CN110287439A - A kind of network behavior method for detecting abnormality based on LSTM - Google Patents
A kind of network behavior method for detecting abnormality based on LSTM Download PDFInfo
- Publication number
- CN110287439A CN110287439A CN201910566453.3A CN201910566453A CN110287439A CN 110287439 A CN110287439 A CN 110287439A CN 201910566453 A CN201910566453 A CN 201910566453A CN 110287439 A CN110287439 A CN 110287439A
- Authority
- CN
- China
- Prior art keywords
- behavior
- network
- user
- sequence
- behavior sequence
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3438—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment monitoring of user actions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- 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
- H04L41/142—Network analysis or design using statistical or mathematical methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
- H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
- H04L63/1425—Traffic logging, e.g. anomaly detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Signal Processing (AREA)
- Mathematical Physics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Algebra (AREA)
- Fuzzy Systems (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a kind of network behavior method for detecting abnormality based on LSTM, collection network data on flows and user behavior sequence is converted into according to the definition of user behavior first, then in view of the otherness between network user's subject behavior mode, therefore the present invention classifies to user behavior sequence according to k- central point algorithm.Then, using sorted behavior sequence data as the input of LSTM shot and long term memory network, neural network model is trained in conjunction with Attention mechanism.The model completed finally by training predicts to determine its intensity of anomaly behavior sequence to be detected.The angle of subordinate act of the present invention, which is set out, handles network flow data, it can fully consider the incidence relation between internal factor, and it establishes network behavior mode and distinguishes the behavior of user, then traditional network abnormality detection is broken through using the artificial method for extracting feature, exception information is distinguished to the development fitting effect of large scale network behavior sequence data flow in conjunction with LSTM shot and long term memory network, significantly improves the precision and efficiency of Network anomaly detection.
Description
Technical field
The invention belongs to technical field of network security, more specifically, it is different to be related to a kind of network behavior based on LSTM
Normal detection method.
Background technique
With the rapid development of global network information industry, various data interactions are more and more frequent, computer increasingly
Today of people's life is incorporated, people also increasingly be unable to do without network.The especially rise of mobile Internet, even more handle in recent years
People have pulled in the Network Information epoch.However in increasingly complicated network environment, for network entity attack increasingly
Frequently, attack pattern also increasingly develops towards diversification with the direction complicated, these network attacks gently then influence to be attacked
The service quality of person, it is heavy then cause information leakage, network paralysis, cause huge economic loss.So how by a kind of high
It imitates and accurately mode detects Network anomalous behaviors, be all considerable for network service provider and user.
Network security system have passed through the two generation systems development of traditional " non-black i.e. white ", be had evolved at present by looking into
The mode of behavior is looked for judge the behavior of user with the presence or absence of abnormal.First generation network security system is the side by " blacklist "
Formula to carry out killing to viral wooden horse.Second generation network security system is the behavior that user is judged using the mechanism of " white list "
It is whether credible.Third generation networks security system is then with technological means such as big data, artificial intelligence, machine learning to user
Behavioral data is acquired, analyzes and studies and judges, and carries out early warning to the abnormal behaviour of user.
LSTM, i.e. Long Short-Term Memory shot and long term memory network are a kind of based on Recognition with Recurrent Neural Network RNN
Time recurrent neural network, suitable for time series analysis be fitted.LSTM algorithm machine translation, sentiment analysis,
Multiple artificial intelligence fields such as image analysis, documentation summary, speech recognition and recommender system, which have, to be widely applied, be it is a kind of at
Ripe machine learning algorithm, but in network behavior abnormality detection field using shot and long term memory network still at an early stage.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, propose a kind of method for detecting Network anomalous behaviors,
By classifying to user behavior pattern, and LSTM neural network model and Attention mechanism are combined, can be obviously improved
To the accuracy rate of Network anomalous behaviors detection.
For achieving the above object, the present invention is based on the network behavior method for detecting abnormality of LSTM, which is characterized in that packet
Include following steps:
(1), network flow data is collected and cleaning arranges
Data on flows for abnormality detection is collected generally by the distributed agent for being deployed in each host terminal,
The data on flows that each distributed agent is collected upward first-level agent's convergence again.Later further according to analysis demand to the stream of collection
Amount data are cleaned.Then it is directed to current network data, the definition of clear user behavior in a network, and to network number again
Each user crawl in is converted to user behavior track sets.
