CN106789297A - Predicting network flow system and its method for predicting based on neutral net - Google Patents
Predicting network flow system and its method for predicting based on neutral net Download PDFInfo
- Publication number
- CN106789297A CN106789297A CN201611249158.8A CN201611249158A CN106789297A CN 106789297 A CN106789297 A CN 106789297A CN 201611249158 A CN201611249158 A CN 201611249158A CN 106789297 A CN106789297 A CN 106789297A
- Authority
- CN
- China
- Prior art keywords
- data
- network
- predicting
- neutral net
- module
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/147—Network analysis or design for predicting network behaviour
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
-
- 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)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Environmental & Geological Engineering (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention discloses a kind of predicting network flow system based on neutral net and network flow prediction method, belongs to field of computer technology.The predicting network flow system includes data acquisition module, data preprocessing module and predicting network flow module, data-acquisition submodule realizes the Real-time Collection to various flow informations in network using the network mode based on Port Mirroring, data preprocessing module respectively preserves the data for collecting and does normalized, so that sampled data values are between 0~1, for prediction module provides pure data.Volume forecasting module determines the topological structure and network parameter of the neutral net for volume forecasting according to the IP network data on flows for collecting, and method using neutral net is predicted, and draws and predicts the outcome.The present invention can be monitored detection and analyze to various backbone networks, and the network exception event in monitor in real time, detection backbone network, realization gives warning in advance to network abnormal situation.
Description
Technical field
The present invention relates to technical field of the computer network, more particularly to a kind of predicting network flow system based on neutral net
System and its method for predicting.
Background technology
With the extreme enrichment developed rapidly with Network of the internet with IP as core technology, internet oneself through by
Gradually develop into bearing multiple service, the global public information transmission platform of service catergories of user colony.But due to IP associations
The intrinsic connectionless characteristic of view and traditional IP are done one's best the Service Principle of (Best-Effort), traditional internet without
Normal direction user provides effective service quality (QoS) and ensures, can not realize the effective monitoring and management of Internet resources.Therefore,
Monitoring and management to network have turned into the hot issue of research.
Realize that network QoS control needs the current operation conditions of timely awareness network, to take appropriate measures in time,
This is accomplished by carrying out network flow programming method.According to detected historical data, by setting up appropriate discharge model, in the future
Data on flows be predicted.By the result predicted, the approximate trend of flow in future can be obtained, can so avoid network event
The generation of barrier.Therefore network flow programming method is the basis for implementing effective QoS controls, and predicting network flow technology is also subject to more next
More attention.
Present Network Traffic Forecast Model is divided into linear prediction and nonlinear prediction, and wherein ARIMA is used as linear prediction
Representative be widely used.But because ARIMA is suitable only for processing linear wide stationary processes, real network traffics are big
Part does not comply with the precondition of ARIMA forecast models, therefore the precision of prediction of network traffics ARIMA forecast models is relatively low,
And whole features of network cannot be accurately depicted.So, people concentrate on research nonlinear method.This method due to
The nonlinear characteristic of network traffics can be depicted and the performance more excellent than linear prediction method is shown.
The content of the invention
The problem to be solved in the present invention is to provide one kind for predicting network flow, from the feature of analysis network traffics
Hand, establishes the method for predicting based on BP neural network, and the time series to network flow data is modeled and prediction.
Devise data acquisition, data prediction and the big module of volume forecasting three, Memorability and powerful using BP neural network
Habit ability carries out short-term forecast to network traffics, is that network monitoring and management provide foundation.
A kind of predicting network flow system based on neutral net of present invention offer, including data acquisition module, data are pre-
Processing module and predicting network flow module；The data acquisition module includes several data collecting card, data collecting card
The information for collecting is sent to data prediction submodule, and the data preprocessing module, will after sample data is pre-processed
Data is activation after treatment is predicted to predicting network flow module using the method for neutral net, is drawn and is predicted the outcome.
Further improvement of these options, the data collecting card includes that 100/1000MFE capture cards, atm link are adopted
Truck and SDH link acquisition cards.
