CN112488149A - Network security data classification method based on 1D-CNN feature reconstruction - Google Patents
Network security data classification method based on 1D-CNN feature reconstruction Download PDFInfo
- Publication number
- CN112488149A CN112488149A CN202011209655.1A CN202011209655A CN112488149A CN 112488149 A CN112488149 A CN 112488149A CN 202011209655 A CN202011209655 A CN 202011209655A CN 112488149 A CN112488149 A CN 112488149A
- Authority
- CN
- China
- Prior art keywords
- data
- cnn
- matrix
- model
- reconstruction
- 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
- 238000000034 method Methods 0.000 title claims abstract description 30
- 239000011159 matrix material Substances 0.000 claims abstract description 58
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000013145 classification model Methods 0.000 claims abstract description 27
- 238000011176 pooling Methods 0.000 claims abstract description 27
- 231100000279 safety data Toxicity 0.000 claims abstract description 19
- 230000009467 reduction Effects 0.000 claims abstract description 15
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000010801 machine learning Methods 0.000 claims abstract description 8
- 238000010276 construction Methods 0.000 claims abstract description 5
- 238000005457 optimization Methods 0.000 claims abstract description 5
- 238000012360 testing method Methods 0.000 claims description 33
- 230000006870 function Effects 0.000 claims description 31
- 230000004913 activation Effects 0.000 claims description 18
- 238000013507 mapping Methods 0.000 claims description 13
- 238000005070 sampling Methods 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 claims description 3
- 238000003066 decision tree Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims description 2
- 238000013527 convolutional neural network Methods 0.000 abstract description 59
- 238000013135 deep learning Methods 0.000 abstract description 5
- 238000001514 detection method Methods 0.000 abstract description 5
- 230000006399 behavior Effects 0.000 abstract 1
- 238000013136 deep learning model Methods 0.000 abstract 1
- 238000012545 processing Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000003042 antagnostic effect Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a network security data classification method based on 1D-CNN feature reconstruction, which comprises the steps of model construction and training optimization, and specifically comprises the steps of firstly constructing a 1D-CNN deep learning model by utilizing a correlation matrix among features, generating low-dimensional reconstruction features through convolution, pooling and fully-connected global convolution operation, and completing dimension reduction reconstruction of data. And then, constructing a safety data classification model by using a traditional shallow machine learning algorithm to realize detection of threat behaviors in the network safety big data. The feature reconstruction method based on the 1D-CNN can control the dimensionality of the reconstructed features, realize data dimensionality reduction, simplify the operation process of deep learning, improve the operation efficiency of a model, improve the relevance between the reconstructed features by utilizing the correlation between the features in a convolutional layer and enable the classification result to be more accurate.
Description
Technical Field
The invention relates to the field of analysis and modeling of network security big data, in particular to a network security data classification method based on 1D-CNN feature reconstruction and dimension reduction.
Background
Various network attack modes exist in a network space, such as malicious codes, fishing mails and websites, traffic attacks, loopholes and the like, the attacks not only cause huge economic loss, but also threaten national security and social stability, and therefore, the detection of the network threat is necessary. In the detection process, a large amount of network data, such as malicious software, phishing mails, network traffic, system logs and the like, need to be collected, and it is difficult to obtain a good effect by constructing a traditional machine learning model to analyze the data. With the continuous development of deep learning and artificial intelligence calculation and the successful application of the deep learning technology in the aspects of computer vision, natural language processing and the like, the deep learning technology is applied to the aspect of network space threat detection, and the method is an effective method for realizing network security data classification and improving network threat detection.
The deep learning technique includes various algorithms such as a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), a antagonistic neural network (GAN), etc., wherein the CNN algorithm learns and re-characterizes data features using nonlinear operations of convolutional layers, and reduces the dimensionality of the data features using compression operations of pooling layers, and thus, the CNN algorithm can be used to process network security data. The CNN algorithm can construct a 1D-CNN model and a 2D-CNN model according to different types of processing data. For example, when processing sequence signal data and natural language, a 1D-CNN model is constructed, and when processing image and video data, a 2D-CNN model is constructed. When the 2D-CNN model is used for classifying the network security data, firstly, the data needs to be converted into an image format and then processed, and the defects of complex operation process and large operation amount exist in the processing process; meanwhile, the problem of poor correlation between features also exists by adopting random numbers in the convolution kernel, so that the overall classification precision is low.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a network security data classification method based on 1D-CNN feature reconstruction, which is characterized in that a 1D-CNN deep layer model is constructed according to the one-dimensional characteristics of a security data sample, the characteristics of original security data are reconstructed by utilizing the nonlinear characteristics of neurons, the random number in the convolution kernel of a convolutional layer is changed into a feature correlation coefficient, the correlation among the reconstruction features is improved, the feature space dimension is reduced, and the classification precision is improved.