(2), behavior sequence is classified
It for the action trail sequence of all users, is clustered according to k- central point algorithm, is classified as k inhomogeneity
Other behavior sequence.User user i.e. to be detected for needing to carry out network behavior abnormality detection, by its behavior sequence and k
The cluster central point of different classes of behavior sequence carries out similarity measurement, and the one kind for taking its most like is as user behavior to be detected
The classification of sequence.
(3), LSTM neural network model is established
Using k class behavior sequence data obtained in step 2 as the input data of k LSTM neural network, in conjunction with
Attention mechanism is trained LSTM neural network, obtains k LSTM neural network model of training completion.Wherein, often
A neural network model corresponds to a kind of user behavior classification.
(4), network behavior abnormality detection
To user to be detected, using its behavior sequence as the input of the LSTM neural network model of corresponding classification, and by mould
Intensity of anomaly of the difference as network behavior between the behavior prediction of type and true behavior.
The object of the present invention is achieved like this.
A kind of network behavior method for detecting abnormality based on LSTM of the present invention, first collection network data on flows and according to
The definition of family behavior is arranged as user behavior track sets.Simultaneously, it is contemplated that between network user's subject behavior mode
Otherness, therefore the present invention classifies to user behavior sequence by k- central point algorithm.Then, by sorted behavior sequence
Input of the column data as LSTM shot and long term memory network, is trained neural network model in conjunction with Attention mechanism.Most
Behavior sequence to be detected is predicted by the model that training is completed to determine its intensity of anomaly afterwards.The angle of subordinate act of the present invention
Degree, which sets out, handles network flow data, can fully consider the incidence relation between internal factor, and establish network row
The behavior of user is distinguished for mode, then breaks through traditional network abnormality detection using the artificial method for extracting feature, knot
It closes LSTM shot and long term memory network and exception information is distinguished to the development fitting effect of large scale network behavior sequence data flow, show
Write the precision and efficiency for improving Network anomaly detection.
Detailed description of the invention
Fig. 1 is a kind of a kind of specific embodiment process of the network behavior method for detecting abnormality based on LSTM of the present invention
Figure；
Fig. 2 is the schematic diagram of user behavior sequence in the present invention；
Fig. 3 is a kind of LSTM model schematic of the network behavior method for detecting abnormality based on LSTM of the present invention.
Specific embodiment
A specific embodiment of the invention is described with reference to the accompanying drawing, preferably so as to those skilled in the art
Understand the present invention.Requiring particular attention is that in the following description, when known function and the detailed description of design perhaps
When can desalinate main contents of the invention, these descriptions will be ignored herein.
Fig. 1 is a kind of a kind of specific embodiment process of the network behavior method for detecting abnormality based on LSTM of the present invention
Figure.
In the present embodiment, as shown in Figure 1, a kind of network behavior method for detecting abnormality based on LSTM of the present invention include with
Lower step:
S1: network flow data is collected and cleaning arranges
The log information that network flow data, that is, user records when accessing specific network entity, such as access time, IP
Address, source port, destination port and operational order etc..
According to the specific network entity that user accesses, being specifically defined for user behavior sequence is specified, by network flow data
Arranging is user behavior sequence.User behavior sequence can also be called " user behavior based on time series ", be in certain a period of time
Between in section, being engaged in certain movable each walking according to the people that chronological order records is.
Such as shown in Fig. 2, on website, in a period of time, a user is from entering website to during leaving website
Each walking record for being, be recorded as a user behavior sequence.
S2: behavior sequence classification
In general, user is usually that a series of movable behavioral agents are carried out on network.Because of everyone identity
And the difference of living habit, the behavior pattern between user is discrepant, so needing to be divided user behavior sequence
Class improves the accuracy of abnormality detection with this.
Firstly the need of to carrying out similarity measurement between behavior sequence, in order in the description present invention that is more clear
User behavior sequence similarity measure provides following several definition:
Definition one: subsequence.If given behavior sequence X=(x1,x2,…xm), then another sequence Z=(Z1,Z2,…Zm) be
The subsequence of X refers to that there are a strictly increasing subscript sequence (i1,i2,…im), so that having for all j=1 ..., kIf being designated as 1 under starting.