The present invention also provides a kind of stream of the predicting network flow system based on neutral net based on described in such scheme
Amount Forecasting Methodology, comprises the following steps：
(1) using the monitoring to network egress data of data acquisition module, the network mode based on Port Mirroring is used
To realize the Real-time Collection of network data, data image server is connected to the mirror port of experimental situation middle outlet interchanger
On, the data for flowing through interchanger can just be arrived by mirror port by data image server Real-time Collection；
(2) in data preprocessing module, the data on flows that data acquisition module is collected is preserved and is normalized
Treatment, for predicting network flow submodule provides pure data；
(3) in predicting network flow submodule, the IP network data on flows collected to data preprocessing module is utilized
The method of neutral net is predicted, and draws and predicts the outcome, and the main task of the module is to determine the nerve for volume forecasting
The structure of network, the neural network structure is BP neural network structure, for determining network topology structure and network parameter
Determine two by comparing the sample training result under different condition to determine predicting network flow value.
Further improvement of these options, in step (2), the formula that the data on flows is normalized
For it neutralizes the maximum and minimum value being respectively in sample data, is raw sample data, is the numerical value after conversion.
Further improvement of these options, in step (3), comprises the following steps that：
(1) BP neural network topological structure and network parameter are determined, using three layers of neural network structure pattern, i.e. hidden layer
Number is 1, and input layer number is 6, the i.e. data on flows at the first six of predicted value moment, and the number of hidden nodes is 11, output layer section
It is 1 to count, i.e. predicted value；Initial weight and threshold values are generated with random function；
(2) parameter initialization is carried out, including maximum frequency of training Max_epoch, error precision, node in hidden layer, just
Beginning weights, threshold value, the isoparametric initialization of initial learning rate；
(3) to one training sample of BP network inputs, the input and output value of each layer is calculated, calculates each layer output valve and expect
The error of result；
(4) according to the connection weight and threshold value of adaptive strain learning rate algorithm amendment neutral net；
(5) if error precision is unsatisfactory for requiring, and frequency of training is less than its maximum frequency of training, then return to (2)；
(6) required if error precision meets, you can predict flow service condition.
The beneficial effects of the present invention are：
(1) data-acquisition submodule of the present invention is realized to various flows in network using the network mode based on Port Mirroring
The Real-time Collection of information, and according to the difference of access network, the function of the module is realized using various network data acquisition cards, can
With the monitor and detection by data acquisition module to 100M/1000M/2.5G/10G backbone networks and analysis.Data prediction mould
Block respectively preserves the data for collecting and does normalized so that sampled data values are that prediction module is carried between 0~1
For pure data.Volume forecasting module determines the neutral net for volume forecasting according to the IP network data on flows for collecting
Topological structure and network parameter, and method using neutral net is predicted, and draws and predicts the outcome.
(2) present invention can be to the network exception event in backbone network monitor in real time, detection backbone network；It is key when finding
Occur abnormal conditions in network, just can in time notify keeper, processed by keeper, it is to avoid appearance is processed not in time, shadow
Ringing other work is carried out.
(3) keeper can predict according to method of the present invention flow distribution and the actual capacity limit of network line
System, so as to realize giving warning in advance Network Abnormal.
Brief description of the drawings
The frame diagram of Fig. 1 predicting network flow systems based on neutral net of the present invention.
The flow of the network flow prediction method of Fig. 2 predicting network flow systems based on neutral net of the present invention
Figure.
Specific embodiment
For technological means, creation characteristic, reached purpose and effect for making present invention realization are easy to understand, with reference to
Specific embodiment, is expanded on further the present invention.
The predicting network flow system based on neutral net described in the present embodiment includes that data acquisition module, data are located in advance
Reason module and predicting network flow module.
Data-acquisition submodule, mainly completes the collection to various flow informations in network, according to the difference of access network,
The function of submodule is realized using various network data acquisition cards.The data collecting card include 100/1000MFE capture cards,
Atm link capture card and SDH link acquisition cards.