A network security data classification method based on 1D-CNN feature reconstruction specifically comprises the following steps:
the method comprises the following steps: building a data set
Performing One-hot coding on original safety data to construct a test set X with the size of N X DtestCalculating a characteristic correlation matrix R of the training set X with the training set X, wherein N is the sample number of the data set, and D represents the dimension of the data set; y is a real class label set corresponding to the training set X,
step two: construction of 1D-CNN Algorithm model
And constructing a 1D-CNN algorithm model for performing dimension reduction and reconstruction on the input data set, wherein the model comprises an input layer, L convolutional layers, L pooling layers, a full connection layer and a Softmax layer.
And the input layer is used for inputting the training set X.
Setting M convolution kernels for each convolution layer according to the characteristic correlation matrix R obtained by calculation in the step one, carrying out convolution operation on data input into the convolution layers and the M convolution kernels to obtain M mapping characteristic matrixes, and obtaining M nonlinear mapping characteristic matrixes S through a nonlinear activation function ReLU (·)l(ii) a Wherein the output of the first convolutional layer is subjected to a non-linear activation function ReLU (·)) Then is connected with the first pooling layer; when l > 1, the convolutional layer input is connected to the output of the input layer, and when l > 1, the convolutional layer input is connected to the output of the l-1 th pooling layer via the nonlinear activation function ReLU ().
The L pooling layers adopt a maximum pooling mode to perform nonlinear mapping on the characteristic matrix S output by the convolutional layerlPerforming down-sampling operation, and performing nonlinear activation function ReLU (-) to obtain pooled nonlinear mapping characteristic matrix Tl(ii) a Wherein the output of the L-th pooling layer is connected to the input of the fully-connected layer via a non-linear activation function ReLU (-).
The full connection layer adopts the global convolution with convolution kernel of M multiplied by D; outputting T to the L-th pooling layerLAfter the nonlinear transformation of the feature space is realized through global convolution, a reconstruction matrix X 'is obtained through a nonlinear activation function ReLU (·)'f。
And the input of the Softmax layer is connected with the output of the full connection layer.
Step three, training and optimizing the 1D-CNN algorithm model
And G, reconstructing an output reconstruction matrix X 'of the 1D-CNN algorithm model obtained in the step two'fInputting the prediction reconstruction feature matrix X 'into the softmax layer'fComparing the predicted sample class label Y' with the real class label Y, deducing a Loss function Loss of the 1D-CNN model, and circularly training the 1D-CNN algorithm model for F times.
In the training iteration process of the 1D-CNN algorithm model, an Adam optimization function is adopted to optimize the Loss function Loss of the 1D-CNN algorithm model to the minimum value. And when the training of the 1D-CNN algorithm model is finished, outputting the 1D-CNN algorithm model obtained from the full connection layer, and obtaining a reconstruction characteristic matrix X ' with the size of N multiplied by D ' after the training of the 1D-CNN algorithm model is finished, wherein D ' is less than or equal to D, and the dimension of the reconstruction matrix is lower than that of the original data matrix, namely the 1D-CNN algorithm model realizes the dimension reduction of the reconstruction matrix to the original matrix.
Preferably, the number of times of cyclic training of the 1D-CNN algorithm model is F1000.
Step four, constructing and training a safety data classification model
And (4) constructing a safety data classification model based on a shallow machine learning algorithm, and inputting the reconstructed feature matrix X 'obtained in the step three into the safety data classification model to obtain a predicted sample class label Y'.
And setting a performance target, comparing the predicted sample class label Y' with the real class label Y, and calculating the performance of the classification model according to the confusion matrix evaluation index. When the performance of the classification model does not reach the preset target, returning to the step three, and retraining and optimizing the 1D-CNN algorithm model; and when the performance of the classification model reaches a preset target, the next step is carried out.
Preferably, the shallow machine learning algorithm is a support vector machine, a decision tree or naive bayes.