Definition two: common subsequence.There are given two behavior sequences X and Y, when another sequence Z is both the subsequence of X
It is the subsequence of Y again, then Z is sequence X and the common subsequence of Y.Wherein the longest sequence of Z is the public sub- sequence of longest of X and Y
Column.
After the definition for having subsequence and common subsequence, so that it may find out two user behaviors by dynamic programming algorithm
Longest common subsequence between sequence.X=(x is saved with c [i] [j]1,x2,…xm) and Y=(y1,y2,…yn) longest it is public
Subsequence altogether, then:
It is possible thereby to acquire the longest common subsequence between two user behavior sequences.
After the behavior sequence for having each user, so that it may indicate two use by calculating the similarity between user
Similarity degree and relationship between the behavior of family.User behavior sequence similarity is realized by user behavior pattern similarity.In order to
The similarity of behavior pattern is calculated, first calculating behavior pattern distance.The calculation method of behavior pattern distance is described below.
Behavior pattern needs to calculate the distance between behavior sequence in calculating process, therefore defines behavior sequence first
The distance between.In order to make the common subsequence of two behavior sequences is longer, similarity is bigger, between two behavior sequences away from
It is from smaller, the distance definition between behavior sequence is as follows:
Wherein | X | and | Y | indicate the length of behavior sequence X and behavior sequence Y, lcs (X, Y) is two behavior sequences X and Y
Longest common subsequence.
In fact, the latter half in above formula can be used to measure the similitude of two behavior sequences X and Y.When X and Y are complete
When exactly the same, D (X, Y)=0；When X and Y do not have any common subsequence, D (X, Y)=1.
After having the distance definition of behavior pattern, so that it may be clustered to user behavior sequence, thus user behavior mould
Formula is distinguished to improve the accuracy of abnormality detection.In order to quickly detect abnormal behaviour, a kind of quick clustering algorithm is needed
Cluster task is completed, in clustering algorithm, k- center point method is simple, quickly, is able to satisfy needs, and in face of presence
It is healthy and strong when the network data of " noise " and isolated point, therefore select k- center point method.
The elementary tactics of k- center point method is: one arbitrarily, which found, for each cluster first represents behavior sequence object,
Other objects then according to them at a distance from these cluster representative objects respectively by they belong to each corresponding cluster centre (according to
Distance calculating method in upper step), and if replacing a cluster representative and can improve obtained clustering result quality, it can
Old cluster representative object is replaced newly to represent object with one.Iteration continues, so that it may will be so behavior sequence is categorized into k not
Same classification.
S3: LSTM neural network model is established
The different LSTM neural network model of k kind is established respectively first against k different classes of user behavior sequences, often
The input of a network is the corresponding behavior sequence data of such network.The operational process of the model are as follows: by corresponding behavior sequence
The preceding n-1 behavior of column is encoded to input layer of the hidden variable as neural network, using Attention mechanism, by hidden change
Divided attention power weight coefficient is measured, hidden variable is generated into the context variable comprising whole behavior sequence traffic flow information；LSTM
The shot and long term memory network number of plies is more, stronger to the study predictive ability of behavior sequence.But the number of plies can make model when excessively high
Training is difficult to restrain, therefore 3 layers of LSTM network is used in the present invention.Meanwhile in last plus one layer of full articulamentum for exporting
As a result dimensionality reduction, as shown in Figure 3.Finally use SOFTMAX function as the output layer of neural network, corresponding label information is
The classification of the last one behavior of behavior sequence.Decline backpropagation by gradient and lose training neural network model, and constantly
The parameter of model is adjusted, the final LSTM neural network model for obtaining training and completing.
S4: network behavior abnormality detection
To the network flow data to be detected being collected into, first progress data prediction, then according to user's access
Specific network entity specifies being specifically defined for user behavior sequence, and pretreated network flow data is arranged as user's row
For sequence.
To the user behavior sequence, according to the method in step 2 by k cluster center behavior sequence object of itself and other into
Row similarity measurement finds the maximum center behavior sequence object of similitude, centered on user behavior sequence mark to be detected
The corresponding classification of behavior sequence object.
It is used as input data after behavior sequence is removed the last one behavior, the corresponding training of the input category is completed
LSTM neural network model.This section of behavior sequence is encoded to hidden variable by model, and by Attention mechanism, by hidden variable
The context variable comprising whole behavior sequence traffic flow information is generated, predicts that the classification of the next behavior of behavior sequence is simultaneously led to
Discrete probability distribution after crossing the output normalization of SOFTMAX function.
