CN111242272A - Wireless sensor network anomaly detection method - Google Patents
Wireless sensor network anomaly detection method Download PDFInfo
- Publication number
- CN111242272A CN111242272A CN202010040468.9A CN202010040468A CN111242272A CN 111242272 A CN111242272 A CN 111242272A CN 202010040468 A CN202010040468 A CN 202010040468A CN 111242272 A CN111242272 A CN 111242272A
- Authority
- CN
- China
- Prior art keywords
- fsvm
- sample
- search
- sensor network
- wireless sensor
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 51
- 241000255588 Tephritidae Species 0.000 claims abstract description 49
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000012706 support-vector machine Methods 0.000 claims abstract description 19
- 230000002159 abnormal effect Effects 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 11
- 238000005457 optimization Methods 0.000 claims abstract description 8
- 241000255581 Drosophila <fruit fly, genus> Species 0.000 claims description 22
- 238000000034 method Methods 0.000 claims description 15
- 230000005856 abnormality Effects 0.000 claims description 7
- 241000255601 Drosophila melanogaster Species 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000002790 cross-validation Methods 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 230000006872 improvement Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000010845 search algorithm Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000007123 defense Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
- H04W12/12—Detection or prevention of fraud
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a wireless sensor network anomaly detection method, which is based on a fruit fly optimization algorithm and applies a fuzzy support vector machine to the field of wireless sensor network anomaly detection and comprises the following steps: a training stage: s1: collecting sensor detection data, and preprocessing the data to form a training data set; s2: establishing a wireless sensor network anomaly detection model based on the FSVM technology; s3: establishing an IFOA-FSVM model, and performing anomaly detection training on a data set; a detection stage: s4: collecting sensor detection data, and preprocessing the data to form a sample to be detected; s5: and inputting the sample to be detected into the IFOA-FSVM model for detection, and judging whether the sample to be detected is abnormal or not.
Description
Technical Field
The invention relates to the field of wireless sensor network information security, in particular to a wireless sensor network anomaly detection method.
Background
The wireless sensor network is a special mobile ad hoc network different from the conventional wired network. The wireless sensor network is widely applied to various fields, such as military fields of national defense, anti-terrorism and the like, due to the advantages of simple node networking, capability of self-organizing to form a network, low node cost and the like. When the wireless network sensor is abnormal, abnormal data in the sensor network can be detected efficiently in real time, and the method has very important significance for early warning and prevention of external emergencies and health condition monitoring of the sensor network.
In recent years, many studies have been made at home and abroad on a method for detecting abnormal data of a wireless sensor network. Among these researches, the anomaly detection technology based on the support vector machine has become a research hotspot of the anomaly detection technology due to the simple structure, the fast training speed, the good prediction accuracy and the sufficient theoretical basis. However, most of the existing wireless sensor network anomaly detection models based on the SVM (support vector machine) do not introduce the fuzzy theory, so that the models are weak in generalization capability on data sets containing noise samples compared with the support vector machine model based on the fuzzy theory.
Like the SVM, the size of the hyperparameter of the FSVM (fuzzy support vector machine) has a large influence on the prediction capability of the model, the hyperparameter of the algorithm is difficult to set, a certain strategy needs to be adopted for setting, and in view of the situation, the drosophila optimization algorithm is adopted for setting.
Disclosure of Invention
To the deficiency of the prior art, the technical problem to be solved by the present patent application is: how to provide a wireless sensor network anomaly detection method which can solve the problem of hyper-parameter setting and improve anomaly detection precision.
In order to achieve the purpose, the invention adopts the following technical scheme:
a wireless sensor network anomaly detection method is based on a fruit fly optimization algorithm, and the method applies a fuzzy support vector machine to the field of wireless sensor network anomaly detection and comprises the following steps:
a training stage:
s1: collecting sensor detection data, and preprocessing the data to form a training data set;
s2: establishing a wireless sensor network anomaly detection model based on the FSVM technology;
s3: establishing an IFOA-FSVM model, and performing anomaly detection training on a data set;
a detection stage:
s4: collecting sensor detection data, and preprocessing the data to form a sample to be detected;
s5: and inputting the sample to be detected into the IFOA-FSVM model for detection, and judging whether the sample to be detected is abnormal or not.