Preferably, the performance of the classification model comprises the accuracy rate, precision rate and call-back rate of the classification model
Step five, safety data classification
Test data set XtestInputting the data into the 1D-CNN algorithm model trained and optimized in the third step to obtain a test data set XtestOf reconstructed feature matrix X'testThen X'testInputting the data into the safety data classification model trained and optimized in the fourth step to obtain a test data set XtestOf prediction category matrix Y'testAnd realizing the prediction of the data classification labels in the network security data set, namely realizing the classification of the network security data.
The invention has the following beneficial effects:
(1) the method comprises the steps that a 1D-CNN algorithm model is constructed according to a 1D structure of input data, wherein a convolution kernel and pooling sampling both adopt 1D matrix patterns, compared with a 2D-CNN model in the prior art, the operation of a 1D-CNN network is simplified, the operation amount is correspondingly reduced, and the operation efficiency of the model can be improved;
(2) in the convolution operation of the 1D-CNN algorithm model convolution layer, a convolution kernel adopts a characteristic correlation matrix instead of a random number, and input data and the correlation convolution kernel are subjected to convolution operation, so that the obtained reconstruction characteristics have better correlation, and the accuracy of a classification result can be improved;
(3) through the convolution operation of the 1D-CNN, the pooling operation and the convolution operation of the full connection layer, the dimensionality of the reconstruction characteristics can be controlled, when the dimensionality is lower than the dimensionality of the original data, the dimensionality reduction of the data is achieved, and the performance of the reconstruction data based on shallow machine learning classification is improved.
Drawings
FIG. 1 is a flowchart of a network security data classification method based on 1D-CNN feature reconstruction.
Detailed Description
The invention is further explained below with reference to the drawings;
the raw security data used in this embodiment is derived from the classic data set KDDCUP99 and the latest data set cic native 2020 in the field of network security.
As shown in fig. 1, the network security data classification method based on 1D-CNN feature reconstruction includes model construction, training optimization and data classification, and the specific process is as follows:
the method comprises the following steps: building a data set
Performing One-hot coding on the original safety data to construct a test set X of size N X DtestAnd training set X, X ═ X (X)1,x2,…,xn,…xN) Where N is the number of samples in the data set, D represents the dimension of the data set, and a feature correlation matrix R of the training set X is calculated from the covariance matrix, where R is (R)1,r2,…,rn,…rN) (ii) a And Y is a real class label set corresponding to the training set X.
Step two: construction of 1D-CNN Algorithm model
And constructing a 1D-CNN algorithm model for performing dimension reduction and reconstruction on the input data set, wherein the model comprises an input layer, L convolutional layers, L pooling layers, a full connection layer and a Softmax layer.
And the input layer is used for inputting the training set X.
The output of the L convolutional layers is connected with the L pooling layer after passing through a nonlinear activation function ReLU (-); when l is 1, the input of the convolution layer is the same as the output of the input layerWhen l > 1, the convolutional layer input is connected to the output of the l-1 pooling layer via the nonlinear activation function ReLU (-). Setting M convolution kernels for each convolution layer, wherein the mth convolution kernel is a 1D matrix randomly extracted from the row matrix of the characteristic correlation matrix R,performing convolution operation on data input into the convolution layer and M convolution kernels to obtain M mapping characteristic matrixes, and obtaining M nonlinear mapping characteristic matrixes S through a nonlinear activation function ReLU (·)l(ii) a The mth nonlinear mapping feature matrix obtained from the first convolutional layerComprises the following steps:
And the output of the L pooling layer is connected with the input of the full connection layer after passing through a nonlinear activation function ReLU (-). The pooling layer adopts a maximum pooling mode and uses a pooling matrixMapping the characteristic matrix of the mth nonlinear of the convolutional layer outputPerforming down-sampling operation, and performing nonlinear activation function ReLU (-) to obtain the mth pooled nonlinear mapping feature matrix
Where maxporoling (. circle.) represents the maximum pooling function.
The full connection layer adopts an MXD convolution kernel KM×DCarrying out global convolution operation to realize nonlinear change of the characteristic space; outputting T to the L-th pooling layerLObtaining a reconstruction matrix X 'through a nonlinear activation function ReLU (-) after global convolution'f:
And the input of the Softmax layer is connected with the output of the full connection layer.