The ProbabilityDistribution Vector x for the next behavior that LSTM neural network prediction is gone out1With true next behavior classification
One-Hot vector x2Manhatton distance is calculated as follows, for the size of distance as abnormality detection index, distance is bigger, it is believed that
The exception of the network behavior may be bigger:
Wherein, d12Indicate vector x1With x2Distance, m be ProbabilityDistribution Vector dimension namely network behavior classification
Number.
In the present invention, a kind of network behavior based on LSTM is proposed for the deficiency in traditional network method for detecting abnormality
Method for detecting abnormality.Wound is made that in the key technologies such as the classification of network behavior sequence and LSTM Network anomaly detection in the present invention
Newly.
Although the illustrative specific embodiment of the present invention is described above, in order to the technology of the art
Personnel understand the present invention, it should be apparent that the present invention is not limited to the range of specific embodiment, to the common skill of the art
For art personnel, if various change the attached claims limit and determine the spirit and scope of the present invention in, these
Variation is it will be apparent that all utilize the innovation and creation of present inventive concept in the column of protection.
Claims (2)
1. a kind of network behavior method for detecting abnormality based on LSTM, which comprises the following steps:
(1), network flow data is collected and cleaning arranges；
By being deployed in the distributed agent collection network data on flows of each host terminal, according to analysis demand to the flow of collection
Data are cleaned, and are then directed to current network data, the definition of clear user behavior in a network, and to network data again
In each user crawl be converted to user behavior sequence；
(2), behavior sequence is classified；
It for the action trail sequence of all users, is clustered according to k- central point algorithm, it is a different classes of to be classified as k
Behavior sequence, the user user i.e. to be detected for needing to carry out network behavior abnormality detection are different from k by its behavior sequence
The cluster central point of the behavior sequence of classification carries out similarity measurement, and the one kind for taking its most like is as user behavior sequence to be detected
Classification；
(3), LSTM neural network model is established；
Using k class behavior sequence data obtained in step 2 as the input data of k LSTM neural network, in conjunction with
Attention mechanism is trained LSTM neural network, obtains k LSTM neural network model of training completion, wherein every
A neural network model corresponds to a kind of user behavior classification；
(4), network behavior abnormality detection；
To user to be detected, using its behavior sequence as the input of the LSTM neural network model of corresponding classification, and by model
Intensity of anomaly of the difference as network behavior between behavior prediction and true behavior.
2. network behavior method for detecting abnormality according to claim 1, which is characterized in that in step (2), described passes through
K- central point algorithm to user behavior sequence classify and step (3) in, described establishes LSTM neural network model:
2.1), for all user behavior sequences to be sorted, step 1: the longest for finding out any two user behavior sequence is public
Subsequence LCS altogether；
X=(x is saved with c [i] [j]1,x2,…xm) and Y=(y1,y2,…yn) longest common subsequence, then:
It is possible thereby to acquire the longest common subsequence between two user behavior sequences；
After the behavior sequence for having each user, so that it may indicate two user's rows by calculating the similarity between user
Similarity degree and relationship between, in order to make, the common subsequence of two behavior sequences is longer, similarity is bigger, two behaviors
The distance between sequence is smaller, and the distance definition between behavior sequence is as follows:
Wherein | X | indicate the length of behavior sequence X, lcs (X, Y) is the longest common subsequence of two behavior sequences X and Y, on
Latter half in formula can be used to measure the similitude of two behavior sequences X and Y, when X is identical with Y, D (X, Y)=