Preferably, in step S3, the building of the IFOA-FSVM model includes the following steps:
a1: improving a fruit fly algorithm;
a2: searching for the FSVM to obtain the optimal parameters by combining the training data set in the step S1 and the improved drosophila algorithm in the step A1;
a3 brings the optimal parameters back to the FSVM to complete the training of the anomaly detection model.
Preferably, in step a1, the improvement of the drosophila algorithm comprises the following steps:
b1: the quantity Num of the fruit flies in the initial search algorithm, the total iteration times T of the algorithm, and the current optimal position H of the fruit fly populationopt,jHistorical optimum odor concentration bestSmell of fruit fly population, step length adjustment parameters rho andextracting zz proportion drosophila individuals from each iteration to search again, wherein j belongs to n, and n is the variable number of the optimization objective function;
b2: calculating a step adjustment factor m (k);
B3: if k < T/2, the fruit fly population is divided into two subgroups group1 and group2, each subgroup executing a random step search strategy with a different range of advance steps. Wherein, group1 carries out search iteration according to the random value of the normal step range, and group2 carries out search iteration according to the random value of the large step range;
Hi,j=Hopt,j+RandomValue×m(k)
Hi,j=Hopt,j+3×RandomValue×m(k)
wherein i is the ith fruit fly individual in the population, and j is the jth variable in the objective function;
b4: and if k is less than T/2, modifying the search step range of partial fruit fly subgroups. Wherein, group1 still performs search iteration according to the random value of the normal step range in B3, and group2 performs search iteration according to the random value of the large step range;
Hi,j=Hopt,j+0.2×RandomValue×m(k)
b5: calculating the odor concentration value smell of each fruit fly according to the fitness function;
smelli=Function(Hi)
wherein, the Function is a fitness Function, namely a target Function to be optimized;
b6: sorting the odor concentration smell of the fruit flies, finding out fruit fly individuals (smllBest) with the largest smell value, namely current optimal fruit fly individuals, judging whether the smell is larger than the historical optimal odor concentration bestsell or not, and if so, updating the bestsell and the current optimal position Hopt,j(ii) a Otherwise, not updating, and finding out the worst drosophila group with a given proportion;
b7: the worst drosophila melanogaster population is searched again according to the random value of the normal step length range in the B3, the corresponding smell value is calculated, the optimal drosophila melanogaster individual in the search is found, if the smell value is larger than bestsell, bestsell and the current optimal position H are updatedopt,j(ii) a Otherwise, the updating is not carried out.
B8: judging whether the maximum search times T is met, if so, stopping searching, and returning the searched optimal objective function solution; otherwise, the execution is repeated from B2 until the termination condition is satisfied.
Preferably, the preprocessing of the data in steps S1 and S4 includes missing value padding and feature value normalization; in the process of carrying out characteristic value standardization, min-max standardization is adopted to carry out pretreatment on characteristic data, and the value range of the treated characteristic is between [0, 1 ]. Given the data:
wherein l is the number of samples, and d is the characteristic number of the samples; .
The minimum and maximum values for each column of features in matrix X use vector XtminAnd xtmaxIt is shown that,
xtmin={x1min,x2min,...,xdmin}
xtmax={x1max,x2max,...,xdmax}
each sample xt∈XAuThe following normalization is performed in order to perform,
wherein x istpIs xtAnd (5) normalizing the feature vector.
Preferably, the specific steps of step S2 are:
c1: given the training set D of the fuzzy support vector machine:
D={(x1,y1,s1),...,(xi,yi,si),...,(xl,yl,sl)},
wherein x isi∈RdD is the feature dimension of each sample, and l is the number of samples; y isiFor each sample category, indicating normal or abnormal; siRepresenting the fuzzy membership degree of each sample as the probability that the corresponding sample belongs to a certain class, and classifying the hyperplane by the FSVM as follows;
wherein z isiThe characteristic value of the original characteristic of the sample after being processed by the kernel function corresponding to the FSVM, namely RdThe feature vector in (1) is passed through a mapping functionA conversion to the feature space z is made,w is a normal vector of the hyperplane, and determines the direction of the hyperplane; b is a displacement term, and determines the distance between the hyperplane and the origin; epsiloniIs a relaxation variable which functions to reduce the influence of outliers on the classification hyperplane; c is a penalty coefficient, has great influence on the final classification capability of the model and is a parameter needing key search;
c2: and establishing an anomaly detection model based on the FSVM.
