation prediction method based on APDE-RBF neutral net
Technical field
The invention belongs to technical field of network security, particularly to a kind of footpath propagating differential evolution algorithm based on attractor To basic function (Affinity Propagation Differential evolution-Radial Basis Function, letter Claim APDE-RBF) neutral net network security situation prediction method.
Background technology
Issued according in January, 2015 China Internet Network Information Center《35th China Internet state of development report》Aobvious Show, cut-off has 46.3% netizen to meet with network security problem in the overall netizen of China in by the end of December, 2014, shows China The safe condition that people internet uses allows of no optimist.With network security problem become increasingly conspicuous with seriously, some traditional safety Defense technique is unable to do what one wishes, and for solving the above problems, the research of network security situation awareness is arisen at the historic moment.
Network safety situation prediction network manager mainly before network loss under attack takes corresponding measure, According to current and conventional network safety situation value, set up the network safe state to following a period of time for the rational Mathematical Modeling It is predicted.Because network attack is random and uncertain, so the prediction to situation value is a complicated non-linear mistake Journey.
The method that researcher proposes a lot of predictions at present, such as statistical method, gray prediction method, neutral net side Method, Markov model, SVMs etc., but all there is respective limitation and deficiency in said method.
The more conventional model of statistical method has：Autoregression model, moving average model and autoregressive moving-average model, But there is following limitation in these models：Seasonal effect in time series data demand steadily, if multiple regression, also require variable it Between be independent；Gray prediction method is suitable for the time series of monotone variation, and for fluctuating, larger time series is difficult to predict； 1987, Application of Neural Network was emulated the study of data and pre- in the time series that produced by computer by Lapdes et al. first Survey, but neutral net haves such problems as that convergence rate is slow, structure choice is difficult and is easily trapped into local minimum, simultaneously because the party Method is affected larger by complicated network structure degree and sample complex, thus study or generalization ability low phenomenon occurs； Markov model needs the mathematical formulae of large amount of complex to derive it is difficult to set up accurate forecast model；SVMs (Support Vector Machine, abbreviation SVM) is difficult to carry out to large-scale training sample, and convergence rate is slow.
Content of the invention
For above-mentioned the deficiencies in the prior art, the present invention proposes a kind of network security state based on APDE-RBF neutral net Gesture Forecasting Methodology is it is intended to while strengthening generalization ability, improve the precision of prediction to network safety situation.
For achieving the above object, a kind of network safety situation prediction based on APDE-RBF neutral net proposed by the present invention Method is it is characterised in that this Forecasting Methodology comprises the following steps：
Step 1：Propagate (Affinity Propagation, abbreviation AP) clustering algorithm using attraction sample data is entered Row partition clustering, thus obtain the center of RBF and the node in hidden layer of network；
Step 2：Draw group difference degree using AP cluster, adaptively change differential evolution (Differential Evolution, vehicle economy) zoom factor of algorithm and crossover probability, the width and connection weight of RBF is optimized；
Step 3：In order to avoid being absorbed in local optimum and jumping out Local Extremum, individual to the elite of every generation population and Group difference degree center carries out Chaos Search；
Step 4：Determine final RBF network model, input test data set, export situation predicted value.
Preferably, described step 1 is further comprising the steps：
Step 11：Calculating the similarity matrix S between input node using Euclidean distance is：S (i, k)=- | | xi-xk| |2, wherein xiAnd xkRepresent RBF neural any two input node, S (i, k) represents point xkAs point xiCluster centre Similarity；
Step 12：Initialization Attraction Degree matrix R and degree of membership matrix A are R (i, k)=0, A (i, k)=0, wherein R (i, k) Represent point xkIt is suitable as data point xiCluster centre degree, A (i, k) represent point xiSelected element xkAs its cluster centre Appropriateness；
Step 13：Determine deflection parameterWherein N represents the quantity of input node, Middle numerical value is occupy in one group of numerical value of median function representation；
Step 14：Calculate Attraction Degree matrix R and degree of membership matrix A according to following formula：
Wherein p (k) represents data point xkAs the point of reference of cluster centre, R (k, k) represents data point xkIt is suitable as certainly The degree of oneself cluster centre, A (k, k') represents data point xkSelect data point xk'As the degree of its cluster centre, S (k, K') represent data point xkSum strong point xk'Similarity degree；From above-mentioned formula can be seen that when p (k) is larger its corresponding R (k, K) also can be larger, and then A (i, k) value also can be larger, thus class to represent k larger as the possibility of final cluster centre；Phase Ying Di, when more p (k) are larger, more classes represent and tend to final cluster centre, therefore, increase or subtract Little p (k) can increase or decrease the clusters number of AP output；