0, when X and Y does not have any common subsequence, D (X, Y)=1；
After having the distance definition of behavior pattern, so that it may be clustered by k- central point algorithm user behavior sequence, base
This strategy is: arbitrarily finding one first for each cluster and represents behavior sequence object, other objects are then according to them and these
They are belonged to each corresponding cluster centre respectively by the distance of cluster representative object, and if one cluster representative of replacement can change
If kind obtained clustering result quality, then can newly represent object with one replaces old cluster representative object, iteration continues, so that it may
All behavior sequences are categorized into k different classifications, this k classification corresponds respectively to the different user behavior pattern of k kind；
2.2), the different LSTM neural network model of k kind is established respectively first against k different classes of user behavior sequences,
The input of each network is the corresponding behavior sequence data of such network；
The operational process of the model are as follows: the preceding n-1 behavior of corresponding behavior sequence is encoded to hidden variable as nerve net
The input layer of network, using Attention mechanism, by including by hidden variable generation to hidden variable divided attention power weight coefficient
The context variable of whole behavior sequence traffic flow information, then using 3 layers of LSTM shot and long term memory network to improve to row
For the study predictive ability of sequence, meanwhile, in last plus one layer of full articulamentum for exporting the dimensionality reduction of result, finally use
Output layer of the SOFTMAX function as neural network, corresponding label information are the classification of the last one behavior of behavior sequence；
Decline backpropagation training neural network model by gradient, and constantly the parameter of model is adjusted, it is final to obtain
The LSTM neural network model that training is completed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910566453.3A CN110287439A (en) | 2019-06-27 | 2019-06-27 | A kind of network behavior method for detecting abnormality based on LSTM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910566453.3A CN110287439A (en) | 2019-06-27 | 2019-06-27 | A kind of network behavior method for detecting abnormality based on LSTM |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110287439A true CN110287439A (en) | 2019-09-27 |
Family
ID=68007716
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910566453.3A Pending CN110287439A (en) | 2019-06-27 | 2019-06-27 | A kind of network behavior method for detecting abnormality based on LSTM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110287439A (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111262849A (en) * | 2020-01-13 | 2020-06-09 | 东南大学 | Method for identifying and blocking network abnormal flow behaviors based on flow table information |
| CN111600750A (en) * | 2020-05-11 | 2020-08-28 | 北京庭宇科技有限公司 | Speed limit detection method and system for PCDN network node flow |
| CN111738335A (en) * | 2020-06-23 | 2020-10-02 | 鲁东大学 | Time series data abnormity detection method based on neural network |
| CN111797978A (en) * | 2020-07-08 | 2020-10-20 | 北京天融信网络安全技术有限公司 | Internal threat detection method and device, electronic equipment and storage medium |
| CN111970169A (en) * | 2020-08-14 | 2020-11-20 | 中山大学 | Protocol flow identification method based on GRU network |
| CN111967011A (en) * | 2020-07-10 | 2020-11-20 | 电子科技大学 | Interpretable internal threat assessment method |
| CN112306982A (en) * | 2020-11-16 | 2021-02-02 | 杭州海康威视数字技术股份有限公司 | Abnormal user detection method and device, computing equipment and storage medium |
| CN112565183A (en) * | 2020-10-29 | 2021-03-26 | 中国船舶重工集团公司第七0九研究所 | Network flow abnormity detection method and device based on flow dynamic time warping algorithm |
| CN112631888A (en) * | 2020-12-30 | 2021-04-09 | 航天信息股份有限公司 | Fault prediction method and device of distributed system, storage medium and electronic equipment |
| CN112671551A (en) * | 2020-11-23 | 2021-04-16 | 中国船舶重工集团公司第七0九研究所 | Network traffic prediction method and system based on event correlation |