Wherein, aiIs a Lagrange factor; k (·,. cndot.) is a kernel function, K (x)i,xj)=zi·zj。
Preferably, the steps a2 and A3 comprise the following sub-steps:
d1: initializing relevant parameters such as iteration times, step length adjustment factors and the like of the IFOA, and selecting an FSVM kernel function according to the data set condition;
d2: searching for the FSVM hyperparameters C and g by improving a drosophila algorithm, inputting the searched hyperparameters into the FSVM algorithm, performing cross validation on a training set, and obtaining the optimal odor concentration smlBest of the group of parameters corresponding to the iterative population of the round by combining a selected error function;
d3: and continuously iterating to find optimal parameters C and g according to an updating strategy of the drosophila algorithm, and further solving w and b of the FSVM.
Preferably, the step S5 uses the following formula when detecting the sample:
advantageous effects
(1) The fuzzy support vector machine is applied to the field of wireless sensor network anomaly detection, and compared with an anomaly detection algorithm based on a support vector, the model has relatively strong robustness.
(2) The fruit fly algorithm is improved, and the advancing step length of the fruit flies can be intelligently adjusted according to the iteration times; dividing the fruit flies into two subgroups, and executing search strategies with different step length ranges in the early stage and the later stage of search; a certain proportion of fruit flies are selected in each iteration process for re-searching, so that the searching diversity of the fruit flies is enriched, and the searching capability of the fruit fly algorithm is improved.
(3) The invention adopts the improved drosophila algorithm to search the hyperparameters of the fuzzy support vector machine, and improves the abnormal detection accuracy of the fuzzy support vector machine.
Description of the drawings:
FIG. 1 is a diagram of a model of the anomaly detection method of the present invention;
FIG. 2 is a flow chart of the improved fruit fly algorithm of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
Referring to fig. 1-2, a method for detecting an abnormality of a wireless sensor network, based on a drosophila optimization algorithm, for applying a fuzzy support vector machine to the field of detecting an abnormality of a wireless sensor network, includes the following steps:
a training stage:
s1: collecting sensor detection data, and preprocessing the data to form a training data set;
s2: establishing a wireless sensor network anomaly detection model based on the FSVM technology;
s3: establishing an IFOA-FSVM model, and performing anomaly detection training on a data set;
a detection stage:
s4: collecting sensor detection data, and preprocessing the data to form a sample to be detected;
s5: and inputting the sample to be detected into the IFOA-FSVM model for detection, and judging whether the sample to be detected is abnormal or not.
In this embodiment, in step S3, the establishing of the IFOA-FSVM model includes the following steps:
a1: improving a fruit fly algorithm;
a2: searching for the FSVM to obtain the optimal parameters by combining the training data set in the step S1 and the improved drosophila algorithm in the step A1;
a3 brings the optimal parameters back to the FSVM to complete the training of the anomaly detection model.