The key of data acquisition module is the monitoring to network egress data.It is main to use the network side based on Port Mirroring
Formula realizes the Real-time Collection of network data.Data image server is connected to the mirror image end of experimental situation middle outlet interchanger
On mouth, the data for flowing through interchanger can just be arrived by mirror port by data image server Real-time Collection.For original number
According to bag capture mainly call WinPcap develop packet interface packet capture is carried out to the machine network interface card port, respectively according to IP,
TCP, UDP pack arrangement definition data structure, and connect the packet that MySQL database will collect and store in MySQL data
In storehouse.The process description of wherein Packet capturing is as follows：
(1) list of devices is obtained first by pcap_findalldevs () function；
(2) equipment interface that user selects is opened using pcap_open_live () function, and is set to promiscuous mode；
(3) pcap_compile () and pcap_setfilter () function setup filter are called；
(4) pcap_loop () is called to be circulated capture to selected device, and defined function is at the bag that captures
Reason；
(5) network interface card is closed.
The data that data acquisition module is collected mainly are preserved and normalized by data preprocessing module, are
Predicting network flow module provides pure data.Mainly give the definition of data structure below, and database table is determined
Details is realized in pretreatment of justice and data etc..
1. data are preserved.In a upper module, we have been able to be captured by the institute of network interface card by WinPcap interfaces
There is packet, and be provided with filter, only receive IP bags, non-IP packet is excluded.Next, needing to the IP bags that capture
Relevant information is stored so that subsequent analysis operation to storage format, it is necessary to be defined.The tables of data being related to mainly is wrapped
Include：Protocol (IP network traffic type list), IPToName (network user's table), UserTraffic (traffic statistics table).Its
In, port and the Apply Names corresponding table of network application are commonly used in Protocol token records, and IPToName tables are used for recording in experiment
To the mapping of title, the data in this two tables are loaded the IP address of each server in system initialization in environment, and
The corresponding interface is provided by system to be safeguarded, kept relative stability in whole system running.UserTraffic tables are pre-
The core of initial data during survey, it is responsible for the record descriptive information related to IP bags, i.e.,：Time, IP address, IP bags
Transmitting-receiving byte number, type of service etc..By the way of being recorded once every 5 seconds, by the statistics and note of network traffics within 5 seconds
In entering database.
2. data prediction.After data are preserved, the IP network flow value in a period of time is have recorded in database,
But, it can be seen from before signal above to BP algorithm to the description of transmittance process and activation primitive characteristic, if neuron
Total input with threshold value far apart, because the activation primitive of each neuron has saturation nonlinearity characteristic, can cause neuron
Output falls in zone of saturation, otherwise the reality output of such network is the maximum of activation primitive, or for activation primitive most
Small value so that the derivative value of output (goes to zero) very little.So as to cause the modification amount very little of weights, not only pace of learning delays
Slowly, and network be difficult convergence.Therefore, before neural network prediction is carried out, to avoid, initial data is excessive to cause network fiber crops
Numbness, will be normalized to initial data, for predicted value, because amplitude of variation is larger, and also should not be directly as nerve
The output of network, therefore still need to carry out necessary normalization.
For monopole type Sigmoid functions, its output changes between 0~1, and only when input is, it is defeated
Go out and just reach (0,1), proper output area is 0.1~0.9.The input of network should also try one's best makes Sigmoid function works
Make in non-saturated region, therefore, following normalized is carried out to sample data：
Wherein, xmaxAnd xminMaximum and minimum value respectively in sample data, xtIt is raw sample data, xt' it is change
Numerical value after changing.So not only avoid input data and fall into zone of saturation, also maintain original feature of data.Work as nerve net
After network treatment terminates, then do renormalization computing.Network is exported, output data is made with original using equation below to output data
Within same region, the formula of renormalization is beginning data：
xt=xt'(xmax-xmin)+xmin
Obviously, this is the inverse process that initial data is normalized.
Predicting network flow module, is mainly carried out pre- to the IP network data on flows for collecting using the method for neutral net
Survey, draw and predict the outcome.The main task of the module is to determine the structure of the neutral net for volume forecasting.
1.BP neural network models
The determination of BP neural network model mainly includes the aspect of determination two of the determination of network topology structure and network parameter,
Determined by comparing the sample training result under different condition.