Step three, training and optimizing the 1D-CNN algorithm model
And G, reconstructing an output reconstruction matrix X 'of the 1D-CNN algorithm model obtained in the step two'fInputting the prediction reconstruction feature matrix X 'into the softmax layer'fSample class label Y' in (1):
Y′=softmax(X′f)
comparing the predicted sample class label Y' with the real class label Y, defining the Loss function Loss of the 1D-CNN algorithm model based on the cross entropy Loss function,
Loss=crossentropy(Y,Y′)
wherein, crossentryprop (-) represents the cross-entropy loss function.
The 1D-CNN algorithm model is trained 1000 times in a circulating way.
In the training iteration process of the 1D-CNN algorithm model, an Adam optimization function is adopted to optimize the Loss function Loss of the 1D-CNN algorithm model to the minimum value. And when the training of the 1D-CNN algorithm model is finished, outputting the 1D-CNN algorithm model obtained from the full connection layer, and obtaining a reconstruction characteristic matrix X ' with the size of N multiplied by D ' after the training of the 1D-CNN algorithm model is finished, wherein D ' is less than or equal to D, and the dimension of the reconstruction matrix is lower than that of the original data matrix, namely the 1D-CNN algorithm model realizes the dimension reduction of the reconstruction matrix to the original matrix.
And (3) carrying out dimension reduction and reconstruction on the data sets KDDCUP99 and CICMalDroid2020 through a 1D-CNN algorithm model to obtain a feature set with a proportion of 15%.
Step four, constructing and training a safety data classification model
And (4) constructing a safety data classification model of the shallow machine learning algorithm based on the support vector machine, and inputting the reconstructed feature matrix X 'obtained in the step three into the safety data classification model to obtain a predicted sample class label Y'.
And setting a performance target, comparing the predicted sample class label Y' with the real class label Y, and calculating the performance of the classification model according to the confusion matrix evaluation index, wherein the performance comprises the accuracy, precision and call-back rate of the classification model. When the performance of the classification model does not reach the preset target, returning to the step three, and retraining and optimizing the 1D-CNN algorithm model; and when the performance of the classification model reaches a preset target, the next step is carried out.
Step five, safety data classification
Test data set XtestInputting the data into the 1D-CNN algorithm model trained and optimized in the third step to obtain a test data set XtestOf reconstructed feature matrix X'testThen X'testInputting the data into the safety data classification model trained and optimized in the fourth step to obtain a test data set XtestOf prediction category matrix Y'testAnd realizing the prediction of the data classification labels in the network security data set, namely realizing the classification of the network security data.
Step six, experimental results
After the steps, the classification accuracy of the data sets KDDCUP99 and CICMalDroid2020 after dimension reduction and reconstruction into the feature set with the proportion of 15% is calculated, and compared with the classification result of the feature set with the proportion of 15% through dimension reduction and reconstruction of the 1D-CNN algorithm model, but the convolution kernel is a random number, and the classification result directly input into the trained security data classification model without dimension reduction and reconstruction of the 1D-CNN algorithm model, the comparison result is shown in the following table:
comparing the data in the second, third and fourth columns of the table, the classification result reconstructed by the 1D-CNN algorithm model is better than the classification result reconstructed without dimensionality reduction; comparing the data in the second column and the data in the third column of the table, it can be seen that the feature correlation matrix is used as the initialization weight of the 1D-CNN algorithm, i.e. the classification result obtained by setting the convolution kernel according to the feature correlation matrix is better than the classification result without using the feature correlation.