In this embodiment, in the step a1, the improvement of the drosophila algorithm includes the following steps:
b1: the quantity Num of the fruit flies in the initial search algorithm, the total iteration times T of the algorithm, and the current optimal position H of the fruit fly populationopt,jHistorical optimum odor concentration bestSmell of fruit fly population, step length adjustment parameters rho andextracting zz proportion drosophila individuals from each iteration to search again, wherein j belongs to n, and n is the variable number of the optimization objective function;
b2: calculating a step adjustment factor m (k);
B3: if k < T/2, the fruit fly population is divided into two subgroups group1 and group2, each subgroup executing a random step search strategy with a different range of advance steps. Wherein, group1 carries out search iteration according to the random value of the normal step range, and group2 carries out search iteration according to the random value of the large step range;
Hi,j=Hopt,j+RandomValue×m(k)
Hi,j=Hopt,j+3×RandomValue×m(k)
wherein i is the ith fruit fly individual in the population, and j is the jth variable in the objective function;
b4: and if k is less than T/2, modifying the search step range of partial fruit fly subgroups. Wherein, group1 still performs search iteration according to the random value of the normal step range in B3, and group2 performs search iteration according to the random value of the large step range;
Hi,j=Hopt,j+0.2×RandomValue×m(k)
b5: calculating the odor concentration value smell of each fruit fly according to the fitness function;
smelli=Function(Hi)
wherein, the Function is a fitness Function, namely a target Function to be optimized;
b6: sorting the odor concentration smell of the fruit flies, finding out fruit fly individuals (smllBest) with the largest smell value, namely current optimal fruit fly individuals, judging whether the smell is larger than the historical optimal odor concentration bestsell or not, and if so, updating the bestsell and the current optimal position Hopt,j(ii) a Otherwise, not updating, and finding out the worst drosophila group with a given proportion;
b7: the worst drosophila melanogaster population is searched again according to the random value of the normal step length range in the B3, the corresponding smell value is calculated, the optimal drosophila melanogaster individual in the search is found, if the smell value is larger than bestsell, bestsell and the current optimal position H are updatedopt,j(ii) a Otherwise, the updating is not carried out.
B8: judging whether the maximum search times T is met, if so, stopping searching, and returning the searched optimal objective function solution; otherwise, the execution is repeated from B2 until the termination condition is satisfied.
In this embodiment, the preprocessing of the data in steps S1 and S4 includes missing value padding and feature value normalization; in the process of carrying out characteristic value standardization, min-max standardization is adopted to carry out pretreatment on characteristic data, and the value range of the treated characteristic is between [0, 1 ]. Given the data:
wherein l is the number of samples, and d is the characteristic number of the samples; .
The minimum and maximum values for each column of features in matrix X use vector XtminAnd xtmaxIt is shown that,
xtmin={x1min,x2min,...,xdmin}
xtmax={x1max,x2max,...,xdmax}
each sample xt∈XAllThe following normalization is performed in order to perform,
wherein x istpIs xtAnd (5) normalizing the feature vector.
In this embodiment, the specific steps of step S2 are as follows:
c1: given the training set D of the fuzzy support vector machine:
D={(x1,y1,s1),...,(xi,yi,si),...,(xl,yl,sl)},
wherein x isi∈RdD is the feature dimension of each sample, and l is the number of samples; y isiFor each sample category, indicating normal or abnormal; siRepresenting the fuzzy membership degree of each sample as the probability that the corresponding sample belongs to a certain class, and classifying the hyperplane by the FSVM as follows;
wherein z isiThe characteristic value of the original characteristic of the sample after being processed by the kernel function corresponding to the FSVM, namely RdThe feature vector in (1) is passed through a mapping functionA conversion to the feature space z is made,w is a normal vector of the hyperplane, and determines the direction of the hyperplane; b is a displacement term, and determines the distance between the hyperplane and the origin; epsiloniIs a relaxation variable which functions to reduce the influence of outliers on the classification hyperplane; c is a penalty coefficient, has great influence on the final classification capability of the model and is a parameter needing key search;
c2: and establishing an anomaly detection model based on the FSVM.
Wherein, aiIs a Lagrange factor; k (·,. cndot.) is a kernel function, K (x)i,xj)=zi·zj。
In this embodiment, the steps a2 and A3 include the following substeps:
d1: initializing relevant parameters such as iteration times and step length adjustment factors of the IFOA, and selecting an FSVM kernel function (selecting a Gaussian kernel for a nonlinear problem) according to the data set condition;
d2: searching for the FSVM hyperparameters C and g (the kernel function is a Gaussian kernel) by improving a drosophila algorithm, inputting the searched hyperparameters into the FSVM algorithm, performing cross validation on a training set, and combining a selected error function (fitness function) to obtain the optimal odor concentration smlBest of the group of parameters corresponding to the iterative population of the round;
d3: and continuously iterating to find optimal parameters C and g according to an updating strategy of the drosophila algorithm, and further solving w and b of the FSVM.