(1) BP neural network topological structure determines.By test of many times, the system prediction submodule uses three layers of nerve
Network structure pattern, i.e. hidden layer number are 1.Input layer number is 6, the i.e. data on flows at the first six of predicted value moment.Hidden layer
Nodes are 11, and output layer nodes are 1, i.e. predicted value.
(2) neural network parameter determines.From trainlm as training function, trasig is learned as hidden layer transmission function
The value for practising speed is 0.025, and initial weight and threshold values are generated with random function, and the value for measuring the factor is 0.9, and cycle-index is
2500, using the average of error sum of squares as accuracy estimating index.
2. volume forecasting step
When being used for modeling and forecasting using neutral net, its step is as follows：
(1) BP neural network topological structure and network parameter are determined, using three layers of neural network structure pattern, i.e. hidden layer
Number is 1, and input layer number is 6, the i.e. data on flows at the first six of predicted value moment, and the number of hidden nodes is 11, output layer section
It is 1 to count, i.e. predicted value；Initial weight and threshold values are generated with random function；
(2) parameter initialization is carried out, including maximum frequency of training Max_epoch, error precision, node in hidden layer, just
Beginning weights, threshold value, the isoparametric initialization of initial learning rate；
(3) to one training sample of BP network inputs, the input and output value of each layer is calculated, calculates each layer output valve and expect
The error of result；
(4) according to the connection weight and threshold value of adaptive strain learning rate algorithm amendment neutral net；
(5) if error precision is unsatisfactory for requiring, and frequency of training is less than its maximum frequency of training, then return to (2)；
(6) required if error precision meets, you can predict flow service condition.
Data-acquisition submodule of the present invention is realized to various flow letters in network using the network mode based on Port Mirroring
The Real-time Collection of breath, and according to the difference of access network, the function of the module is realized using various network data acquisition cards, can be with
The monitor and detection to 100M/1000M/2.5G/10G backbone networks and analysis by data acquisition module.Data preprocessing module
The data for collecting are preserved respectively and normalized is done so that sampled data values between 0~1, for prediction module is provided
Pure data.Volume forecasting module determines the neutral net for volume forecasting according to the IP network data on flows for collecting
Topological structure and network parameter, and method using neutral net is predicted, and draws and predicts the outcome.And the present invention can be to bone
Network exception event in dry network real-time monitoring, detection backbone network；There are abnormal conditions in backbone network is found, just may be used
Keeper is notified in time, is processed by keeper, it is to avoid appearance is processed not in time, the other work of influence are carried out.Keeper
Flow distribution is provided, realization gives warning in advance to network abnormal situation,
Those of ordinary skill in the art it should be appreciated that the embodiment of the above be intended merely to explanation the present invention,
And be not used as being limitation of the invention, as long as in spirit of the invention, the change to embodiment described above
Change, modification will all fall in scope of the presently claimed invention.
Claims (5)
1. a kind of predicting network flow system based on neutral net, it is characterised in that：Locate in advance including data acquisition module, data
Reason module and predicting network flow module；The data acquisition module includes several data collecting card, and data collecting card is adopted
The information for collecting is sent to data prediction submodule, and the data preprocessing module is located in advance by preserving and carrying out sample data
After reason, by the data is activation after treatment to predicting network flow module, volume forecasting is carried out using the method for neutral net, drawn
Predict the outcome.
2. a kind of predicting network flow system based on neutral net according to claim 1, it is characterised in that：The number
Include 100/1000MFE capture cards, atm link capture card and SDH link acquisition cards according to capture card.
3. a kind of method for predicting of the predicting network flow system based on neutral net as claimed in claim 1, it is special
Levy and be, comprise the following steps：
(1) using the monitoring to network egress data of data acquisition module, using the network mode based on Port Mirroring come real
The Real-time Collection of existing network data, data image server is connected on the mirror port of experimental situation middle outlet interchanger,
The data for flowing through interchanger can just be arrived by mirror port by data image server Real-time Collection；
(2) in data preprocessing module, the data on flows that data acquisition module is collected is preserved and normalized,
For predicting network flow submodule provides pure data；
(3) in predicting network flow submodule, the IP network data on flows collected to data preprocessing module is using nerve
The method of network is predicted, and draws and predicts the outcome, and the main task of the module is to determine the neutral net for volume forecasting
Structure and network parameter, the neural network structure be BP neural network structure, by compare the sample under different condition instruct
Practice result to determine predicting network flow value.