Claims (4)
1. A network security data classification method based on 1D-CNN feature reconstruction is characterized in that: the method specifically comprises the following steps:
the method comprises the following steps: constructing a training set
Performing One-hot coding on original safety data to construct a test set X with the size of N X DtestCalculating a characteristic correlation matrix R of the training set X with the training set X, wherein N is the sample number of the data set, and D represents the dimension of the data set; y is a real category label set corresponding to the training set X;
step two: construction of 1D-CNN Algorithm model
Constructing a 1D-CNN algorithm model for performing dimension reduction and reconstruction on an input data set, wherein the model comprises an input layer, L convolutional layers, L pooling layers, a full connection layer and a Softmax layer;
the input layer is used for inputting a training set X;
setting M convolution kernels for each convolution layer according to the characteristic correlation matrix R obtained by calculation in the step one, carrying out convolution operation on data input into the convolution layers and the M convolution kernels to obtain M mapping characteristic matrixes, and obtaining M nonlinear mapping characteristic matrixes S through a nonlinear activation function ReLU (·)l(ii) a Wherein the output of the first convolutional layer is connected with the first pooling layer after passing through a nonlinear activation function ReLU (-); when l is 1, the input of the convolutional layer is connected with the output of the input layer, and when l is more than 1, the input of the convolutional layer is connected with the output of the l-1 st pooling layer passing through the nonlinear activation function ReLU (·);
the L pooling layersThe nonlinear mapping characteristic matrix S of the convolutional layer output is processed by adopting a maximum pooling modelPerforming down-sampling operation, and performing nonlinear activation function ReLU (-) to obtain pooled nonlinear mapping characteristic matrix Tl(ii) a Wherein the output of the L-th pooling layer is connected with the input of the full-connection layer after passing through a nonlinear activation function ReLU (-);
the full connection layer adopts the global convolution with convolution kernel of M multiplied by D; outputting T to the L-th pooling layerLAfter the nonlinear transformation of the feature space is realized through global convolution, a reconstruction matrix X 'is obtained through a nonlinear activation function ReLU (·)'f;
The input of the Softmax layer is connected with the output of the full connection layer;
step three, training and optimizing the 1D-CNN algorithm model
And G, reconstructing an output reconstruction matrix X 'of the 1D-CNN algorithm model obtained in the step two'fInputting the prediction reconstruction feature matrix X 'into the softmax layer'fComparing the predicted sample class label Y' with a real class label Y, deducing a Loss function Loss of the 1D-CNN model, and circularly training the 1D-CNN algorithm model for F times;
in the training iteration process of the 1D-CNN algorithm model, optimizing the Loss function Loss of the 1D-CNN algorithm model to the minimum value by adopting an Adam optimization function; when the training of the 1D-CNN algorithm model is finished, outputting the 1D-CNN algorithm model obtained from the full connection layer, and obtaining a reconstruction characteristic matrix X ' with the size of N multiplied by D ' after the training of the 1D-CNN algorithm model is finished, wherein D ' is less than or equal to D, and the dimension of the reconstruction matrix is lower than that of the original data matrix, namely the 1D-CNN algorithm model realizes the dimension reduction of the reconstruction matrix to the original matrix;
step four, constructing and training a safety data classification model
Constructing a safety data classification model based on a shallow machine learning algorithm, and inputting the reconstructed feature matrix X 'obtained in the step three into the safety data classification model to obtain a predicted sample class label Y';
setting a performance target, comparing the predicted sample class label Y' with the real class label Y, and calculating the performance of the classification model according to the confusion matrix evaluation index; when the performance of the classification model does not reach the preset target, returning to the step three, and retraining and optimizing the 1D-CNN algorithm model; when the performance of the classifier reaches a preset target, entering the next step;
step five, safety data classification
Test data set XtestInputting the data into the 1D-CNN algorithm model trained and optimized in the third step to obtain a test data set XtestOf reconstructed feature matrix X'testThen X'testInputting the data into the safety data classification model trained and optimized in the fourth step to obtain a test data set XtestOf prediction category matrix Y'testAnd realizing the prediction of the data classification labels in the network security data set, namely realizing the classification of the network security data.
2. The method for classifying network security data based on 1D-CNN feature reconstruction as claimed in claim 1, wherein: the number of times of circularly training the 1D-CNN algorithm model is F-1000.
3. The method for classifying network security data based on 1D-CNN feature reconstruction as claimed in claim 1, wherein: the shallow machine learning algorithm is a support vector machine, a decision tree or naive Bayes.