In this embodiment, in the step S5, a formula is adopted when the sample is detected:
advantageous effects
(1) The fuzzy support vector machine is applied to the field of wireless sensor network anomaly detection, and compared with an anomaly detection algorithm based on a support vector, the model has relatively strong robustness.
(2) The fruit fly algorithm is improved, and the advancing step length of the fruit flies can be intelligently adjusted according to the iteration times; dividing the fruit flies into two subgroups, and executing search strategies with different step length ranges in the early stage and the later stage of search; a certain proportion of fruit flies are selected in each iteration process for re-searching, so that the searching diversity of the fruit flies is enriched, and the searching capability of the fruit fly algorithm is improved.
(3) The invention adopts the improved drosophila algorithm to search the hyperparameters of the fuzzy support vector machine, and improves the abnormal detection accuracy of the fuzzy support vector machine.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (7)
1. A wireless sensor network anomaly detection method is characterized in that: based on a fruit fly optimization algorithm, the application of the fuzzy support vector machine to the field of wireless sensor network anomaly detection comprises the following steps:
a training stage:
s1: collecting sensor detection data, and preprocessing the data to form a training data set;
s2: establishing a wireless sensor network anomaly detection model based on the FSVM technology;
s3: establishing an IFOA-FSVM model, and performing anomaly detection training on a data set;
a detection stage:
s4: collecting sensor detection data, and preprocessing the data to form a sample to be detected;
s5: and inputting the sample to be detected into the IFOA-FSVM model for detection, and judging whether the sample to be detected is abnormal or not.
2. The method for detecting the abnormality in the wireless sensor network according to claim 1, wherein the step S3 of establishing the IFOA-FSVM model includes the following steps:
a1: improving a fruit fly algorithm;
a2: searching for the FSVM to obtain the optimal parameters by combining the training data set in the step S1 and the improved drosophila algorithm in the step A1;
a3 brings the optimal parameters back to the FSVM to complete the training of the anomaly detection model.
3. The method for detecting the abnormality of the wireless sensor network according to claim 2, wherein in the step A1, the improvement of the drosophila algorithm comprises the following steps:
b1: the quantity Num of the fruit flies in the initial search algorithm, the total iteration times T of the algorithm, and the current optimal position H of the fruit fly populationopt,jHistorical optimum odor concentration bestSmell of fruit fly population, step length adjustment parameters rho andextracting zz proportion drosophila individuals from each iteration to search again, wherein j belongs to n, and n is the variable number of the optimization objective function;
b2: calculating a step adjustment factor m (k);
B3: if k < T/2, the fruit fly population is divided into two subgroups group1 and group2, each subgroup executing a random step search strategy with a different range of advance steps. Wherein, group1 carries out search iteration according to the random value of the normal step range, and group2 carries out search iteration according to the random value of the large step range;
Hi,j=Hopt,j+RandomValue×m(k)
Hi,j=Hopt,j+3×RandomValue×m(k)
wherein i is the ith fruit fly individual in the population, and j is the jth variable in the objective function;
b4: and if k is less than T/2, modifying the search step range of partial fruit fly subgroups. Wherein, group1 still performs search iteration according to the random value of the normal step range in B3, and group2 performs search iteration according to the random value of the large step range;
Hi,j=Hopt,j+0.2×RandomValue×m(k)
b5: calculating the odor concentration value smell of each fruit fly according to the fitness function;
smelli=Function(Hi)
wherein, the Function is a fitness Function, namely a target Function to be optimized;
b6: sorting the odor concentration smell of the fruit flies, finding out fruit fly individuals (smllBest) with the largest smell value, namely current optimal fruit fly individuals, judging whether the smell is larger than the historical optimal odor concentration bestsell or not, and if so, updating the bestsell and the current optimal position Hopt,j(ii) a Otherwise, not updating, and finding out the worst drosophila group with a given proportion;
b7: the worst drosophila melanogaster population is searched again according to the random value of the normal step length range in the B3, the corresponding smell value is calculated, the optimal drosophila melanogaster individual in the search is found, if the smell value is larger than bestsell, bestsell and the current optimal position H are updatedopt,j(ii) a Otherwise, the updating is not carried out.
