CN109359686A - A kind of user's portrait method and system based on Campus Network Traffic - Google Patents
A kind of user's portrait method and system based on Campus Network Traffic Download PDFInfo
- Publication number
- CN109359686A CN109359686A CN201811217041.0A CN201811217041A CN109359686A CN 109359686 A CN109359686 A CN 109359686A CN 201811217041 A CN201811217041 A CN 201811217041A CN 109359686 A CN109359686 A CN 109359686A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- network
- portrait
- campus
- 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
- 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
- G06F18/21—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
- G06F18/2415—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
Abstract
User's portrait method and system based on Campus Network Traffic that the invention proposes a kind of, acquire the data on flows in campus network, data are cleaned and are pre-processed, classifier is established using machine learning algorithm, training and Optimum Classification device, analyze the integrality of network in data set, user's portrait is carried out to the user in campus network using trained classifier, user's portrait result visualization is presented to administrator, improve the stability of network, administrator is facilitated to carry out the monitoring and maintenance of network, timely investigation and response can be carried out to the Cyberthreat or abnormal user of burst.
Description
Technical field
It is the invention belongs to data analysis and the crossing domain of Internet technology, in particular to a kind of based on Campus Network Traffic
User's portrait method and system.
Background technique
With the development of internet, the data generated in network are more and more, and Feature Engineering and machine learning algorithm go out
It is existing, so that analysis data become increasingly to facilitate, in addition, also more and more valuable by analyzing the result that data obtain.
User's portrait is actually a kind of application of Feature Engineering, and the data that the purpose is to be generated by analysis user are dug
The value information that pick is wherein hidden, analysis and the characteristic behavior for summarizing user.The key point for constructing user's portrait is to utilize storage
It arranged, excavated and is analyzed in the user data of the magnanimity of lane database, then by training pattern, user behavior is carried out
Classification, labels " " to user, to reach the method for user's portrait.
Currently used user's portrait method is according to user demographic characteristics, network access behavior, Social behaviors and to disappear
The information such as expense behavior and the user model of a labeling taken out.
But traditional method, there are still some shortcomings, firstly, being authenticity guarantee's problem of data, in many methods
Data source will lead to data source confusion in multiple channel in this way, the problems such as data format disunity, information source is inaccurate,
To influence the result of analysis.Secondly, there are isolated problems for the data source of Part Methods, due between operator or application
Closure, many methods are difficult to obtain all data of user, so that analysis can not be in user's whole network scope of activities
It carries out, so that user's portrait can not be carried out accurately.In addition the base also extremely important, traditional to the selected method of data classification
All cannot accurately it be classified to data well in the method for statistics and single machine learning classification method.
Summary of the invention
The method and system in view of the above problems, user that the present invention provides a kind of based on Campus Network Traffic draws a portrait, solve
Above-mentioned problems faced.The present invention be on the basis of campus network service, by acquisition, the network row of user in Integration of Campus Network
For data, a variety of machine learning methods are then used, establish and train user's portrait model for Campus Network Traffic.And by picture
As result visualization is presented to network administrator, facilitates network administrator quickly to check network state, find Cyberthreat in time
And it makes a response.
To achieve the above object, The technical solution adopted by the invention is as follows:
First by traffic capture tool, acquire flow on the server of campus network, the access data including user and
Behavioral data, such as access time, access frequency, access address, access link etc.；Then by collected data carry out cleaning and
Pretreatment, selection can most reflect the feature of network state and user behavior；It is selected in the data pre-processed some with generation
The data of table carry out handmarking, generate training sample, training sample is put into Bayes classifier and SVM classifier,
Machine learning model is established, and is trained and optimizes.Then other data are trained and are divided using trained model
Class generates user's portrait according to result, the result visualization that user draws a portrait is presented to administrator, and indicate whether in a network
There are security risks.
The method specifically, a kind of user based on Campus Network Traffic of the present invention draws a portrait, comprising the following steps:
1) data acquisition phase: by the data on flows in acquisition Campus Network Server, collected data are carried out clear
It washes and pre-processes；
2) modelling phase: analyzing and handles data, the feature of selection energy accurate description network state and user behavior, in conjunction with
Correlation machine learning algorithm generates model, and model is trained and is optimized；
3) data are drawn a portrait the stage: overall operation state and user data are analyzed respectively using trained model,
Network state and user behavior characteristics are described from multiple directions；
4) application and analysis phase: analyzing the result that data are drawn a portrait, and analysis result visualization is presented
To network administrator, facilitate administrator that can preferably manage network, monitor network state, while can be to there is abnormal behaviour
User carries out analysis and checks and make a response.