| CN112818868A (en) * | 2021-02-03 | 2021-05-18 | 招联消费金融有限公司 | Behavior sequence characteristic data-based violation user identification method and device |
| CN113409105A (en) * | 2021-06-04 | 2021-09-17 | 山西大学 | E-commerce network abnormal user detection method and system |
| CN113472742A (en) * | 2021-05-28 | 2021-10-01 | 中国科学院信息工程研究所 | Internal threat detection method and device based on gated cyclic unit |
| CN113569879A (en) * | 2020-04-28 | 2021-10-29 | 中国移动通信集团浙江有限公司 | Training method of abnormal recognition model, abnormal account recognition method and related device |
| CN114154034A (en) * | 2021-11-04 | 2022-03-08 | 杭州电魂网络科技股份有限公司 | Game studio user detection method and system based on longest public subsequence |
| CN114221816A (en) * | 2021-12-17 | 2022-03-22 | 恒安嘉新（北京）科技股份公司 | Flow detection method, device, equipment and storage medium |
| CN114238625A (en) * | 2021-10-11 | 2022-03-25 | 电子科技大学长三角研究院(湖州) | Network water army behavior early warning method based on inconsistency of user dynamic chart characterization |
| CN115086043A (en) * | 2022-06-17 | 2022-09-20 | 电子科技大学 | Encryption network flow classification and identification method based on minimum public subsequence |
| CN115426124A (en) * | 2022-05-13 | 2022-12-02 | 湖北大学 | User abnormal network behavior prediction method and device |
| CN116232761A (en) * | 2023-05-04 | 2023-06-06 | 华东交通大学 | Method and system for detecting abnormal network traffic based on shapelet |
| WO2023243036A1 (en) * | 2022-06-16 | 2023-12-21 | 三菱電機株式会社 | Information processing device, program, and information processing method |
| WO2024009390A1 (en) * | 2022-07-05 | 2024-01-11 | 三菱電機株式会社 | Information processing device, program, and information processing method |
| CN117573480A (en) * | 2023-12-14 | 2024-02-20 | 杭州丽冠科技有限公司 | Data security monitoring method and device based on artificial intelligence |
| CN118413405A (en) * | 2024-07-03 | 2024-07-30 | 本溪钢铁（集团）信息自动化有限责任公司 | Industrial control auditing method and device based on industrial Internet |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106657410A (en) * | 2017-02-28 | 2017-05-10 | 国家电网公司 | Detection method for abnormal behaviors based on user access sequence |
| CN106815639A (en) * | 2016-12-27 | 2017-06-09 | 东软集团股份有限公司 | The abnormal point detecting method and device of flow data |
| CN107070683A (en) * | 2016-12-12 | 2017-08-18 | 国网北京市电力公司 | The method and apparatus of data prediction |
| CN108718291A (en) * | 2018-02-28 | 2018-10-30 | 北京微智信业科技有限公司 | A kind of malice URL detection methods based on big data |
| CN108900546A (en) * | 2018-08-13 | 2018-11-27 | 杭州安恒信息技术股份有限公司 | The method and apparatus of time series Network anomaly detection based on LSTM |
| CN109615019A (en) * | 2018-12-25 | 2019-04-12 | 吉林大学 | Anomaly detection method based on space-time autocoder |
| CN109641603A (en) * | 2017-07-19 | 2019-04-16 | 株式会社东芝 | Abnormal detector, method for detecting abnormality and computer program |
| CN109685376A (en) * | 2018-12-26 | 2019-04-26 | 国家电网公司华中分部 | A kind of power customer abnormal behaviour method for early warning based on similarity analysis theory |
| CN109871976A (en) * | 2018-12-20 | 2019-06-11 | 浙江工业大学 | A kind of prediction technique of power quality containing distributed power distribution network based on cluster and neural network |
| CN109886833A (en) * | 2019-01-21 | 2019-06-14 | 广东电网有限责任公司信息中心 | A kind of deep learning method of smart grid-oriented server traffic abnormality detection |
-
2019
- 2019-06-27 CN CN201910566453.3A patent/CN110287439A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107070683A (en) * | 2016-12-12 | 2017-08-18 | 国网北京市电力公司 | The method and apparatus of data prediction |
| CN106815639A (en) * | 2016-12-27 | 2017-06-09 | 东软集团股份有限公司 | The abnormal point detecting method and device of flow data |
| CN106657410A (en) * | 2017-02-28 | 2017-05-10 | 国家电网公司 | Detection method for abnormal behaviors based on user access sequence |
| CN109641603A (en) * | 2017-07-19 | 2019-04-16 | 株式会社东芝 | Abnormal detector, method for detecting abnormality and computer program |
| CN108718291A (en) * | 2018-02-28 | 2018-10-30 | 北京微智信业科技有限公司 | A kind of malice URL detection methods based on big data |
| CN108900546A (en) * | 2018-08-13 | 2018-11-27 | 杭州安恒信息技术股份有限公司 | The method and apparatus of time series Network anomaly detection based on LSTM |