B8: judging whether the maximum search times T is met, if so, stopping searching, and returning the searched optimal objective function solution; otherwise, the execution is repeated from B2 until the termination condition is satisfied.
4. The method of claim 3, wherein the preprocessing of the data in steps S1 and S4 includes missing value padding and feature value normalization; in the process of carrying out characteristic value standardization, min-max standardization is adopted to carry out pretreatment on characteristic data, and the value range of the treated characteristic is between [0, 1 ]. Given the data:
wherein l is the number of samples, and d is the characteristic number of the samples; .
The minimum and maximum values for each column of features in matrix X use vector XtminAnd xtmaxIt is shown that,
xtmin={x1min,x2min,…,xdmin}
xtmax={x1max,x2max,…,xdmax}
each sample xt∈XAllThe following normalization is performed in order to perform,
wherein x istpIs xtAnd (5) normalizing the feature vector.
5. The method for detecting the abnormality of the wireless sensor network according to claim 4, wherein the step S2 includes the following steps:
c1: given the training set D of the fuzzy support vector machine:
D={(x1,y1,s1),...,(xi,yi,si),...,(xl,yl,sl)},
wherein x isi∈RdD is the feature dimension of each sample, and l is the number of samples; y isiFor each sample category, indicating normal or abnormal; siRepresenting the fuzzy membership degree of each sample as the probability that the corresponding sample belongs to a certain class, and classifying the hyperplane by the FSVM as follows;
wherein z isiThe characteristic value of the original characteristic of the sample after being processed by the kernel function corresponding to the FSVM, namely RdThe feature vector in (1) is passed through a mapping functionA conversion to the feature space z is made,w is a normal vector of the hyperplane, and determines the direction of the hyperplane; b is a displacement term, and determines the distance between the hyperplane and the origin; epsiloniIs a relaxation variable which functions to reduce the influence of outliers on the classification hyperplane; c is a penalty coefficient, has great influence on the final classification capability of the model and is a parameter needing key search;
c2: and establishing an anomaly detection model based on the FSVM.
Wherein, aiIs a Lagrange factor; k (·,. cndot.) is a kernel function, K (x)i,xj)=zi·zj。
6. The method for detecting the abnormality of the wireless sensor network according to claim 5, wherein the steps A2 and A3 comprise the following sub-steps:
d1: initializing relevant parameters such as iteration times, step length adjustment factors and the like of the IFOA, and selecting an FSVM kernel function according to the data set condition;
d2: searching for the FSVM hyperparameters C and g by improving a drosophila algorithm, inputting the searched hyperparameters into the FSVM algorithm, performing cross validation on a training set, and obtaining the optimal odor concentration smlBest of the group of parameters corresponding to the iterative population of the round by combining a selected error function;
d3: and continuously iterating to find optimal parameters C and g according to an updating strategy of the drosophila algorithm, and further solving w and b of the FSVM.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010040468.9A CN111242272A (en) | 2020-01-15 | 2020-01-15 | Wireless sensor network anomaly detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010040468.9A CN111242272A (en) | 2020-01-15 | 2020-01-15 | Wireless sensor network anomaly detection method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111242272A true CN111242272A (en) | 2020-06-05 |
Family
ID=70869009
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010040468.9A Pending CN111242272A (en) | 2020-01-15 | 2020-01-15 | Wireless sensor network anomaly detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111242272A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113433913A (en) * | 2021-07-06 | 2021-09-24 | 上海新氦类脑智能科技有限公司 | System monitoring model generation and monitoring method, processor chip and industrial system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2369529A1 (en) * | 2010-03-24 | 2011-09-28 | Alcatel Lucent | A method of detecting anomalies in a message exchange, corresponding computer program product, and data storage device therefor |