Step 15：Update Attraction Degree matrix R and the formula of degree of membership matrix A is：
R (i, k)=λ * R (i, k)old+(1-λ)*R(i,k)new
A (i, k)=λ * A (i, k)old+(1-λ)*A(i,k)new
Above-mentioned when more new formula represents each iteration, new Attraction Degree matrix R (i, k)newWith degree of membership matrix A (i, k)new Will respectively with last R (i, k)oldWith A (i, k)oldIt is weighted updating, obtain Attraction Degree matrix and the ownership of this iteration Degree matrix, wherein λ represents updating factor；
Step 16：If meeting one of following condition：1. the class center selecting keeps stable, 2. exceedes greatest iteration time Number, then go to step 17, otherwise go to step 14；
Step 17, exports cluster result.
Preferably, described step 2 is further comprising the steps：
Step 21：Execute initialization, process is as follows：
σi=σmin+rand(0,1)*(σmax-σmin)
wi=rand (0,1)
Wherein σiFor RBF neural basic function width, σmaxRepresent two farthest data points in all sample number strong points Apart from width, its computing formula is：σminRepresent in all sample number strong points two Nearly data point apart from width, its computing formula is：wiRepresent hidden layer to output Layer connection weight, rand (0,1) represents equally distributed random number between (0,1)；
Step 22：Execution mutation process, by g+1 for the individual V that makes a variation in populationi(g+1) it is modeled as g in population three Individual function：
Vi(g+1)=Xr1(g)+F*(Xr2(g)-Xr3(g))
i≠r1≠r2≠r3
Wherein XiG () is that g is individual for i-th in population, i.e. Xr1(g)、Xr2(g) and Xr3G () represents that g generation plants respectively In group r1, r2 and the r3 individual, F is zoom factor；
Step 23：Execution crossover process, produces the individual u of i-th jth reform of g+1 generationij(g+1) formula is：
Wherein vij(g+1) represent that g carries out the individuality after mutation operation, x for i-th jth dimension individuality of populationijG () represents G is individual for i-th jth dimension of population, and rand is equally distributed random number, j between (0,1)randIt is whole at random between [1, n] Number, CR represents crossover probability；Above-mentioned formula implication is when stochastic variable rand is less than element pair in crossover probability CR or individuality Ordinal number j is answered to be equal to stochastic variable jrand, that is, using variation individuality in element as new individual it is intended to improve individual variation can Can property；Otherwise, still keep target individual xijG () is constant；
Step 24：Execution selection course is as follows：
Wherein Ui(g+1) it is candidate individual, XiG () is corresponding individual, f () is individual fitness function, is used herein as Mean square error (mean square error, abbreviation MSE) is as fitness function.
Further, in described step 22, zoom factor F is entered Mobile state adjustment formula be：
Wherein FmaxAnd FminRepresent the bound of zoom factor respectively, PD (g) is the group difference degree in g generation, and plants Constellation variance degree represents the cluster number obtained by individualities all in population space are clustered, when group difference degree is bigger, Individuality is distributed more uniform in population space, tries to achieve globally optimal solution possibility bigger；τ1For iteration threshold, gmaxChange for maximum Generation number.
Further, the crossover probability CR that makes in described step 23 can the formula of self-adaptative adjustment be：
Wherein CRminAnd CRmaxRepresent the bound of crossover probability, τ respectively2Iteration threshold for setting.
Preferably, in described step 3, Chaos Search is implemented as：Model 1 dimensional Logistic Map chaos mould first Type, its expression formula is：Zt+1=μ Zt(1-Zt), wherein μ is control parameter,It is a random generation D dimensional vector, t represents chaos iteration number of times；
Secondly, optimum individual and diversity factor center iteration more new formula in modeling population：
Wherein XiRepresent optimum individual or the diversity factor center of population,Represent the new individual after Chaos Search, α table Show chaos regulation parameter, r is the random number between [0,1].
The beneficial effects of the present invention is：Group difference degree is drawn using AP clustering algorithm proposed by the present invention, and adaptive Answer ground to change zoom factor and the crossover probability of DE algorithm, not only optimize width and the connection weight of RBF, and to every generation The elite of population is individual and group difference degree center all carries out Chaos Search, it is to avoid be absorbed in local optimum, is not only able to strengthen Generalization ability, and precision of prediction can be improved.
Brief description
Fig. 1 is the preferred embodiment based on APDE-RBF neutral net network security situation prediction method that the present invention provides Flow chart；
Fig. 2 is algorithms of different situation value prediction contrast simulation figure；
Fig. 3 is the different error contrast simulation figures of algorithms of different；
Fig. 4 is that the different DE algorithms that improve predict contrast simulation figure；
Fig. 5 is the different different error contrast simulation figures improving DE algorithm.
Specific embodiment