4. the method for predicting of a kind of predicting network flow system based on neutral net according to claim 3, its
It is characterised by：In step (2), the formula that the data on flows is normalized isWherein xmaxWith
xminMaximum and minimum value respectively in sample data, xtIt is raw sample data, x 'tIt is the numerical value after conversion.
5. the method for predicting of a kind of predicting network flow system based on neutral net according to claim 3, its
It is characterised by：In step (3), comprise the following steps that：
(1) BP neural network topological structure and network parameter are determined, using three layers of neural network structure pattern, i.e. hidden layer number is
1, input layer number is 6, the i.e. data on flows at the first six of predicted value moment, and the number of hidden nodes is 11, output layer nodes
Be 1, i.e. predicted value；Initial weight and threshold values are generated with random function；
(2) parameter initialization, including maximum frequency of training Max_epoch, error precision, node in hidden layer, initial power are carried out
Value, threshold value, the isoparametric initialization of initial learning rate；
(3) to one training sample of BP network inputs, the input and output value of each layer is calculated, calculates each layer output valve and expected result
Error；
(4) according to the connection weight and threshold value of adaptive strain learning rate algorithm amendment neutral net；
(5) if error precision is unsatisfactory for requiring, and frequency of training is less than its maximum frequency of training, then return to (2)；
(6) required if error precision meets, you can predict flow service condition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611249158.8A CN106789297A (en) | 2016-12-29 | 2016-12-29 | Predicting network flow system and its method for predicting based on neutral net |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611249158.8A CN106789297A (en) | 2016-12-29 | 2016-12-29 | Predicting network flow system and its method for predicting based on neutral net |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106789297A true CN106789297A (en) | 2017-05-31 |
Family
ID=58928247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611249158.8A Pending CN106789297A (en) | 2016-12-29 | 2016-12-29 | Predicting network flow system and its method for predicting based on neutral net |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106789297A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107370732A (en) * | 2017-07-14 | 2017-11-21 | 成都信息工程大学 | System is found based on neutral net and the industrial control system abnormal behaviour of optimal recommendation |
| CN108173704A (en) * | 2017-11-24 | 2018-06-15 | 中国科学院声学研究所 | A kind of method and device of the net flow assorted based on representative learning |
| CN108494746A (en) * | 2018-03-07 | 2018-09-04 | 长安通信科技有限责任公司 | A kind of network port Traffic anomaly detection method and system |
| CN108540331A (en) * | 2018-04-26 | 2018-09-14 | 南京邮电大学 | Based on the network flow prediction method for improving ESN |
| CN108647778A (en) * | 2018-05-09 | 2018-10-12 | 天津大学 | Dynamic prediction method for drainage flow of drainage port of urban rainwater system |
| CN108985446A (en) * | 2018-07-24 | 2018-12-11 | 百度在线网络技术（北京）有限公司 | method and device for alarm |
| CN109547251A (en) * | 2018-11-27 | 2019-03-29 | 广东电网有限责任公司 | A kind of operation system failure and performance prediction method based on monitoring data |
| CN109639524A (en) * | 2018-12-13 | 2019-04-16 | 国网上海市电力公司 | Communication network data method for visualizing, device and equipment based on volume forecasting |
| CN109981332A (en) * | 2018-12-03 | 2019-07-05 | 绥化学院 | Network flow prediction method and device based on wavelet neural network |
| CN110545208A (en) * | 2019-09-23 | 2019-12-06 | 电子科技大学 | Network traffic prediction method based on LSTM |
| CN111130890A (en) * | 2019-12-26 | 2020-05-08 | 深圳市高德信通信股份有限公司 | Network flow dynamic prediction system |