| CN109871976A (en) * | 2018-12-20 | 2019-06-11 | 浙江工业大学 | A kind of prediction technique of power quality containing distributed power distribution network based on cluster and neural network |
| CN109615019A (en) * | 2018-12-25 | 2019-04-12 | 吉林大学 | Anomaly detection method based on space-time autocoder |
| CN109685376A (en) * | 2018-12-26 | 2019-04-26 | 国家电网公司华中分部 | A kind of power customer abnormal behaviour method for early warning based on similarity analysis theory |
| CN109886833A (en) * | 2019-01-21 | 2019-06-14 | 广东电网有限责任公司信息中心 | A kind of deep learning method of smart grid-oriented server traffic abnormality detection |
Cited By (37)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111262849A (en) * | 2020-01-13 | 2020-06-09 | 东南大学 | Method for identifying and blocking network abnormal flow behaviors based on flow table information |
| CN113569879B (en) * | 2020-04-28 | 2024-03-19 | 中国移动通信集团浙江有限公司 | Training method of abnormal recognition model, abnormal account recognition method and related device |
| CN113569879A (en) * | 2020-04-28 | 2021-10-29 | 中国移动通信集团浙江有限公司 | Training method of abnormal recognition model, abnormal account recognition method and related device |
| CN111600750A (en) * | 2020-05-11 | 2020-08-28 | 北京庭宇科技有限公司 | Speed limit detection method and system for PCDN network node flow |
| CN111600750B (en) * | 2020-05-11 | 2022-10-21 | 北京庭宇科技有限公司 | Speed limit detection method and system for PCDN network node flow |
| CN111738335A (en) * | 2020-06-23 | 2020-10-02 | 鲁东大学 | Time series data abnormity detection method based on neural network |
| CN111797978A (en) * | 2020-07-08 | 2020-10-20 | 北京天融信网络安全技术有限公司 | Internal threat detection method and device, electronic equipment and storage medium |
| CN111967011A (en) * | 2020-07-10 | 2020-11-20 | 电子科技大学 | Interpretable internal threat assessment method |
| CN111970169A (en) * | 2020-08-14 | 2020-11-20 | 中山大学 | Protocol flow identification method based on GRU network |
| CN112565183A (en) * | 2020-10-29 | 2021-03-26 | 中国船舶重工集团公司第七0九研究所 | Network flow abnormity detection method and device based on flow dynamic time warping algorithm |
| CN112306982B (en) * | 2020-11-16 | 2021-07-16 | 杭州海康威视数字技术股份有限公司 | Abnormal user detection method and device, computing equipment and storage medium |
| CN112306982A (en) * | 2020-11-16 | 2021-02-02 | 杭州海康威视数字技术股份有限公司 | Abnormal user detection method and device, computing equipment and storage medium |
| CN112671551A (en) * | 2020-11-23 | 2021-04-16 | 中国船舶重工集团公司第七0九研究所 | Network traffic prediction method and system based on event correlation |
| CN112671551B (en) * | 2020-11-23 | 2022-11-18 | 中国船舶重工集团公司第七0九研究所 | Network traffic prediction method and system based on event correlation |
| CN112631888A (en) * | 2020-12-30 | 2021-04-09 | 航天信息股份有限公司 | Fault prediction method and device of distributed system, storage medium and electronic equipment |
| CN112818868A (en) * | 2021-02-03 | 2021-05-18 | 招联消费金融有限公司 | Behavior sequence characteristic data-based violation user identification method and device |
| CN112818868B (en) * | 2021-02-03 | 2024-05-28 | 招联消费金融股份有限公司 | Method and device for identifying illegal user based on behavior sequence characteristic data |
| CN113472742A (en) * | 2021-05-28 | 2021-10-01 | 中国科学院信息工程研究所 | Internal threat detection method and device based on gated cyclic unit |
| CN113472742B (en) * | 2021-05-28 | 2022-09-27 | 中国科学院信息工程研究所 | Internal threat detection method and device based on gated cyclic unit |
| CN113409105A (en) * | 2021-06-04 | 2021-09-17 | 山西大学 | E-commerce network abnormal user detection method and system |
| CN113409105B (en) * | 2021-06-04 | 2023-09-26 | 山西大学 | Method and system for detecting abnormal users of e-commerce network |
| CN114238625A (en) * | 2021-10-11 | 2022-03-25 | 电子科技大学长三角研究院(湖州) | Network water army behavior early warning method based on inconsistency of user dynamic chart characterization |
| CN114154034A (en) * | 2021-11-04 | 2022-03-08 | 杭州电魂网络科技股份有限公司 | Game studio user detection method and system based on longest public subsequence |
| CN114221816A (en) * | 2021-12-17 | 2022-03-22 | 恒安嘉新（北京）科技股份公司 | Flow detection method, device, equipment and storage medium |
| CN114221816B (en) * | 2021-12-17 | 2024-05-03 | 恒安嘉新（北京）科技股份公司 | Flow detection method, device, equipment and storage medium |
| CN115426124B (en) * | 2022-05-13 | 2024-10-25 | 湖北大学 | Method and device for predicting abnormal network behavior of user |