For making the object, technical solutions and advantages of the present invention become more apparent, the tool to the present invention below in conjunction with the accompanying drawings Body embodiment is described in further detail.
Fig. 1 is being preferable to carry out of the network security situation prediction method based on APDE-RBF neutral net that the present invention provides Example flow chart, the method specifically includes following steps：
Step 1：Using AP clustering algorithm, partition clustering is carried out to sample data, thus obtaining the center of RBF and network Node in hidden layer；
Step 2：Draw group difference degree using AP cluster, adaptively change the zoom factor of DE algorithm and intersect general Rate, is optimized to the width and connection weight of RBF RBF；
Step 3：In order to avoid being absorbed in local optimum and jumping out Local Extremum, individual to the elite of every generation population and Group difference degree center carries out Chaos Search；
Step 4：Determine final RBF network model, input test data set, export situation predicted value.
According to the present invention, further comprising the steps in described step 1：
Step 11：Calculating the similarity matrix S between input node using Euclidean distance is：S (i, k)=- | | xi-xk| |2, wherein xiAnd xkRepresent RBF neural any two input node, S (i, k) represents point xkAs point xiCluster centre Similarity；
Step 12：Initialization Attraction Degree matrix R and degree of membership matrix A are R (i, k)=0, A (i, k)=0, wherein R (i, k) Represent point xkIt is suitable as data point xiCluster centre degree, A (i, k) represent point xiSelected element xkAs its cluster centre Appropriateness；
Step 13：Determine deflection parameterpkRepresent that each sample data point is selected as clustering The possibility at center, is the value of element on similar matrix S diagonal, and k=1 ..., N, N represent the quantity of input node, Middle numerical value is occupy in one group of numerical value of median function representation；
Step 14：Calculate Attraction Degree matrix R and degree of membership matrix A according to following formula：
Wherein p (k) represents data point xkAs the point of reference of cluster centre, R (k, k) represents data point xkIt is suitable as certainly The degree of oneself cluster centre, A (k, k') represents data point xkSelect data point xk' as its cluster centre degree, S (k, K') represent data point xkSum strong point xk' similarity degree；From above-mentioned formula can be seen that when p (k) is larger its corresponding R (k, K) also can be larger, and then A (i, k) value also can be larger, thus class to represent k larger as the possibility of final cluster centre；Phase Ying Di, when more p (k) are larger, more classes represent and tend to final cluster centre, therefore, increase or subtract Little p (k) can increase or decrease the clusters number of AP output；
Step 15：Update Attraction Degree matrix R and the formula of degree of membership matrix A is：
R (i, k)=λ * R (i, k)old+(1-λ)*R(i,k)new
A (i, k)=λ * A (i, k)old+(1-λ)*A(i,k)new
Above-mentioned when more new formula represents each iteration, new Attraction Degree matrix R (i, k)newWith degree of membership matrix A (i, k)new Will respectively with last R (i, k)oldWith A (i, k)oldIt is weighted updating, obtain Attraction Degree matrix and the ownership of this iteration Degree matrix, wherein λ represents updating factor；
Step 16：If meeting one of following condition：1. the class center selecting keeps stable, 2. exceedes greatest iteration time Number, then go to step 17, otherwise go to step 14；
Step 17, exports cluster result.
According to the present invention, further comprising the steps in described step 2：
Step 21：Execute initialization, process is as follows：
σi=σmin+rand(0,1)*(σmax-σmin)
wi=rand (0,1)
Wherein σiFor RBF neural basic function width, σmaxRepresent two farthest data points in all sample number strong points Apart from width, its computing formula is：σminRepresent in all sample number strong points two Nearly data point apart from width, its computing formula is：ci、cjRepresent any two not Same hidden layer node, wiRepresent hidden layer arrive output layer connection weight, rand (0,1) expression (0,1) between equally distributed with Machine number；
Step 22：Execution mutation process, by g+1 for the individual V that makes a variation in populationi(g+1) it is modeled as g in population three Individual function：
Vi(g+1)=Xr1(g)+F*(Xr2(g)-Xr3(g))i≠r1≠r2≠r3
Wherein Xr1(g)、Xr2(g) and Xr3G () represents g for r1 in population, r2 and r3 respectively Body, F is zoom factor；
Step 23：Execution crossover process, produces new individual uij(g+1) formula is as follows：
Wherein vij(g+1) represent that g carries out the individuality after mutation operation, x for i-th jth dimension individuality of populationijG () represents G is individual for i-th jth dimension of population, and rand is equally distributed random number, j between (0,1)randIt is whole at random between [1, n] Number, CR represents crossover probability；Above-mentioned formula implication is when stochastic variable rand is less than element pair in crossover probability CR or individuality Ordinal number j is answered to be equal to stochastic variable jrand, that is, using variation individuality in element as new individual it is intended to improve individual variation can Can property；Otherwise, still keep target individual xijG () is constant；
Step 24：Execution selection course, specific as follows：
Wherein Ui(g+1) it is candidate individual, XiG () is corresponding individual, f () is individual fitness function, is used herein as Mean square error (mean square error, abbreviation MSE) is as fitness function.