| CN111884883A (en) * | 2020-07-29 | 2020-11-03 | 北京宏达隆和科技有限公司 | Quick auditing processing method for service interface |
| CN111988239A (en) * | 2020-08-21 | 2020-11-24 | 哈尔滨工业大学 | Method for acquiring pure software flow for Android application |
| CN112532643A (en) * | 2020-12-07 | 2021-03-19 | 长春工程学院 | Deep learning-based traffic anomaly detection method, system, terminal and medium |
| CN113453096A (en) * | 2021-06-04 | 2021-09-28 | 中国联合网络通信集团有限公司 | Method and device for predicting PON port flow of passive optical network |
| CN114205164A (en) * | 2021-12-16 | 2022-03-18 | 北京百度网讯科技有限公司 | Traffic classification method and device, training method and device, equipment and medium |
| CN116996403A (en) * | 2023-09-26 | 2023-11-03 | 深圳市乙辰科技股份有限公司 | Network traffic diagnosis method and system applying AI model |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102032935A (en) * | 2010-12-07 | 2011-04-27 | 杭州电子科技大学 | Soft measurement method for sewage pumping station flow of urban drainage converged network |
| CN103095496A (en) * | 2013-01-10 | 2013-05-08 | 周亚建 | Prediction method and device for network flow |
-
2016
- 2016-12-29 CN CN201611249158.8A patent/CN106789297A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102032935A (en) * | 2010-12-07 | 2011-04-27 | 杭州电子科技大学 | Soft measurement method for sewage pumping station flow of urban drainage converged network |
| CN103095496A (en) * | 2013-01-10 | 2013-05-08 | 周亚建 | Prediction method and device for network flow |
Non-Patent Citations (2)
| Title |
|---|
| 刘杰等: ""基于BP神经网络的非线性网络流量预测"", 《计算机应用》 * |
| 王艳萍等: ""基于改进BP神经网络的刮板输送机负载预测方法研究"", 《矿山机械》 * |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107370732A (en) * | 2017-07-14 | 2017-11-21 | 成都信息工程大学 | System is found based on neutral net and the industrial control system abnormal behaviour of optimal recommendation |
| CN107370732B (en) * | 2017-07-14 | 2021-08-17 | 成都信息工程大学 | Abnormal behavior discovery system of industrial control system based on neural network and optimal recommendation |
| CN108173704A (en) * | 2017-11-24 | 2018-06-15 | 中国科学院声学研究所 | A kind of method and device of the net flow assorted based on representative learning |
| CN108494746A (en) * | 2018-03-07 | 2018-09-04 | 长安通信科技有限责任公司 | A kind of network port Traffic anomaly detection method and system |
| CN108540331B (en) * | 2018-04-26 | 2021-07-13 | 南京邮电大学 | Network flow prediction method based on improved ESN |
| CN108540331A (en) * | 2018-04-26 | 2018-09-14 | 南京邮电大学 | Based on the network flow prediction method for improving ESN |
| CN108647778A (en) * | 2018-05-09 | 2018-10-12 | 天津大学 | Dynamic prediction method for drainage flow of drainage port of urban rainwater system |
| CN108647778B (en) * | 2018-05-09 | 2022-02-25 | 天津大学 | Dynamic prediction method for drainage flow of drainage port of urban rainwater system |
| CN108985446A (en) * | 2018-07-24 | 2018-12-11 | 百度在线网络技术（北京）有限公司 | method and device for alarm |
| CN109547251A (en) * | 2018-11-27 | 2019-03-29 | 广东电网有限责任公司 | A kind of operation system failure and performance prediction method based on monitoring data |
| CN109547251B (en) * | 2018-11-27 | 2022-05-10 | 广东电网有限责任公司 | Service system fault and performance prediction method based on monitoring data |
| CN109981332A (en) * | 2018-12-03 | 2019-07-05 | 绥化学院 | Network flow prediction method and device based on wavelet neural network |
| CN109639524A (en) * | 2018-12-13 | 2019-04-16 | 国网上海市电力公司 | Communication network data method for visualizing, device and equipment based on volume forecasting |
| CN110545208A (en) * | 2019-09-23 | 2019-12-06 | 电子科技大学 | Network traffic prediction method based on LSTM |
| CN110545208B (en) * | 2019-09-23 | 2021-10-15 | 电子科技大学 | Network traffic prediction method based on LSTM |
| CN111130890A (en) * | 2019-12-26 | 2020-05-08 | 深圳市高德信通信股份有限公司 | Network flow dynamic prediction system |
| CN111884883A (en) * | 2020-07-29 | 2020-11-03 | 北京宏达隆和科技有限公司 | Quick auditing processing method for service interface |