4. The method for classifying network security data based on 1D-CNN feature reconstruction as claimed in claim 1, wherein: the performance of the classification model comprises the accuracy rate, the precision rate and the call-back rate of the classification model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011209655.1A CN112488149B (en) | 2020-11-03 | 2020-11-03 | Network security data classification method based on 1D-CNN feature reconstruction |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011209655.1A CN112488149B (en) | 2020-11-03 | 2020-11-03 | Network security data classification method based on 1D-CNN feature reconstruction |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112488149A true CN112488149A (en) | 2021-03-12 |
CN112488149B CN112488149B (en) | 2024-02-27 |
Family
ID=74927762
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011209655.1A Active CN112488149B (en) | 2020-11-03 | 2020-11-03 | Network security data classification method based on 1D-CNN feature reconstruction |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112488149B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113239949A (en) * | 2021-03-15 | 2021-08-10 | 杭州电子科技大学 | Data reconstruction method based on 1D packet convolutional neural network |
CN114781875A (en) * | 2022-04-25 | 2022-07-22 | 四川中电启明星信息技术有限公司 | Micro-grid economic operation state evaluation method based on deep convolutional network |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108388917A (en) * | 2018-02-26 | 2018-08-10 | 东北大学 | A kind of hyperspectral image classification method based on improvement deep learning model |
US20190236411A1 (en) * | 2016-09-14 | 2019-08-01 | Konica Minolta Laboratory U.S.A., Inc. | Method and system for multi-scale cell image segmentation using multiple parallel convolutional neural networks |
CN110619049A (en) * | 2019-09-25 | 2019-12-27 | 北京工业大学 | Message anomaly detection method based on deep learning |
CN111245820A (en) * | 2020-01-08 | 2020-06-05 | 北京工业大学 | Phishing website detection method based on deep learning |
-
2020
- 2020-11-03 CN CN202011209655.1A patent/CN112488149B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190236411A1 (en) * | 2016-09-14 | 2019-08-01 | Konica Minolta Laboratory U.S.A., Inc. | Method and system for multi-scale cell image segmentation using multiple parallel convolutional neural networks |
CN108388917A (en) * | 2018-02-26 | 2018-08-10 | 东北大学 | A kind of hyperspectral image classification method based on improvement deep learning model |
CN110619049A (en) * | 2019-09-25 | 2019-12-27 | 北京工业大学 | Message anomaly detection method based on deep learning |
CN111245820A (en) * | 2020-01-08 | 2020-06-05 | 北京工业大学 | Phishing website detection method based on deep learning |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113239949A (en) * | 2021-03-15 | 2021-08-10 | 杭州电子科技大学 | Data reconstruction method based on 1D packet convolutional neural network |
CN114781875A (en) * | 2022-04-25 | 2022-07-22 | 四川中电启明星信息技术有限公司 | Micro-grid economic operation state evaluation method based on deep convolutional network |
CN114781875B (en) * | 2022-04-25 | 2024-08-02 | 四川中电启明星信息技术有限公司 | Micro-grid economic operation state evaluation method based on deep convolution network |
Also Published As
Publication number | Publication date |
---|---|
CN112488149B (en) | 2024-02-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109101552B (en) | Phishing website URL detection method based on deep learning | |
CN111753881B (en) | Concept sensitivity-based quantitative recognition defending method against attacks | |
CN110048827B (en) | Class template attack method based on deep learning convolutional neural network | |
CN110287983B (en) | Single-classifier anomaly detection method based on maximum correlation entropy deep neural network | |
CN111785329B (en) | Single-cell RNA sequencing clustering method based on countermeasure automatic encoder | |
CN113806746B (en) | Malicious code detection method based on improved CNN (CNN) network | |
CN110263227A (en) | Clique based on figure neural network finds method and system | |
CN114120041B (en) | Small sample classification method based on double-countermeasure variable self-encoder | |
CN112488149B (en) | Network security data classification method based on 1D-CNN feature reconstruction | |
CN113239949A (en) | Data reconstruction method based on 1D packet convolutional neural network | |
CN115422369B (en) | Knowledge graph completion method and device based on improved TextRank | |
CN113901448A (en) | Intrusion detection method based on convolutional neural network and lightweight gradient elevator | |
CN113157919A (en) | Sentence text aspect level emotion classification method and system | |
CN112487406A (en) | Network behavior analysis method based on machine learning | |
CN113283400B (en) | Skeleton action identification method based on selective hypergraph convolutional network | |
Adaïmé et al. | Deep learning approaches to the phylogenetic placement of extinct pollen morphotypes | |
CN113868647A (en) | Network unknown threat detection method based on feature extension CNN | |
CN113420289A (en) | Hidden poisoning attack defense method and device for deep learning model | |
CN115937567B (en) | Image classification method based on wavelet scattering network and ViT | |
Kavzoglu et al. | The use of feature selection techniques in the context of artificial neural networks | |
CN116977694A (en) | Hyperspectral countermeasure sample defense method based on invariant feature extraction | |
CN115310083A (en) | Malicious software robust identification method based on contrast learning | |
CN114741532A (en) | Knowledge graph completion method and system based on multi-scale decentralized dynamic routing | |
CN113283520A (en) | Member reasoning attack-oriented depth model privacy protection method and device based on feature enhancement | |
CN111191717A (en) | Black box confrontation sample generation algorithm based on hidden space clustering |
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 |