According to the present invention, the formula that F is entered with Mobile state adjustment in described step 22 is：
Wherein FmaxAnd FminRepresent the bound of zoom factor respectively, PD (g) is the group difference degree in g generation, and plants Constellation variance degree represents the cluster number obtained by individualities all in population space are clustered, when group difference degree is bigger, Individuality is distributed more uniform in population space, tries to achieve globally optimal solution possibility bigger；τ1For the iteration threshold of setting, gmaxFor Maximum iteration time.
According to the present invention, further, the crossover probability CR that makes in described step 23 can the formula of self-adaptative adjustment be：
Wherein CRminAnd CRmaxIt is the bound of crossover probability, τ2It is the iteration threshold of setting.
According to the present invention, further, in described step 3, Chaos Search is implemented as：Model dimensional Logistic first Map chaotic model, its expression formula is：Zt+1=μ Zt(1-Zt) this formula is iterative formula in mathematical meaning, its valueIt is a random D dimensional vector generating, wherein, μ is control parameter, and t represents chaos iteration number of times；
Secondly, optimum individual and diversity factor center iteration more new formula in modeling population：
Wherein XiRepresent optimum individual or the diversity factor center of population,Represent the new individual after Chaos Search, α table Show chaos regulation parameter, r is the random number between [0,1].
Beneficial effect to illustrate the invention, is further analyzed according to simulation result.
Fig. 2 describes the network safety situation value obtaining based on algorithms of different.Figure it is seen that autoregression is slided putting down All model (Auto-Regressive and Moving Average Model, abbreviation ARMA) mainly at random smoothly when Between sequence, but be because randomness and the complexity of network attack, network safety situation sequence be non-smoothly；Gray model (Grey Model, abbreviation GM) is good for the time series forecasting effect of monotone variation, otherwise error is big；Least square support to The supporting vector of amount machine (Least Squares Support Vector Machines, abbreviation LSSVM) becomes all data Point, loses the openness feature of SVM；Kmeans-RBF needs to preset hidden layer node, have ignored the spy of data itself Point, weakens the generalization ability of RBF, however, using actual value as criterion, different degrees of compared to what said method occurred Error and defect, APDE-RBF Neural Network model predictive precision highest proposed by the present invention.
Fig. 3 shows the different errors contrast of algorithms of different, as can be seen from Figure 3 whether average relative error, all square Root error or relative mean square error, APDE-RBF neural network model is held at less error level, embodies higher Precision of prediction.
Fig. 4 display is different to improve the network safety situation value in different time points for the DE algorithm.DE algorithm be fixing F and CR, is easily absorbed in local optimum；Differential evolution version (the Simplified Differential Evolution simplifying Version, abbreviation SDE) algorithm F adopt simple random number；Differential evolution algorithm based on all parameters and Mutation Strategy (Differential evolution algorithm with ensemble of parameters and mutation Strategies, abbreviation EPSDE) algorithm using Mutation Strategy pond and parameter pond random combine be iterated evolve；Self adaptation is poor F and CR dividing (Self-Adaptive Differential Evolution, abbreviation jDE) algorithm of evolving relies on random number differentiation Thus obtaining different results；Differential evolution (Evolution based on composite testing vector generation strategy and control parameter With Composite Trial Vector Generation Strategies and Control Parameters, referred to as CoDE) algorithm is the Mutation Strategies different using three kinds and parameter setting competition coupling is iterated evolving.Although said method Mutation Strategy to DE algorithm and parameter setting carry out adaptive impovement, but are all mostly random numbers or rely on random number and carry out Differentiate and choose, lead to evolve unstable.Further, it can be seen that using actual value as criterion, compared to above-mentioned Algorithm, APDE algorithm generally maintains relatively low absolute error, and its reason is that APDE-RBF neural network model relies on and plants Constellation variance degree and iterative evolution degree carry out self-adaptative adjustment to F and CR, so that population is evolved to beneficial direction, accelerate algorithm Convergence rate.
Fig. 5 is the different error contrasts of algorithms of different, and whether average relative error, root mean square miss as can be seen from Figure 5 Difference or relative mean square error, APDE-RBF neural network model proposed by the present invention is held at less error level, body Show higher precision of prediction.
The lifted embodiment of the present invention or embodiment have been carried out to the object, technical solutions and advantages of the present invention further Detailed description, be should be understood that embodiment provided above or embodiment be only the preferred embodiment of the present invention and , not in order to limit the present invention, all any modifications made for the present invention within the spirit and principles in the present invention, equivalent replace Change, improve, should be included within the scope of the present invention.