As a further improvement of the present invention, in step 1), the data on flows includes rising for the user in campus network
Begin access time point, access duration time, access target, the size for accessing data volume and access content.
As a further improvement of the present invention, in step 2), when handling data, representative data are chosen first
Handmarking is carried out, training sample is generated, later using the data training classifier marked, while using multiple classifiers pair
Different types of data set is classified respectively, until classifier generates preferable classifying quality, then by trained model
It is saved in disk.
As a further improvement of the present invention, in step 3), many-sided description network state includes when describing this
Network flow total amount, network access total degree and the network generated in section accesses peak value；The user behavior characteristics include using
Network address, network access time, network access frequency, network access link, network access content and the network access quantity at family.
As a further improvement of the present invention, in step 3), what is be substantially carried out is for each use in campus network data
Family carries out user's portrait, the data on flows of user and website visiting historical data is put into respectively trained classifier,
Analysis obtains the behavioural characteristic of each user, and stamps corresponding label and classification is marked, and each user can be according to classification
Result stamp multiple labels.
As a further improvement of the present invention, the result that the data are drawn a portrait includes user behavior label, user
Network behavior feature and the potential security threat of user.
A kind of user's portrait system based on Campus Network Traffic, comprising:
Data acquisition module is acquired data, cleans and pre- for acquiring the data on flows in Campus Network Server
Processing；
Modeling module finds the feature that can most describe network and user behavior, in conjunction with phase for analyzing simultaneously preprocessed data
The learning algorithm that shuts down generates model, and model is trained and is optimized；
Data portrait module, for being analyzed respectively using trained model entirety and user data, from multiple sides
To description network state and user behavior characteristics；
Using and analysis module, for analyzing the result that data are drawn a portrait, and by analysis result visualization be in
Network administrator is now given, facilitates administrator that can preferably manage network, monitors network, while can analyze abnormal user
It checks and makes a response.
As a further improvement of the present invention, the data acquisition module includes:
Campus network data processing unit: different types of in campus network for that will be acquired during capturing data on flows
Data；
Data storage unit:, will be collected a variety of for data to be stored and accessed using Hadoop Distributed Architecture
Data distribution formula is stored in Hadoop cluster, is checked and is called at any time using Hbase, in combination with other in Hadoop ecology
Software coordinates cooperation manage data jointly.
As a further improvement of the present invention, the data acquisition module includes:
Preprocessed data unit converts partial data format for removing to useless field is analyzed；
Training pattern unit generates classifier, and preliminary artificial mark is carried out to flow for using machine learning algorithm
Note, then using the data training marked and Optimized model.
As a further improvement of the present invention, data portrait module includes:
Overall data analytical unit, for carrying out whole analysis to the network data under current state；
User's portrait and labeling unit are generated for carrying out labeling to user using trained model and data
User's portrait.
Compared with prior art, the present invention has the advantages that
Method of the invention mainly accesses the base that the flow that internet generates is drawn a portrait as analysis using the user in campus network
The portrait of network state and campus network users is depicted to come by data processing, machine learning algorithm, data visualization for plinth.
Its data source is in Campus Network Server, so that data are true and reliable, and contains the various nets of user in campus network comprehensively
The data of network behavior are drawn a portrait so as to obtain complete user from many aspects by analysis；In addition, passing through visualization side
Method, so that the user of analysis draws a portrait, result is more comprehensively intuitive, and administrator is facilitated to check and make decisions at any time.
System of the invention is realized by data acquisition module, modeling module, data portrait module, application and analysis module
Data Layer, process layer are connected with the hardware of presentation layer, eventually by the data on flows in acquisition campus network, are carried out to data clear
It washes and pre-processes, establish classifier, training and Optimum Classification device using machine learning algorithm, analyze the entirety of network in data set
State carries out user's portrait to the user in campus network using trained classifier, and user's portrait result visualization is presented
To administrator, the stability of network is improved, administrator is facilitated to carry out the monitoring and maintenance of network, it can be to the Cyberthreat of burst
Or abnormal user carries out timely investigation and response.