CN109002918A (en) * | 2018-07-16 | 2018-12-14 | 国网浙江省电力有限公司经济技术研究院 | Based on drosophila optimization algorithm-support vector machines electricity sales amount prediction technique |
CN109062180A (en) * | 2018-07-25 | 2018-12-21 | 国网江苏省电力有限公司检修分公司 | A kind of oil-immersed electric reactor method for diagnosing faults based on IFOA optimization SVM model |
CN110008983A (en) * | 2019-01-17 | 2019-07-12 | 西安交通大学 | A kind of net flow assorted method of the adaptive model based on distributed fuzzy support vector machine |
-
2020
- 2020-01-15 CN CN202010040468.9A patent/CN111242272A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2369529A1 (en) * | 2010-03-24 | 2011-09-28 | Alcatel Lucent | A method of detecting anomalies in a message exchange, corresponding computer program product, and data storage device therefor |
CN109002918A (en) * | 2018-07-16 | 2018-12-14 | 国网浙江省电力有限公司经济技术研究院 | Based on drosophila optimization algorithm-support vector machines electricity sales amount prediction technique |
CN109062180A (en) * | 2018-07-25 | 2018-12-21 | 国网江苏省电力有限公司检修分公司 | A kind of oil-immersed electric reactor method for diagnosing faults based on IFOA optimization SVM model |
CN110008983A (en) * | 2019-01-17 | 2019-07-12 | 西安交通大学 | A kind of net flow assorted method of the adaptive model based on distributed fuzzy support vector machine |
Non-Patent Citations (6)
Title |
---|
GANG LI等: "An Enhanced Fruit Fly Optimization Algorithm Based on Elitist Learning and Differential Perturbation Strategy" * |
王林等: "果蝇优化算法研究综述" * |
祝伟康: "面向高铁信号系统的入侵检测研究" * |
赵新等: "基于改进果蝇算法优化SVM的模拟电路故障诊断及对比分析" * |
雷李彪: "面向监测数据的无线传感器网络入侵检测算法研究" * |
顾成杰等: "一种具有特征有效度的模糊支持向量机" * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113433913A (en) * | 2021-07-06 | 2021-09-24 | 上海新氦类脑智能科技有限公司 | System monitoring model generation and monitoring method, processor chip and industrial system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106600140B (en) | Gas pipeline fault prediction early warning system and method based on improved support vector machine | |
CN108846512B (en) | Water quality prediction method based on preferential classification | |
CN107203810A (en) | A kind of precipitation Forecasting Methodology based on depth network | |
Luo et al. | Air pollutant prediction based on ARIMA-WOA-LSTM model | |
CN110020712B (en) | Optimized particle swarm BP network prediction method and system based on clustering | |
Shi et al. | A fast Pearson graph convolutional network combined with electronic nose to identify the origin of rice | |
CN106919980A (en) | A kind of increment type target identification system based on neuromere differentiation | |
Khashei et al. | Hybridization of the probabilistic neural networks with feed-forward neural networks for forecasting | |
CN112580050A (en) | XSS intrusion identification method based on semantic analysis and vectorization big data | |
CN116015967B (en) | Industrial Internet intrusion detection method based on improved whale algorithm optimization DELM | |
CN111222689A (en) | LSTM load prediction method, medium, and electronic device based on multi-scale temporal features | |
CN116484289A (en) | Carbon emission abnormal data detection method, terminal and storage medium | |
Hu et al. | An efficient Long Short-Term Memory model based on Laplacian Eigenmap in artificial neural networks | |
CN111242272A (en) | Wireless sensor network anomaly detection method | |
CN113344243B (en) | Wind speed prediction method and system for optimizing ELM (ELM) based on improved Harris eagle algorithm | |
Chen et al. | Anomaly detection for drinking water quality via deep biLSTM ensemble | |
Chen et al. | Short-term Wind Speed Forecasting Based on Singular Spectrum Analysis, Fuzzy C-Means Clustering, and Improved POABP | |
US11816556B1 (en) | Method for predicting air quality index (AQI) based on a fusion model | |
Wang et al. | An improved deep learning based intrusion detection method | |
Ji et al. | A network intrusion detection approach based on asymmetric convolutional autoencoder | |
Kye et al. | Hierarchical autoencoder for network intrusion detection | |
CN113852612B (en) | Network intrusion detection method based on random forest | |
Suresh et al. | Design an efficient disease monitoring system for paddy leaves based on big data mining | |
Shi et al. | Key Process Protection of High Dimensional Process Data in Complex Production. | |
CN117688504B (en) | Internet of things abnormality detection method and device based on graph structure learning |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200605 |