Claims (6)

1. a kind of network security state of the RBF APDE-RBF neutral net propagating differential evolution algorithm based on attractor Gesture Forecasting Methodology is it is characterised in that the method comprises the following steps：
Step 1：Propagate AP clustering algorithm using attraction and partition clustering is carried out to sample data, thus obtaining RBF The center of RBF and the node in hidden layer of network；
Step 2：Draw group difference degree using AP cluster, adaptively change zoom factor and the intersection of differential evolution DE algorithm Probability, is optimized to the width and connection weight of RBF；
Step 3：In order to avoid being absorbed in local optimum and jumping out Local Extremum, and population individual to the elite of every generation population Diversity factor center carries out Chaos Search；
Step 4：Determine final RBF network model, input test data set, export situation predicted value.
2. the network security situation prediction method based on APDE-RBF neutral net according to claim 1, its feature exists In described step 1 is further comprising the steps：
Step 11：Calculating the similarity matrix S between input node using Euclidean distance is：S (i, k)=- | | xi-xk||2, its Middle xiAnd xkRepresent RBF neural any two input node, S (i, k) represents point xkAs point xiCluster centre similar Degree, its value is stored in similar matrix S；
Step 12：Initialization Attraction Degree matrix R and degree of membership matrix A are R (i, k)=0, A (i, k)=0, and wherein R (i, k) represents Point xkIt is suitable as data point xiCluster centre degree, A (i, k) represent point xiSelected element xkSuitable as its cluster centre Conjunction degree；
Step 13：Determine deflection parameterPk represents that each sample data point is selected as cluster centre Possibility, is the value of element on similar matrix S diagonal, and k=1 ..., N, N represent the quantity of input node, median function Represent and take the numerical value occuping middle in one group of numerical value；
Step 14：Calculate Attraction Degree matrix R and degree of membership matrix A according to following formula：
R ( i , k ) = S ( i , k ) - m a x k ≠ k ′ { A ( i , k ′ ) + S ( i , k ′ ) }
A ( i , k ) = m i n { 0 , R ( k , k ) + Σ i ′ ∉ { i , k } m a x { 0 , R ( i ′ , k ) } }
R ( k , k ) = p ( k ) - m a x k ≠ k ′ { A ( k , k ′ ) + S ( k , k ′ ) }
Wherein p (k) represents data point xkAs the point of reference of cluster centre, R (k, k) represents data point xkIt is suitable as oneself The degree of cluster centre, A (k, k') represents data point xkSelect data point xk'As the degree of its cluster centre, S (k, k') table Registration strong point xkSum strong point xk'Similarity degree；
Step 15：Update Attraction Degree matrix R and the formula of degree of membership matrix A is：
R (i, k)=λ * R (i, k)old+(1-λ)*R(i,k)new
A (i, k)=λ * A (i, k)old+(1-λ)*A(i,k)new
Above-mentioned when more new formula represents each iteration, new Attraction Degree matrix R (i, k)newWith degree of membership matrix A (i, k)newDivide Not with last Attraction Degree matrix R (i, k)oldWith degree of membership matrix A (i, k)oldIt is weighted updating, obtain this iteration Attraction Degree matrix and degree of membership matrix, wherein λ represents updating factor；
Step 16：If meeting one of following condition：1. the class center selecting keeps stable, 2. exceedes maximum iteration time, then Go to step 17, otherwise go to step 14；
Step 17, exports cluster result.
3. the network security situation prediction method based on APDE-RBF neutral net according to claim 1, its feature exists In further comprising the steps in described step 2：
Step 21：Execute initialization, process is as follows：
σi=σmin+rand(0,1)*(σmax-σmin)
wi=rand (0,1)
Wherein σiFor RBF neural basic function width, σmaxRepresent the distance of two farthest data points in all sample number strong points Width, its computing formula is：σminRepresent that in all sample number strong points, two count recently Strong point apart from width, its computing formula is：ci、cjRepresent that any two is different Hidden layer node, wiRepresent hidden layer to output layer connection weight, equally distributed random number between rand (0,1) expression (0,1)；