Detailed description of the invention
Fig. 1 is the system framework figure of building user portrait method in the present invention；
Fig. 2 is the flow chart of foundation and training user's portrait model in the present invention；
Fig. 3 is the flow chart of Bayes classifier training and classification used in user's portrait in the present invention；
Fig. 4 is user's Figure Characteristics and label schematic diagram in the present invention.
Specific embodiment
Below by drawings and examples, technical scheme of the present invention will be described in further detail.
The present invention provides a kind of user's portrait method, system based on Campus Network Traffic, passes through building data on flows analysis system
System collects the flow of user in campus network, a model is generated and trained in conjunction with a variety of machine learning algorithms, and use the model
The method for constructing multi-analysis user portrait, analysis obtain the network behavior feature of user in campus network, network state,
With the presence or absence of potential Cyberthreat, the data validity of existing user's portrait method and the disadvantage of isolated problem are overcome
End provides the decision basis based on analysis user's portrait, system framework reference for the status monitoring of campus network, user management
Fig. 1.
Data Layer is acquired for data: by the data on flows in acquisition Campus Network Server, data is acquired,
Cleaning and pretreatment；Data on flows includes the starting access time point of the user in campus network, access duration time, access data
The size and access content of amount.Comprising:
1) campus network acquires data cell: different types of in campus network for acquiring during capturing data on flows
Flow, downloading flow, game flow etc. is broadcast live in data, including network flow data, such as video flow；Website visitation data, such as
Portal website, entertainment sites, shopping website, Educational website, social network sites etc..
2) data storage unit: storing using Hadoop Distributed Architecture and access data, by collected a variety of numbers
It according to distributed storage in Hadoop cluster, checks and calls at any time using Hbase, in combination with other in Hadoop ecology
Software coordinates cooperation manages data jointly.
Process layer, for modeling: analyzing and preprocessed data, searching can most describe the feature of network and user behavior, knot
It closes correlation machine learning algorithm and generates model, and model is trained and is optimized；It analyzes and preprocessed data includes deleting nothing
With field, merges similar features, add new data characteristics.
It is also used to data portrait: network entirety and user data being analyzed respectively using trained model, from multiple
Direction describes network state and user behavior characteristics respectively；Various aspects description network state includes describing network flow in the period
Total amount, network access total degree and network access peak value.When user behavior characteristics include the network address of user, network access
Between, network access frequency, network access link, network access content and network access quantity.
Comprising:
1) preprocessed data unit: removing to useless field is analyzed, carry out conversion unification to partial data format, convenient
Next analysis；
2) training pattern unit is manually marked using a variety of machine learning algorithms, and to part exemplary traffic data,
Then using the data training classifier marked, model is continuouslyd optimize；
3) the whole analysis based on statistics overall data analytical unit: is carried out to the network data under current state.
4) user's portrait and labeling unit, classify to data using trained model, while utilizing classification knot
Fruit carries out labeling to user, generates user's portrait；
Presentation layer, for applying and analyzing: analyzing the result that data are drawn a portrait, and result visualization will be analyzed
It is presented to network administrator, facilitates administrator that can preferably manage network, monitors network, while can divide abnormal user
Analysis is checked and is made a response.
Using and the analysis phase include that result visualization is presented to network administrator, administrator is learnt at once
Current network state, network congestion degree, if there are potential Cyberthreats, and can check according to demand designated user or
The portrait result of user group.
User's portrait result includes user behavior label, user network behaviors feature and the potential security threat of user.
The main purpose of present system is as follows:
First, the utility value of campus network data is given full play to, using campus network reliable data source advantage comprehensively, is collected
And Campus Network Traffic data are stored, next to carry out user's portrait and analysis using being collected into data.
Second, in campus network user individual and network integrally carry out portrait analysis, the network supervision for after provides
Data basis, the analysis and decision after facilitating.
User's portrait method based on Campus Network Traffic that the present invention also provides a kind of, the flow chart of this method refer to Fig. 2, under
Mask body illustrates this method, comprising the following steps:
Step S1: acquisition data on flows: the flow of Campus Network Server is collected；
In step sl, when accessing internet in campus network as user, the content and history requested access to can all pass through
The server of campus network forwards, therefore permission can be used directly in server in the network administrator for possessing server admin permission
The traffic captures tools such as middle deployment tcpdump capture the data generated in network, while ensure that the true of the data captured
Property, integrality and real-time.
Meanwhile different types of data in campus network, including network flow will be acquired during capturing data on flows
Flow, downloading flow, game flow etc. is broadcast live in data, such as video flow；Website visitation data, such as portal website entertain net
It stands, shopping website, Educational website, social network sites etc..Since the data volume of generation is more huge, the storage of common single host is empty
Between be unable to satisfy storage demand, therefore data will be stored and accessed using Hadoop Distributed Architecture, will be collected a variety of
Data distribution formula is stored in Hadoop cluster, is checked and is called at any time using Hbase, in combination with other in Hadoop ecology
Software coordinates cooperation manage data jointly.
Step S2: preprocessed data: cleaning is removed to useless field is analyzed, carries out arrangement system to partial data format
One, facilitate next analysis；It analyzes and preprocessed data includes deleting hashed field, merge similar features, add new number
According to feature.
Then, the data of storage are being cleared up and are being pre-processed, removing the field or invalid data useless to analysis,
Such as TTL, web site requests method, User-Agent etc..Retain to useful field is analyzed, such as user's source IP address, source port, mesh
IP address, destination port, the number of the packet of transmission, the number of the packet received, the total flow of inflow, the total flow of outflow, visit
The web site url asked, the attributes such as access time.And by the data uniform output format after cleaning, then adds some pairs of analyses and have
The field of help can be submitted the website domain name in data in the form of web form such as website visitation data
Web crawlers is write to the website Alexa ranking, and using Python+Selenium tool, acquires the temperature of the website of return
And websites collection, the analysis statistics after facilitating, and the data of the uniform format after arrangement are newly stored into Hbase distribution number
According in library.
Step S3: training pattern generates classifier using machine learning algorithm, and preliminary artificial mark is carried out to flow
Note, then using the data training marked and Optimized model；
Since the essence of user's portrait is to characterize, that is, excavate the behavioural characteristic of user, therefore the process that user can be drawn a portrait
Regard the process of classification as, classifier is constructed using a variety of machine learning algorithms in the method, according to different network flows
Feature is classified using different classifiers.Due to needing to be to solve labeling problem in this method, and unsupervised learning
In only analysis data itself, be the aggregation classification carried out according to the physical feature of data, obtained result possibly can not correspond to
It is required that purpose feature, therefore in the method select supervised learning algorithm, in the method select supervised learning in
Bayesian Classification Arithmetic and support vector machines (SVM).
Wherein the classification process of Bayes classifier is as shown in figure 3, whole flow process can be divided into three phases: preparing data
Collection stage, training classifier stage and application class stage.It is the preparation stage first, is selected in the data pre-processed and true
Then fixed most representational attributive character selects the apparent data set of some features to carry out handmarking, generates training set；
The frequency that each characteristic of division occurs in the sample is calculated, then to every using the training set marked followed by the training stage
The conditional probability of one characteristic attribute computation partition, and record result；It is finally the application stage, data on flows to be sorted is put
Into the trained classifier of previous step, classify, generates the label of user's classification.In addition, SVM classifier is generally used for solving
Certainly two classification problem, and user's portrait in method presented herein is classification problem more than one, it is therefore desirable to use SVM pairs
The label and other labels of each classification carry out one time two classification, and the process of svm classifier is similar with Bayes classifier.
When handling data, some representative data are chosen first and carry out handmarking, generate training sample, it
Model is continuouslyd optimize using the data training marked afterwards, until classifier has preferable classifying quality.This stage is due to relating to
And the classification of the network data arrived is different, the standard and result of classification are also different, be such as related to net flow assorted feature and
Network access behavior classification is characterized in different, therefore multiple classifiers may be needed to different during classification
Data set is classified, and then trained model is saved in disk, facilitates next use.
Step S4: whole analysis overall data analysis: is carried out to the network data under current state；
In step s 4, what is carried out first is whole network data analysis, is mainly concerned with statistics network overall condition
Constant, the beginning and ending time including the data set, total packet number of transmission, uplink total flow and peak value, downlink total flow and peak value,
The information such as total number of users in network have a primary statistics and assessment to whole network state, if it find that having abnormal
Data are regarded as network and abnormal conditions occur, need to take phase if certain period peak flow is higher by normal level for a long time
Answer measure.