| CN111988239A (en) * | 2020-08-21 | 2020-11-24 | 哈尔滨工业大学 | Method for acquiring pure software flow for Android application |
| CN112532643A (en) * | 2020-12-07 | 2021-03-19 | 长春工程学院 | Deep learning-based traffic anomaly detection method, system, terminal and medium |
| CN112532643B (en) * | 2020-12-07 | 2024-02-20 | 长春工程学院 | Flow anomaly detection method, system, terminal and medium based on deep learning |
| CN113453096A (en) * | 2021-06-04 | 2021-09-28 | 中国联合网络通信集团有限公司 | Method and device for predicting PON port flow of passive optical network |
| CN113453096B (en) * | 2021-06-04 | 2022-12-13 | 中国联合网络通信集团有限公司 | Method and device for predicting PON port flow of passive optical network |
| CN114205164A (en) * | 2021-12-16 | 2022-03-18 | 北京百度网讯科技有限公司 | Traffic classification method and device, training method and device, equipment and medium |
| CN116996403A (en) * | 2023-09-26 | 2023-11-03 | 深圳市乙辰科技股份有限公司 | Network traffic diagnosis method and system applying AI model |
| CN116996403B (en) * | 2023-09-26 | 2023-12-15 | 深圳市乙辰科技股份有限公司 | Network traffic diagnosis method and system applying AI model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106789297A (en) | Predicting network flow system and its method for predicting based on neutral net | |
| CN102957579B (en) | A kind of exception flow of network monitoring method and device | |
| Kayacik et al. | On the capability of an SOM based intrusion detection system | |
| CN110138787A (en) | A kind of anomalous traffic detection method and system based on hybrid neural networks | |
| CN109905696B (en) | Method for identifying video service experience quality based on encrypted flow data | |
| CN108877839A (en) | The method and system of perceptual evaluation of speech quality based on voice semantics recognition technology | |
| CN109873726B (en) | Robust service quality prediction and guarantee method based on deep learning in SDN | |
| CN101355504A (en) | Method and apparatus for confirming user behavior | |
| CN109325638A (en) | A kind of SDN method for predicting based on RBF neural | |
| CN106941493A (en) | A kind of network security situation awareness result output intent and device | |
| CN110138786A (en) | Web method for detecting abnormality and system based on SMOTETomek and LightGBM | |
| CN107040551A (en) | A kind of industry control network safe early warning method and system | |
| CN107483473A (en) | A kind of low speed Denial of Service attack data-flow detection method of cloud environment | |
| CN108632269A (en) | Detecting method of distributed denial of service attacking based on C4.5 decision Tree algorithms | |
| CN107767055A (en) | A kind of mass-rent result assemblage method and device based on collusion detection | |
| CN107332715A (en) | Active Performance test plus the network application system and its implementation of passive shunting control | |
| CN106656616A (en) | Whole network flow analysis method of computer network | |
| CN109150859A (en) | A kind of Botnet detection method flowing to similitude based on network flow | |
| CN106780248A (en) | City internet of things equipment operation control system and method based on big data technology | |
| CN114499979A (en) | SDN abnormal flow cooperative detection method based on federal learning | |
| TW200522627A (en) | Methodology of predicting distributed denial of service based on gray theory | |
| CN109284740A (en) | Method, apparatus, equipment and the storage medium that mouse feelings are counted | |
| CN110266603A (en) | Authentication business network flow analysis system and method based on http protocol | |
| CN112734699A (en) | Article state warning method and device, storage medium and electronic device | |
| CN114244691A (en) | Video service fault positioning method and device and electronic equipment |
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: 20170531 |
|
| RJ01 | Rejection of invention patent application after publication |