Step 22：Execution mutation process, by g+1 for the individual V that makes a variation in populationi(g+1) it is modeled as g for three in population The function of body：
Vi(g+1)=Xr1(g)+F*(Xr2(g)-Xr3(g))
i≠r1≠r2≠r3
Wherein XiG () is that g is individual for i-th in population, i.e. Xr1(g)、Xr2(g) and Xr3G () represents g in population respectively R1, r2 and the r3 are individual, and F is zoom factor；
Step 23：Execution crossover process, produces the individual u of i-th jth reform of g+1 generationij(g+1) formula is：
Wherein vij(g+1) represent that g carries out the individuality after mutation operation, x for i-th jth dimension individuality of populationijG () represents g Individual for i-th jth dimension of population, rand is equally distributed random number, j between (0,1)randIt is the random integers between [1, n], CR Represent crossover probability；Above-mentioned formula implication is：Correspond to sequence when stochastic variable rand is less than element in crossover probability CR or individuality Number j is equal to stochastic variable jrand, then adopt the element in variation individuality as new individual it is intended to improve the possibility of individual variation Property；Otherwise, still keep target individual xijG () is constant；
Step 24：Execution selection course is as follows：
Wherein Ui(g+1) it is candidate individual, XiG () is corresponding individual, f () is individual fitness function, is used herein as all Square error is as fitness function.
4. the network security situation prediction method based on APDE-RBF neutral net according to claim 3, its feature exists In the formula that in described step 22, zoom factor F is entered with Mobile state adjustment is：
Wherein FmaxAnd FminRepresent the bound of zoom factor respectively, PD (g) is the group difference degree in g generation, and population is poor Different degree represents the cluster number obtained by individualities all in population space are clustered, when group difference degree is bigger, individual Population space is distributed more uniform, tries to achieve globally optimal solution possibility bigger；τ1For the iteration threshold of setting, gmaxFor maximum Iterations.
5. the network security situation prediction method based on APDE-RBF neutral net according to claim 3, its feature exists In the crossover probability CR that makes in described step 23 can the formula of self-adaptative adjustment be：
Wherein CRminAnd CRmaxRepresent the bound of crossover probability, τ respectively2For the iteration threshold of setting, gmaxFor greatest iteration time Number.
6. the network security situation prediction method based on APDE-RBF neutral net according to claim 1, its feature exists In in described step 3, Chaos Search is specially：Model 1 dimensional Logistic Map chaotic model first, its expression formula is：Zt+1 =μ Zt(1-Zt), wherein, ZtIt is a D dimensional vector, μ is control parameter, and t represents chaos iteration number of times；Secondly, in modeling population Optimum individual and diversity factor center iteration more new formula：
X i t + 1 = X i + αZ t + 1
Wherein XiRepresent optimum individual or the diversity factor center of population,Represent the new individual after Chaos Search, α represents mixed Ignorant regulation parameter, r is the random number between [0,1].
CN201610871705.XA 2016-09-30 2016-09-30 Network security situation prediction method based on APDE-RBF neural network Active CN106411896B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610871705.XA CN106411896B (en) 2016-09-30 2016-09-30 Network security situation prediction method based on APDE-RBF neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610871705.XA CN106411896B (en) 2016-09-30 2016-09-30 Network security situation prediction method based on APDE-RBF neural network