Step S5: user's portrait and labeling carry out labeling to user using trained model and data, generate and use
Family portrait；
In step s 5, what is be substantially carried out is to carry out user's portrait for each user in campus network data, by user
Data on flows and website visiting historical data be put into respectively trained classifier, analysis obtains the behavior of each user
Feature, and stamp corresponding label and classification is marked, as shown in figure 4, according to different classification results, in the online period, always
One or more labels are stamped in the directions such as used time, flow dosage, preference access, such as label of a certain user are as follows: online period: evening
On, morning.Daily online used time 1-5 hours, dosage 5-10GB, preference access: video, information.
Step S6: interpretation of result and application.
Finally, the analysis of overall data is generated phase as a result, summarize and carry out visualization presentation with the portrait of each user
The chart answered, is presented to administrator, and administrator is allow to view whole network state, makes administrator to each user or use
Family group has and is understood more intuitively, while can be according to the better monitoring network state of these data, when occurring network in network
When threat, administrator can quickly be positioned by abnormal user in portrait, to respond and decision, prevent from threatening
Further expansion.
To sum up locating, this method collects data on flows and user's access history data in campus network first, generates and instructs
Practice sorter model, user's portrait is carried out to user using classifier, and provide visual present as a result, enabling administrator
Enough more easily monitoring network states are analyzed using data traffic accurate in campus network by this method, improve school
The stability of garden Webweb network has a degree of understanding to every user in network, and being capable of positioning net as early as possible
Threat in network.
What said above is merely illustrative of the invention's technical idea, and this does not limit the scope of protection of the present invention,
It is all without departing from technical idea proposed by the present invention, any changes made on the basis of the technical scheme each falls within power of the present invention
Within the protection scope of sharp claim.
Claims (10)
- A kind of method 1. user based on Campus Network Traffic draws a portrait, which comprises the following steps:1) data acquisition phase: by acquisition Campus Network Server in data on flows, to collected data carry out cleaning and Pretreatment；2) modelling phase: analyzing and handles data, the feature of selection energy accurate description network state and user behavior, in conjunction with correlation Machine learning algorithm generates model, and model is trained and is optimized；3) data are drawn a portrait the stage: are analyzed respectively using trained model overall operation state and user data, from more A direction describes network state and user behavior characteristics；4) application and analysis phase: analyzing the result that data are drawn a portrait, and analysis result visualization is presented to net Network administrator facilitates administrator that can preferably manage network, monitors network state, while can be to the user for having abnormal behaviour Analysis is carried out to check and make a response.
- The method 2. user according to claim 1 based on Campus Network Traffic draws a portrait, which is characterized in that in step 1), institute The data on flows stated includes the starting access time point of the user in campus network, access duration time, access target, access data The size and access content of amount.
- The method 3. user according to claim 1 based on Campus Network Traffic draws a portrait, which is characterized in that in step 2), When handling data, representative data are chosen first and carry out handmarking, generate training sample, use has marked later Data train classifier, while being classified respectively using multiple classifiers to different types of data set, until classifier produces Trained classifier, is then saved in disk by raw preferable classifying quality.
- The method 4. a kind of user based on Campus Network Traffic according to claim 1 draws a portrait, which is characterized in that step 3) In, many-sided description network state includes the network flow total amount for describing to generate in the period, network access total degree And network accesses peak value；The user behavior characteristics include the network address of user, network access time, network access frequency Rate, network access link, network access content and network access quantity.
- The method 5. a kind of user based on Campus Network Traffic according to claim 1 draws a portrait, which is characterized in that step 3) In, what is be substantially carried out is to carry out user's portrait for each user in campus network data, by the data on flows of user and website Access history data is put into respectively trained classifier, and analysis obtains the behavioural characteristic of each user, and stamps correspondence Label classification is marked, each user can stamp multiple labels according to the result of classification.
- The method 6. a kind of user based on Campus Network Traffic according to claim 1 draws a portrait, which is characterized in that the number It include user behavior label, user network behaviors feature and the potential security threat of user according to the result that portrait obtains.
- The system 7. a kind of user based on Campus Network Traffic draws a portrait characterized by comprisingData acquisition module is acquired data, cleaning and pre- place for acquiring the data on flows in Campus Network Server Reason；Modeling module, for analyzing and preprocessed data, the feature of network and user behavior can most be described by finding, in conjunction with mutually shutting down Device learning algorithm generates model, and model is trained and is optimized；Data portrait module is retouched for being analyzed respectively using trained model entirety and user data from multiple directions State network state and user behavior characteristics；Using and analysis module, for analyzing the result that data are drawn a portrait, and by analysis result visualization be presented to Network administrator facilitates administrator that can preferably manage network, monitors network, while can carry out analysis to abnormal user and check And it makes a response.
- The system 8. user according to claim 7 based on Campus Network Traffic draws a portrait, which is characterized in that the data are adopted Collecting module includes:Campus network data processing unit: for different types of number in campus network will to be acquired during capturing data on flows According to；Data storage unit: for data to be stored and accessed using Hadoop Distributed Architecture, by collected a variety of data Distributed storage is checked and is called at any time in Hadoop cluster, using Hbase, in combination with other soft in Hadoop ecology The common management data of part coordination.
- The system 9. user according to claim 7 based on Campus Network Traffic draws a portrait, which is characterized in that the data are adopted Collecting module includes:Preprocessed data unit converts partial data format for removing to useless field is analyzed；Training pattern unit for using machine learning algorithm, generation classifier, and carries out preliminary artificial mark to flow, Then using the data training marked and Optimized model.
- The system 10. user according to claim 7 based on Campus Network Traffic draws a portrait, which is characterized in that the data Portrait module include:Overall data analytical unit, for carrying out whole analysis to the network data under current state；User's portrait and labeling unit generate user for carrying out labeling to user using trained model and data Portrait.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811217041.0A CN109359686A (en) | 2018-10-18 | 2018-10-18 | A kind of user's portrait method and system based on Campus Network Traffic |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811217041.0A CN109359686A (en) | 2018-10-18 | 2018-10-18 | A kind of user's portrait method and system based on Campus Network Traffic |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109359686A true CN109359686A (en) | 2019-02-19 |
Family
ID=65345787
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811217041.0A Pending CN109359686A (en) | 2018-10-18 | 2018-10-18 | A kind of user's portrait method and system based on Campus Network Traffic |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109359686A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213112A (en) * | 2019-06-14 | 2019-09-06 | 广州志浩信念网络科技有限公司 | A kind of user behavior analysis method and system |
| CN110245296A (en) * | 2019-06-14 | 2019-09-17 | 浙江华坤道威数据科技有限公司 | A kind of PAS user's portrait analysis system and its method based on big data |
| CN111614614A (en) * | 2020-04-14 | 2020-09-01 | 瑞数信息技术（上海）有限公司 | Safety monitoring method and device applied to Internet of things |
| CN113965389A (en) * | 2021-10-26 | 2022-01-21 | 天元大数据信用管理有限公司 | Network security management method, equipment and medium based on firewall log |
| CN115632839A (en) * | 2022-10-10 | 2023-01-20 | 江苏海洋大学 | Smart campus environment network supervision method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104239351A (en) * | 2013-06-20 | 2014-12-24 | 阿里巴巴集团控股有限公司 | User behavior machine learning model training method and device |
| CN106911806A (en) * | 2017-04-26 | 2017-06-30 | 努比亚技术有限公司 | A kind of method of PUSH message, terminal, server and system |
| US20170206204A1 (en) * | 2016-01-20 | 2017-07-20 | Alibaba Group Holding Limited | System, method, and device for generating a geographic area heat map |
| CN107404398A (en) * | 2017-05-31 | 2017-11-28 | 中山大学 | A kind of networks congestion control judgement system |
| CN107612938A (en) * | 2017-10-27 | 2018-01-19 | 朱秋华 | A kind of network user's anomaly detection method, device, equipment and storage medium |
| CN108596738A (en) * | 2018-05-08 | 2018-09-28 | 新华三信息安全技术有限公司 | A kind of user behavior detection method and device |
-
2018
- 2018-10-18 CN CN201811217041.0A patent/CN109359686A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104239351A (en) * | 2013-06-20 | 2014-12-24 | 阿里巴巴集团控股有限公司 | User behavior machine learning model training method and device |
| US20170206204A1 (en) * | 2016-01-20 | 2017-07-20 | Alibaba Group Holding Limited | System, method, and device for generating a geographic area heat map |
| CN106911806A (en) * | 2017-04-26 | 2017-06-30 | 努比亚技术有限公司 | A kind of method of PUSH message, terminal, server and system |
| CN107404398A (en) * | 2017-05-31 | 2017-11-28 | 中山大学 | A kind of networks congestion control judgement system |
| CN107612938A (en) * | 2017-10-27 | 2018-01-19 | 朱秋华 | A kind of network user's anomaly detection method, device, equipment and storage medium |
| CN108596738A (en) * | 2018-05-08 | 2018-09-28 | 新华三信息安全技术有限公司 | A kind of user behavior detection method and device |
Non-Patent Citations (2)
| Title |
|---|
| 徐龙章等: "《Python 网络数据爬取及分析从入门到精通 爬取篇》", 30 June 2018, 北京航空航天大学出版社 * |
| 马飒飒等: "《物联网基础技术及应用》", 31 January 2018, 西安电子科技大学出版社 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213112A (en) * | 2019-06-14 | 2019-09-06 | 广州志浩信念网络科技有限公司 | A kind of user behavior analysis method and system |
| CN110245296A (en) * | 2019-06-14 | 2019-09-17 | 浙江华坤道威数据科技有限公司 | A kind of PAS user's portrait analysis system and its method based on big data |
| CN111614614A (en) * | 2020-04-14 | 2020-09-01 | 瑞数信息技术（上海）有限公司 | Safety monitoring method and device applied to Internet of things |
| CN111614614B (en) * | 2020-04-14 | 2022-08-05 | 瑞数信息技术（上海）有限公司 | Safety monitoring method and device applied to Internet of things |
| CN113965389A (en) * | 2021-10-26 | 2022-01-21 | 天元大数据信用管理有限公司 | Network security management method, equipment and medium based on firewall log |
| CN115632839A (en) * | 2022-10-10 | 2023-01-20 | 江苏海洋大学 | Smart campus environment network supervision method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109359686A (en) | A kind of user's portrait method and system based on Campus Network Traffic | |
| US11538064B2 (en) | System and method of providing a platform for managing data content campaign on social networks | |
| Liu et al. | Monitoring and analyzing big traffic data of a large-scale cellular network with Hadoop | |
| Chae et al. | Spatiotemporal social media analytics for abnormal event detection and examination using seasonal-trend decomposition | |
| CN102035698B (en) | HTTP tunnel detection method based on decision tree classification algorithm | |
| US20160359880A1 (en) | Geo visualization of network flows | |
| CN108595461B (en) | Interest exploration method, storage medium, electronic device and system | |
| WO2021082589A1 (en) | Content check model training method and apparatus, video content check method and apparatus, computer device, and storage medium | |
| CN105447147B (en) | A kind of data processing method and device | |
| CN102724059A (en) | Website operation state monitoring and abnormal detection based on MapReduce | |
| CN104488231A (en) | Real-time network monitoring and subscriber identification with an on-demand appliance | |
| CN109905696A (en) | A kind of recognition methods of the Video service Quality of experience based on encryption data on flows | |
| CN102315974A (en) | Stratification characteristic analysis-based method and apparatus thereof for on-line identification for TCP, UDP flows | |
| CN103595576A (en) | Interconnection port ICP flow statistical system and method based on content provider identifications | |
| CN103186600A (en) | Specific analysis method and device of Internet public sentiment | |
| CN110417729A (en) | A kind of service and application class method and system encrypting flow | |
| CN106789242A (en) | A kind of identification application intellectual analysis engine based on mobile phone client software behavioral characteristics storehouse | |
| CN112000889A (en) | Information gathering and presenting system | |
| Fan et al. | An interactive visual analytics approach for network anomaly detection through smart labeling | |
| CN106559498A (en) | Air control data collection platform and its collection method | |
| CN114422211B (en) | HTTP malicious traffic detection method and device based on graph attention network | |
| US20170272453A1 (en) | User interface for displaying network analytics | |
| CN108270637A (en) | A kind of Website quality multilayer drills through system and method | |
| CN106372171A (en) | Real-time data processing method of monitoring platform | |
| CN106910082A (en) | It is a kind of to transfer the method and apparatus that advertisement is pushed |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190219 |
|
| RJ01 | Rejection of invention patent application after publication |