| CN115426124A (en) * | 2022-05-13 | 2022-12-02 | 湖北大学 | User abnormal network behavior prediction method and device |
| WO2023243036A1 (en) * | 2022-06-16 | 2023-12-21 | 三菱電機株式会社 | Information processing device, program, and information processing method |
| CN115086043B (en) * | 2022-06-17 | 2023-03-21 | 电子科技大学 | Encryption network flow classification and identification method based on minimum public subsequence |
| CN115086043A (en) * | 2022-06-17 | 2022-09-20 | 电子科技大学 | Encryption network flow classification and identification method based on minimum public subsequence |
| WO2024009390A1 (en) * | 2022-07-05 | 2024-01-11 | 三菱電機株式会社 | Information processing device, program, and information processing method |
| JP7558459B2 (en) | 2022-07-05 | 2024-09-30 | 三菱電機株式会社 | Information processing device, program, and information processing method |
| CN116232761B (en) * | 2023-05-04 | 2023-07-14 | 华东交通大学 | Method and system for detecting abnormal network traffic based on shapelet |
| CN116232761A (en) * | 2023-05-04 | 2023-06-06 | 华东交通大学 | Method and system for detecting abnormal network traffic based on shapelet |
| CN117573480A (en) * | 2023-12-14 | 2024-02-20 | 杭州丽冠科技有限公司 | Data security monitoring method and device based on artificial intelligence |
| CN118413405A (en) * | 2024-07-03 | 2024-07-30 | 本溪钢铁（集团）信息自动化有限责任公司 | Industrial control auditing method and device based on industrial Internet |
| CN118413405B (en) * | 2024-07-03 | 2024-09-17 | 本溪钢铁（集团）信息自动化有限责任公司 | Industrial control auditing method and device based on industrial Internet |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110287439A (en) | A kind of network behavior method for detecting abnormality based on LSTM | |
| CN108023876A (en) | Intrusion detection method and intruding detection system based on sustainability integrated study | |
| CN112491796A (en) | Intrusion detection and semantic decision tree quantitative interpretation method based on convolutional neural network | |
| CN106250442A (en) | The feature selection approach of a kind of network security data and system | |
| Cui et al. | Learning global pairwise interactions with Bayesian neural networks | |
| KR102362872B1 (en) | Method for refining clean labeled data for artificial intelligence training | |
| CN112700324A (en) | User loan default prediction method based on combination of Catboost and restricted Boltzmann machine | |
| Zhu et al. | Traffic monitoring and anomaly detection based on simulation of luxembourg road network | |
| WO2019200739A1 (en) | Data fraud identification method, apparatus, computer device, and storage medium | |
| CN109462578A (en) | Threat intelligence use and propagation method based on statistical learning | |
| Yeh et al. | Merchant category identification using credit card transactions | |
| Naranjo-Alcazar et al. | An open-set recognition and few-shot learning dataset for audio event classification in domestic environments | |
| Jha et al. | Criminal behaviour analysis and segmentation using k-means clustering | |
| CN114897085A (en) | Clustering method based on closed subgraph link prediction and computer equipment | |
| CN113837266A (en) | Software defect prediction method based on feature extraction and Stacking ensemble learning | |
| Wang et al. | Early diagnosis of Parkinson's disease with Speech Pronunciation features based on XGBoost model | |
| CN104636636B (en) | The long-range homology detection method of protein and device | |
| CN105930430B (en) | Real-time fraud detection method and device based on non-accumulative attribute | |
| CN113159976B (en) | Identification method for important users of microblog network | |
| Dash et al. | Statistical analysis and comparison of deep convolutional neural network models for the identification and classification of maize leaf diseases | |
| Shahane et al. | A Survey on Classification Techniques to Determine Fake vs. Real Identities on Social Media Platforms | |
| CN115329838A (en) | Attribute graph anomaly detection method considering class imbalance | |
| Munikoti et al. | Bayesian graph neural network for fast identification of critical nodes in uncertain complex networks | |
| CN112422505A (en) | Network malicious traffic identification method based on high-dimensional extended key feature vector | |
| CN114519605A (en) | Advertisement click fraud detection method, system, server and storage medium |
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: 20190927 |