Publications (2)

Publication Number Publication Date
CN106411896A true CN106411896A (en) 2017-02-15
CN106411896B CN106411896B (en) 2019-04-23

Family

ID=59229451

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610871705.XA Active CN106411896B (en) 2016-09-30 2016-09-30 Network security situation prediction method based on APDE-RBF neural network

Country Status (1)

Country Link
CN (1) CN106411896B (en)

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107194524A (en) * 2017-07-28 2017-09-22 合肥工业大学 A kind of Coal and Gas Outbursts Prediction method based on RBF neural
CN107222333A (en) * 2017-05-11 2017-09-29 中国民航大学 A kind of network node safety situation evaluation method based on BP neural network
CN108009241A (en) * 2017-11-30 2018-05-08 昆山青石计算机有限公司 A kind of industrial Internet of Things safe polymeric correlating method based on PSO parameter optimizations
CN108548846A (en) * 2018-06-21 2018-09-18 电子科技大学 Bearing device subsurface defect extracting method based on thermal image AP clusters
CN108594793A (en) * 2018-04-17 2018-09-28 电子科技大学 A kind of improved RBF flight control systems fault diagnosis network training method
CN108881250A (en) * 2018-06-28 2018-11-23 广东电网有限责任公司 Powerline network security postures prediction technique, device, equipment and storage medium
CN109040027A (en) * 2018-07-12 2018-12-18 北京理工大学 The active predicting method of network vulnerability node based on gray model
CN109102062A (en) * 2018-08-15 2018-12-28 桂林电子科技大学 3D NoC test-schedule method based on Petri network Yu chaos difference glowworm swarm algorithm
CN109117872A (en) * 2018-07-24 2019-01-01 贵州电网有限责任公司信息中心 A kind of user power utilization behavior analysis method based on automatic Optimal Clustering
CN109543914A (en) * 2018-11-29 2019-03-29 北京化工大学 The prediction technique of RBF constructing operation energy consumption based on AP clustering algorithm
CN110362558A (en) * 2019-06-12 2019-10-22 广东工业大学 A kind of energy consumption data cleaning method based on neighborhood propagation clustering
CN110474904A (en) * 2019-08-16 2019-11-19 武汉思普崚技术有限公司 A kind of Situation Awareness method and system improving prediction
CN111783988A (en) * 2020-06-16 2020-10-16 长江大学 Multi-target grey prediction evolution method and system for solving environmental economic scheduling problem
CN112509017A (en) * 2020-11-18 2021-03-16 西北工业大学 Remote sensing image change detection method based on learnable difference algorithm
CN113141272A (en) * 2021-04-20 2021-07-20 浙江大学 Network security situation analysis method based on iteration optimization RBF neural network
CN113191917A (en) * 2021-03-09 2021-07-30 中国大唐集团科学技术研究院有限公司 Power plant industrial control system network security threat classification method based on radial basis function algorithm
CN114155390A (en) * 2021-12-02 2022-03-08 北京中科智易科技有限公司 Clustering device for classifying metadata of classified objects
CN116843368A (en) * 2023-07-17 2023-10-03 杭州火奴数据科技有限公司 Marketing data processing method based on ARMA model

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101975092A (en) * 2010-11-05 2011-02-16 中北大学 Real-time prediction method of mine gas concentration in short and medium terms based on radial basis function neural network integration
US20150188823A1 (en) * 2013-12-03 2015-07-02 Akamai Technologies, Inc. Virtual private network (VPN)-as-a-service with load-balanced tunnel endpoints
CN104978857A (en) * 2015-05-26 2015-10-14 重庆邮电大学 Traffic state prediction method based on chaos theory and device thereof
CN105678380A (en) * 2016-01-08 2016-06-15 浙江工业大学 Ecological niche and adaptive negative correlation learning-based evolutionary neural network integration method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101975092A (en) * 2010-11-05 2011-02-16 中北大学 Real-time prediction method of mine gas concentration in short and medium terms based on radial basis function neural network integration
US20150188823A1 (en) * 2013-12-03 2015-07-02 Akamai Technologies, Inc. Virtual private network (VPN)-as-a-service with load-balanced tunnel endpoints
CN104978857A (en) * 2015-05-26 2015-10-14 重庆邮电大学 Traffic state prediction method based on chaos theory and device thereof
CN105678380A (en) * 2016-01-08 2016-06-15 浙江工业大学 Ecological niche and adaptive negative correlation learning-based evolutionary neural network integration method

Cited By (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107222333A (en) * 2017-05-11 2017-09-29 中国民航大学 A kind of network node safety situation evaluation method based on BP neural network
CN107194524A (en) * 2017-07-28 2017-09-22 合肥工业大学 A kind of Coal and Gas Outbursts Prediction method based on RBF neural
CN107194524B (en) * 2017-07-28 2020-05-22 合肥工业大学 RBF neural network-based coal and gas outburst prediction method
CN108009241A (en) * 2017-11-30 2018-05-08 昆山青石计算机有限公司 A kind of industrial Internet of Things safe polymeric correlating method based on PSO parameter optimizations
CN108594793A (en) * 2018-04-17 2018-09-28 电子科技大学 A kind of improved RBF flight control systems fault diagnosis network training method
CN108548846A (en) * 2018-06-21 2018-09-18 电子科技大学 Bearing device subsurface defect extracting method based on thermal image AP clusters
CN108881250A (en) * 2018-06-28 2018-11-23 广东电网有限责任公司 Powerline network security postures prediction technique, device, equipment and storage medium
CN109040027A (en) * 2018-07-12 2018-12-18 北京理工大学 The active predicting method of network vulnerability node based on gray model
CN109117872A (en) * 2018-07-24 2019-01-01 贵州电网有限责任公司信息中心 A kind of user power utilization behavior analysis method based on automatic Optimal Clustering
CN109102062B (en) * 2018-08-15 2020-03-13 桂林电子科技大学 3D NoC test planning method based on Petri network and chaotic differential firefly algorithm
CN109102062A (en) * 2018-08-15 2018-12-28 桂林电子科技大学 3D NoC test-schedule method based on Petri network Yu chaos difference glowworm swarm algorithm
CN109543914A (en) * 2018-11-29 2019-03-29 北京化工大学 The prediction technique of RBF constructing operation energy consumption based on AP clustering algorithm
CN109543914B (en) * 2018-11-29 2023-02-28 北京化工大学 Method for predicting RBF building operation energy consumption based on AP clustering algorithm
CN110362558A (en) * 2019-06-12 2019-10-22 广东工业大学 A kind of energy consumption data cleaning method based on neighborhood propagation clustering
CN110362558B (en) * 2019-06-12 2022-12-16 广东工业大学 Energy consumption data cleaning method based on neighborhood propagation clustering
CN110474904B (en) * 2019-08-16 2022-04-12 武汉思普崚技术有限公司 Situation awareness method and system for improving prediction
CN110474904A (en) * 2019-08-16 2019-11-19 武汉思普崚技术有限公司 A kind of Situation Awareness method and system improving prediction
CN111783988A (en) * 2020-06-16 2020-10-16 长江大学 Multi-target grey prediction evolution method and system for solving environmental economic scheduling problem
CN111783988B (en) * 2020-06-16 2024-03-26 长江大学 Multi-target gray prediction evolution method and system for solving environmental economic scheduling problem
CN112509017A (en) * 2020-11-18 2021-03-16 西北工业大学 Remote sensing image change detection method based on learnable difference algorithm
CN113191917A (en) * 2021-03-09 2021-07-30 中国大唐集团科学技术研究院有限公司 Power plant industrial control system network security threat classification method based on radial basis function algorithm
CN113141272B (en) * 2021-04-20 2022-03-15 浙江大学 Network security situation analysis method based on iteration optimization RBF neural network
CN113141272A (en) * 2021-04-20 2021-07-20 浙江大学 Network security situation analysis method based on iteration optimization RBF neural network
CN114155390A (en) * 2021-12-02 2022-03-08 北京中科智易科技有限公司 Clustering device for classifying metadata of classified objects
CN114155390B (en) * 2021-12-02 2022-05-06 北京中科智易科技有限公司 Clustering device for classifying metadata of classified objects
CN116843368A (en) * 2023-07-17 2023-10-03 杭州火奴数据科技有限公司 Marketing data processing method based on ARMA model
CN116843368B (en) * 2023-07-17 2024-01-26 杭州火奴数据科技有限公司 Marketing data processing method based on ARMA model

Also Published As

Publication number Publication date
CN106411896B (en) 2019-04-23

Similar Documents

Publication Publication Date Title
CN106411896A (en) APDE-RBF neural network based network security situation prediction method
Zhong et al. Super efficiency SBM-DEA and neural network for performance evaluation
Lobato et al. Multi-objective genetic algorithm for missing data imputation
Cheng et al. Accurately predicting building energy performance using evolutionary multivariate adaptive regression splines
Ergezer et al. Oppositional biogeography-based optimization
CN108647233A (en) A kind of answer sort method for question answering system
CN113190688B (en) Complex network link prediction method and system based on logical reasoning and graph convolution
CN111416797B (en) Intrusion detection method for optimizing regularization extreme learning machine by improving longicorn herd algorithm
CN107241213A (en) A kind of web service composition method learnt based on deeply
CN106910337A (en) A kind of traffic flow forecasting method based on glowworm swarm algorithm Yu RBF neural
CN110276679A (en) A kind of network individual credit fraud detection method towards deep learning
Kobayashi et al. Bayesian analysis of quantile regression for censored dynamic panel data
Donate et al. Evolutionary optimization of sparsely connected and time-lagged neural networks for time series forecasting
CN112733997A (en) Hydrological time series prediction optimization method based on WOA-LSTM-MC
Ketipi et al. A flexible nonlinear approach to represent cause–effect relationships in FCMs
CN114780879A (en) Interpretable link prediction method for knowledge hypergraph
Kim et al. Knowledge extraction and representation using quantum mechanics and intelligent models
CN107704944B (en) Construction method of stock market fluctuation interval prediction model based on information theory learning
CN102708298B (en) A kind of Vehicular communication system electromagnetic compatibility index distribution method
Liu et al. Research on multi-digital twin and its application in wind power forecasting
CN114707655A (en) Quantum line conversion method, quantum line conversion system, storage medium and electronic equipment
CN114819151A (en) Biochemical path planning method based on improved agent-assisted shuffled frog leaping algorithm
CN107451943A (en) The site selecting method of Urban renewal
CN105512754A (en) Conjugate prior-based single-mode distribution estimation optimization method
CN111563767A (en) Stock price prediction method and device

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant