CN109697332A - A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method - Google Patents
A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method Download PDFInfo
- Publication number
- CN109697332A CN109697332A CN201910031830.3A CN201910031830A CN109697332A CN 109697332 A CN109697332 A CN 109697332A CN 201910031830 A CN201910031830 A CN 201910031830A CN 109697332 A CN109697332 A CN 109697332A
- Authority
- CN
- China
- Prior art keywords
- state
- exception monitoring
- point
- stream calculation
- calculation system
- 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
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention proposes a kind of exception monitoring scheme of stream calculation system based on unsupervised learning method, belong to distributed real-time systems exception monitoring field, specifically include: system action describing module constructs exception monitoring module, the big module of online adaptive exception monitoring three based on small sample constraint condition.First, this method is based on carrying out on the basis of conversion operation obtains compound event primitive event using event processing, to obtain state-event data target and physical state data target, system action state index space is obtained by the collected data target of time window technological incorporation again, realizes the behavior description of stream calculation system；Secondly, proposing a kind of non-supervisory statistical analysis technique, building realizes the exception monitoring of the unbalanced data of stream calculation system based on the exception monitoring model under small sample constraint condition；Finally, proposing a kind of online adaptive exception monitoring model, adjust automatically network structure updates cluster centre, realizes online adaptive exception monitoring.
Description
Technical field
The invention belongs to distributed real-time systems exception monitoring fields, relate generally to distributed real-time systems on-line monitoring mould
The foundation of type, the particularly normal clustering method of a kind of polyisocyanate based on system action state.
Background technique
With the continuous continuous expansion expanded with big data application field of big data industry size, stream calculation system conduct
A kind of emerging big data tupe is increasingly becoming a kind of important tool of people's production application.In large-scale production application,
Simultaneously real-time online handles data to thousands of a nodes, this makes the safety problem of stream calculation system pay close attention to weight as everybody
Point, safety accident will bring huge economic loss and immeasurable invisible loss, therefore, stream calculation system to society
The emphasis of credibility, reliability, safety as academia and industry research.The exception of this paper Main Analysis stream calculation system
Monitoring problem proposes a kind of online exception monitoring scheme.
At this stage, the exception monitoring of stream calculation system has become the research hotspot of academia.For stream calculation system
Exception monitoring is generally basede on statistical method, its advantage is that calculating light weight.Including abnormality test, such as ESD method and the side K-sigma
Method slides threshold method, and change point monitoring method etc. is all a kind of efficient method, and does not require user's custom parameter, but this
A little methods are concentrated on free air anomaly, need to limit their dependences for the time.And time-based exception monitoring side
Method can widely be divided into three classes: supervised learning method, semi-supervised learning method and unsupervised learning method.Based on measure of supervision
The characteristics of carrying out exception monitoring is analyzed for entire data set, and the precision of exception monitoring is higher, but can not achieve
Line exception monitoring.Such as the GPCA that document " Generalized principal component analysis " is proposed, document
" HOT SAX:Efficiently finding the most unusual time se-ries subsequence " is proposed
SAX method is to generate symbol by decomposing all time serieses, is finally carrying out exception monitoring, document " Generic and
Scalable Framework for Au-tomated Time-series Anomaly Detection " propose the side EGADS
Method, although effect may very well under certain circumstances for these technologies, they are batch processing modes.Based on semi-supervised learning
It is the improvement done on measure of supervision that method, which carries out exception monitoring, and using less training set training classifier, but it is lacked
It is insensitive for novel class or unknown exception for putting, such as document " A Novel Transductive SVM for the
Semisupervised Classification of Remote-Sensing Images " propose S3VM.Unsupervised learning
The shortcomings that being marked to data, overcoming supervised learning is not needed, becomes and the one of exception monitoring is carried out for stream calculation system
A new direction, such as document " Self-adaptive and dynamic clustering for online anomaly
Detection " it proposes to carry out exception monitoring using Kmeans, while this article introduces neural network algorithm SOM and solves unevenness
The undesirable problem of the Clustering Effect of weighing apparatus sample, but the topological structure of neural network is unstable, the superiority and inferiority of exception monitoring is very
Mostly rely on the quality of weight setting.Therefore, a kind of non_monitor algorithm suitable for Small Sample Database exception monitoring is found
It is extremely important for the exception monitoring of stream calculation system.
Summary of the invention
In view of this, being directed to the defect of above-mentioned existing stream calculation system exception monitoring technology, present invention mainly solves flowing
Small sample method for monitoring abnormality in computing system, the present invention propose a kind of stream calculation system based on unsupervised learning method
Online exception monitoring scheme, comprising:
S1: acquisition stream calculation system action feature is analyzed and describes stream calculation system action state；
S2: according to the unbalance response of abnormal data, building is based on exception monitoring model online under small sample constraint condition；
S3: being based on initial exception monitoring model, and dynamic adjusts exception monitoring network structure size and cluster centre, realizes stream
The adaptive exception monitoring of computing system.
Preferably, stream calculation system action feature is acquired, analyzes and describes stream calculation system action state, specifically include:
S11: by direct fault location mode, the stream calculation working state of system under true environment is simulated；
S12: conversion operation is carried out to primitive event using event processing and obtains compound event, and collection event state
Data target, wherein primitive event, also referred to as simple event refer to and directly generated by equipment, can be compound with data measured directly
Event refers to be combined by simple event, the general level information with service logic or rule, event processing packet
The operating technologies such as tree, figure, automatic machine, petri net and Complex event processing are included, it, will be simple by user's customized event stream
Event is converted to complicated event；
S13: under operation, stream calculation system performance measure, i.e. physical state data target are acquired；
S14: the state-event data target and physical state data target that will acquire using time window technology are melted
It is combined into system features state space.
Preferably, the unbalance response of abnormal data, building are based on online exception monitoring model under small sample constraint condition,
It specifically includes:
S21: building calculates some time etching system state and other shapes based on the exception monitoring model under small sample constraint condition
The moment state is determined as the state consistency with its maximum similarity, that is, solves this problem by the similarity between state:
S22: in view of that there should be larger similarity between any two state similitude, therefore a minimum phase is set
Like degree threshold value, then have:
si,j≥smin, wherein i, j ∈ Cm；
S23: since the jitter of system be easy to cause state singular value occur, a minimum state number threshold is set
Value, then has:
|Cm|≥Nmin, wherein Cm∈C；
S24: since system normal operation belongs to ordinary circumstance, and abnormality belongs to sample present event, is consequently belonging to same
A kind of state similarity is far longer than the similarity between inhomogeneity, then has:
Wherein, si,jIndicate the state similarity of i state and j state, m indicates status categories, sminIndicate minimum similarity degree
Threshold value, CmIndicate that there is state set similar with state m, | Cm| indicate the number with state set similar with m, NminTable
Show that state minimum state quantity threshold, C indicate that all state sets, d indicate state；
S25: being clustered using the realization of cluster centre algorithm is quickly searched, and realizes that quickly lookup gathers using statistical analysis technique
The automatic cluster of class CENTER ALGORITHM finds maximum similarity classification；
Preferably, it is based on initial exception monitoring model, dynamic adjusts exception monitoring network structure size and cluster centre, real
Existing stream calculation system self-adaption exception monitoring, specifically includes:
S31: if the data point newly to arrive and neighbours' similarity are greater than similarity threshold si,j≥smi n, i.e. ρi≤e-nWhen, then
Directly judge that the point is similar with nearest-neighbors point, if the data newly to arrive and neighbours' similarity are less than similarity threshold si,j<
smin, i.e. ρi>e-nWhen, its classification can not be judged, then adjust each point weight matrix size, adjust network structure size；
S32: the minimum range to local density and between higher density point carries out descending sequence, to preceding k
Point carries out hypothesis testing, is judged as cluster centre point if null hypothesis is rejected, is otherwise non-central point, updates cluster centre point
Set；
S33: it for non-cluster central point, assigns them to the higher minimum range point of density and carries out reunion class.The present invention
The advantages of and have the beneficial effect that:
1. the present invention can preferably portray stream calculation system action by building High Dimensional Systems state index；2. hair
A kind of bright method for monitoring abnormality proposed under the constraint condition based on small sample, can effectively monitor unbalanced data,
Improve exception monitoring efficiency；3. the present invention is by introducing statistical analysis side on the basis of quickly searching cluster centre (DPC) algorithm
Method, so that DPC algorithm has automatic cluster effect；4. the present invention proposes a kind of online adaptive exception monitoring algorithm, automatic to adjust
Whole network structure updates cluster centre, realizes online adaptive exception monitoring.Using the method, stream calculation system is being considered
Requirement of real-time under, improve system anomaly detection efficiency.
Detailed description of the invention
Fig. 1 is overall procedure block diagram of the invention.
Fig. 2 is the system features state diagram of stream calculation system.
Fig. 3 is based on small sample exception monitoring flow chart.
Fig. 4 is online adaptive exception monitoring flow chart.
Specific embodiment
It is clear to be more clear the purpose of the present invention, technical solution and advantage, referring to Figure of description, to hair
Bright specific implementation, which is made, further to be elaborated.It is described to implement to be only a part of the embodiments of the present invention.
The technical scheme to solve the above technical problems is that
It is as shown in Figure 1 overview flow chart of the invention, specifically includes: system action describing module, about based on small sample
Beam condition constructs exception monitoring module, the big module of online adaptive exception monitoring three.It illustrates and of the invention implemented in detail
Journey, including following three step:
S1: acquisition stream calculation system action feature is analyzed and describes stream calculation system action state；
S2: according to the unbalance response of abnormal data, building is based on exception monitoring model online under small sample constraint condition；
S3: being based on initial exception monitoring model, and dynamic adjusts exception monitoring network structure size and cluster centre, realizes stream
The adaptive exception monitoring of computing system.
Preferably, stream calculation system action feature is acquired, analyzes and describes stream calculation system action state, specifically include:
S11: by direct fault location mode, the stream calculation working state of system under true environment is simulated；
S12: conversion operation is carried out to primitive event using event processing and obtains compound event, and collection event state
Data target, wherein primitive event, also referred to as simple event refer to and directly generated by equipment, can be compound with data measured directly
Event refers to be combined by simple event, the general level information with service logic or rule, event processing packet
The operating technologies such as tree, figure, automatic machine, petri net and Complex event processing are included, it, will be simple by user's customized event stream
Event is converted to complicated event；
S13: under operation, stream calculation system performance measure, i.e. physical state data target are acquired；
S14: the state-event data target and physical state data target that will acquire using time window technology are melted
It is combined into system features state space, index is carried out using time window technology and merges as shown in Figure 2, wherein t indicates time, W
(t) time window function is indicated.
Preferably, according to the unbalance response of abnormal data, building is based on online exception monitoring under small sample constraint condition
Model specifically includes:
S21: building calculates some time etching system state and other shapes based on the exception monitoring model under small sample constraint condition
The moment state is determined as the state consistency with its maximum similarity, that is, solves this problem by the similarity between state:
S22: in view of that there should be larger similarity between any two state similitude, therefore a minimum phase is set
Like degree threshold value, then have:
si,j≥smin, wherein i, j ∈ Cm；
S23: since the jitter of system be easy to cause state singular value occur, a minimum state number threshold is set
Value, then has:
|Cm|≥Nmin, wherein Cm∈C；
S24: since system normal operation belongs to ordinary circumstance, and abnormality belongs to sample present event, is consequently belonging to same
A kind of state similarity is far longer than the similarity between inhomogeneity, then has:
Wherein, si,jIndicate the state similarity of i state and j state, m indicates status categories, sminIndicate minimum similarity degree
Threshold value, CmIndicate that there is state set similar with state m, | Cm| indicate the number with state set similar with m, NminTable
Show that state minimum state quantity threshold, C indicate that all state sets, d indicate state；
S25: being clustered using the realization of cluster centre algorithm is quickly searched, and realizes that quickly lookup gathers using statistical analysis technique
The automatic cluster of class CENTER ALGORITHM finds maximum similarity classification；
Preferably, as shown in figure 4, be based on initial exception monitoring model, dynamic adjust exception monitoring network structure size and
Cluster centre is realized stream calculation system self-adaption exception monitoring, is specifically included:
S31: if the data point newly to arrive and neighbours' similarity are greater than similarity threshold si,j≥smin, i.e. ρi≤e-nWhen, then
Directly judge that the point is similar with nearest-neighbors point, if the data point newly to arrive and neighbours' similarity are less than similarity threshold si,j<
smin, i.e. ρi>e-nWhen, its classification can not be judged, then adjust each point weight matrix size, adjust network structure size；
S32: the minimum range to local density and between higher density point carries out descending sequence, to preceding k
Point carries out hypothesis testing, is judged as cluster centre point if null hypothesis is rejected, is otherwise non-central point, updates cluster centre point
Set；
S33: it for non-cluster central point, assigns them to the higher minimum range point of density and carries out reunion class.
Preferably, as shown in figure 3, realizing cluster using cluster centre algorithm is quickly searched, and statistical analysis technique is utilized
It realizes the automatic cluster for quickly searching cluster centre algorithm, finds maximum similarity classification and specifically include:
S251: the index value local density of computing system significant condition
S252: the minimum range apart from higher density point is calculated
S253: building statistic γ=ρ * δ, and γ is ranked up according to sequence from big to small, obtain order statistics
Measure X1≥…≥Xn, wherein XiIndicate i-th of maximum γ value (1≤i≤n)；
S254: Η is enabled0,1,…,Η0,mFor X1,…,XmCorresponding null hypothesis, and assuming X not is extreme value abnormal point, i.e., is not
Cluster centre point；
S255: it enables and assumes statistic Rt=γt/γt+1If the individual key value of t-th of null hypothesis is rt=[1- (1- α)1/m]-1/λt,
Wherein 1≤t≤m, α are significance, and λ is power-law distribution coefficient；
S256: Γ is enabledt=Rt-rt, work as Γt> 0, then Η0,tRefusal, otherwise receives；
S227: according to ΓmJudge null hypothesis Η0,mWhether receive, continues to judge Η if receiving0,m-1, until null hypothesis
Η0,kRefuse (1≤k≤m), that is to say, that Η0,kIt is first refusal it is assumed that then showing X1,…,XkFor extreme value abnormal point,
Corresponding γ is cluster centre point；
S258: distributing to minimum range point apart from higher density point for remaining point, i.e., with apart from higher density point most
Small distance point is similar, and the classification of most of data set is normal class in cluster, and abnormal class is corresponding in turn to abnormal 1, abnormal 2 ...,
Abnormal n.
The above embodiment is interpreted as being merely to illustrate the present invention rather than limit the scope of the invention.?
After having read contents of the invention, technical staff can be made various changes or modifications the present invention, these equivalence changes
Right limited range of the present invention is equally fallen into modification.
Claims (5)
1. a kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method, feature the following steps are included:
S1: acquisition stream calculation system action feature is analyzed and describes stream calculation system action state；
S2: according to the unbalance response of abnormal data, building is based on exception monitoring model online under small sample constraint condition；
S3: being based on initial exception monitoring model, and dynamic adjusts exception monitoring network structure size and cluster centre, realizes stream calculation
System self-adaption exception monitoring.
2. a kind of exception monitoring scheme of stream calculation system based on unsupervised learning method according to claim 1,
It is characterized in that, the acquisition stream calculation system action feature, analyzes and describe stream calculation system action state and include:
S11: by direct fault location mode, the stream calculation working state of system under true environment is simulated；
S12: conversion operation is carried out to primitive event using event processing and obtains compound event, and collection event status data
Index；
S13: under operation, stream calculation system performance measure, i.e. physical state data target are acquired；
S14: the state-event data target and physical state data target that will acquire using time window technology are fused to
System features state space.
3. a kind of exception monitoring scheme of stream calculation system based on unsupervised learning method according to claim 1,
It is characterized in that, the unbalance response according to abnormal data, building is based on exception monitoring mould online under small sample constraint condition
Type includes:
S21: building is calculated similar between certain state and other states based on the exception monitoring model under small sample constraint condition
The state is determined as the state consistency with its maximum similarity, that is, solves this problem by degree:
S22: in view of that there should be larger similarity between any two state similitude, therefore a minimum similarity degree is set
Threshold value then has:
si,j≥smin, wherein i, j ∈ Cm；
S23: since the jitter of system be easy to cause state singular value occur, being arranged a minimum state quantity threshold,
Then have:
|Cm|≥Nmin, wherein Cm∈C；
S24: since system normal operation belongs to ordinary circumstance, and abnormality belongs to sample present event, is consequently belonging to same class
State similarity be far longer than the similarity between inhomogeneity, then have:
Wherein, si,jIndicate the state similarity of i state and j state, m indicates status categories, sminIndicate minimum similarity degree threshold value,
CmIndicate that there is state set similar with state m, | Cm| indicate the number with state set similar with m, NminIndicate shape
State minimum state quantity threshold, C indicate that all state sets, d indicate state；
S25: cluster is realized using cluster centre algorithm (DPC) is quickly searched, and is realized using statistical analysis technique and is quickly searched
The automatic cluster of cluster centre algorithm finds maximum similarity classification.
4. a kind of exception monitoring scheme of stream calculation system based on unsupervised learning method according to claim 1,
It is characterized in that, be based on initial exception monitoring model, dynamic adjusts exception monitoring network structure size and cluster centre, realizes flowmeter
Calculating system self-adaption exception monitoring includes:
S31: if the data point newly to arrive and neighbours' similarity are greater than similarity threshold, directly judge the point and nearest-neighbors point
It is similar, if the data point newly to arrive and neighbours' similarity are less than similarity threshold, its classification can not be judged, then adjustment is each
Point weight matrix size, adjusts network structure size；
S32: the minimum range to local density and between higher density point carries out descending sequence, clicks through to first k
Row hypothesis testing is judged as cluster centre point if null hypothesis is rejected, and is otherwise determined as non-central point, updates cluster centre point
Set；
S33: it for non-cluster central point, assigns them to the higher minimum range point of density and carries out reunion class.
5. a kind of exception monitoring scheme of stream calculation system based on unsupervised learning method according to claim 3,
It is characterized in that, step S25 includes:
S251: the index value of computing system significant condition, local density ρ, the minimum range δ apart from higher density point；
S252: building γ=ρ * δ, and γ is ranked up according to sequence from big to small, obtain order statistic X1≥…≥
Xn, wherein XiIndicate i-th of maximum γ value (1≤i≤n)；
S253: Η is enabled0,1,…,Η0,mFor X1,…,XmCorresponding null hypothesis, and assuming X not is extreme value abnormal point, i.e., is not cluster
Central point；
S254: it enables and assumes statistic Rt=γt/γt+1If the individual key value of t-th of null hypothesis is rt=[1- (1- α)1/m]-1/λt,
Wherein 1≤t≤m, α are significance, and λ is power-law distribution coefficient；
S255: Γ is enabledt=Rt-rt, work as Γt> 0, then Η0,tRefusal, otherwise receives；
S226: according to ΓmJudge null hypothesis Η0,mWhether receive, continues to judge Η if receiving0,m-1, until null hypothesis Η0,kIt refuses
(1≤k≤m) absolutely, that is to say, that Η0,kIt is first refusal it is assumed that then showing X1,…,XkIt is corresponding for extreme value abnormal point
γ is cluster centre point；
S257: distributing to minimum range point apart from higher density point for remaining point, i.e., with the most narrow spacing apart from higher density point
Similar from putting, the classification of most of data set is normal class in cluster, and abnormal class is corresponding in turn to abnormal 1, and exception 2 ... is abnormal
n。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910031830.3A CN109697332A (en) | 2019-01-14 | 2019-01-14 | A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910031830.3A CN109697332A (en) | 2019-01-14 | 2019-01-14 | A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109697332A true CN109697332A (en) | 2019-04-30 |
Family
ID=66233254
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910031830.3A Pending CN109697332A (en) | 2019-01-14 | 2019-01-14 | A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109697332A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111061711A (en) * | 2019-11-28 | 2020-04-24 | 同济大学 | Large data flow unloading method and device based on data processing behavior |
| CN111949700A (en) * | 2020-06-24 | 2020-11-17 | 浙江中控技术股份有限公司 | Intelligent safety guarantee real-time optimization method and system for petrochemical device |
| CN112070225A (en) * | 2020-09-01 | 2020-12-11 | 多点(深圳)数字科技有限公司 | Entity card abnormal binding alarm method based on unsupervised learning |
| CN116684202A (en) * | 2023-08-01 | 2023-09-01 | 光谷技术有限公司 | Internet of things information security transmission method |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090150136A1 (en) * | 2005-10-10 | 2009-06-11 | Sei Yang Yang | Dynamic-based verification apparatus for verification from electronic system level to gate level, and verification method using the same |
| CN102714560A (en) * | 2010-01-13 | 2012-10-03 | 松下电器产业株式会社 | Transmitter, transmission method, receiver, reception method, program, and integrated circuit |
| US20150199224A1 (en) * | 2014-01-10 | 2015-07-16 | Instep Software, Llc | Method and Apparatus for Detection of Anomalies in Integrated Parameter Systems |
| US20160183351A1 (en) * | 2013-03-25 | 2016-06-23 | Ids-Ip Holdings Llc | System, method, and apparatus for powering intelligent lighting networks |
| CN105912726A (en) * | 2016-05-13 | 2016-08-31 | 北京邮电大学 | Density centrality based sampling and detecting methods of unusual transaction data of virtual assets |
| CN106844161A (en) * | 2017-02-20 | 2017-06-13 | 重庆邮电大学 | Abnormal monitoring and Forecasting Methodology and system in a kind of carrier state stream calculation system |
| CN107729799A (en) * | 2017-06-13 | 2018-02-23 | 银江股份有限公司 | Crowd's abnormal behaviour vision-based detection and analyzing and alarming system based on depth convolutional neural networks |
| CN107831285A (en) * | 2017-01-19 | 2018-03-23 | 江苏省金威测绘服务中心 | A kind of dystrophication monitoring system and its method based on Internet of Things |
| CN107846472A (en) * | 2017-11-24 | 2018-03-27 | 华北电力大学（保定） | The fleet anomaly detection method of extensive power transmission and transforming equipment Monitoring data flow |
-
2019
- 2019-01-14 CN CN201910031830.3A patent/CN109697332A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090150136A1 (en) * | 2005-10-10 | 2009-06-11 | Sei Yang Yang | Dynamic-based verification apparatus for verification from electronic system level to gate level, and verification method using the same |
| CN102714560A (en) * | 2010-01-13 | 2012-10-03 | 松下电器产业株式会社 | Transmitter, transmission method, receiver, reception method, program, and integrated circuit |
| US20160183351A1 (en) * | 2013-03-25 | 2016-06-23 | Ids-Ip Holdings Llc | System, method, and apparatus for powering intelligent lighting networks |
| US20150199224A1 (en) * | 2014-01-10 | 2015-07-16 | Instep Software, Llc | Method and Apparatus for Detection of Anomalies in Integrated Parameter Systems |
| CN105912726A (en) * | 2016-05-13 | 2016-08-31 | 北京邮电大学 | Density centrality based sampling and detecting methods of unusual transaction data of virtual assets |
| CN107831285A (en) * | 2017-01-19 | 2018-03-23 | 江苏省金威测绘服务中心 | A kind of dystrophication monitoring system and its method based on Internet of Things |
| CN106844161A (en) * | 2017-02-20 | 2017-06-13 | 重庆邮电大学 | Abnormal monitoring and Forecasting Methodology and system in a kind of carrier state stream calculation system |
| CN107729799A (en) * | 2017-06-13 | 2018-02-23 | 银江股份有限公司 | Crowd's abnormal behaviour vision-based detection and analyzing and alarming system based on depth convolutional neural networks |
| CN107846472A (en) * | 2017-11-24 | 2018-03-27 | 华北电力大学（保定） | The fleet anomaly detection method of extensive power transmission and transforming equipment Monitoring data flow |
Non-Patent Citations (3)
| Title |
|---|
| VENISHA MARIA TELLIS 等: "Detecting Anomalies in Data Stream Using Efficient Techniques: A Review", 《2018 INTERNATIONAL CONFERENCE ON CONTROL, POWER, COMMUNICATION AND COMPUTING TECHNOLOGIES (ICCPCCT)》 * |
| 王德文 等: "智能电网大数据流式处理方法与状态监测异常监测", 《电力系统自动化》 * |
| 罗杰: "基于事件流计算系统异常监测研究", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111061711A (en) * | 2019-11-28 | 2020-04-24 | 同济大学 | Large data flow unloading method and device based on data processing behavior |
| CN111061711B (en) * | 2019-11-28 | 2023-09-01 | 同济大学 | Big data stream unloading method and device based on data processing behavior |
| CN111949700A (en) * | 2020-06-24 | 2020-11-17 | 浙江中控技术股份有限公司 | Intelligent safety guarantee real-time optimization method and system for petrochemical device |
| CN111949700B (en) * | 2020-06-24 | 2024-04-09 | 浙江中控技术股份有限公司 | Intelligent safety guarantee real-time optimization method and system for petrochemical device |
| CN112070225A (en) * | 2020-09-01 | 2020-12-11 | 多点(深圳)数字科技有限公司 | Entity card abnormal binding alarm method based on unsupervised learning |
| CN112070225B (en) * | 2020-09-01 | 2023-10-10 | 多点(深圳)数字科技有限公司 | Entity card abnormal binding alarm method based on unsupervised learning |
| CN116684202A (en) * | 2023-08-01 | 2023-09-01 | 光谷技术有限公司 | Internet of things information security transmission method |
| CN116684202B (en) * | 2023-08-01 | 2023-10-24 | 光谷技术有限公司 | Internet of things information security transmission method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109697332A (en) | A kind of exception monitoring scheme of the stream calculation system based on unsupervised learning method | |
| Zhou et al. | A novel transfer learning-based intelligent nonintrusive load-monitoring with limited measurements | |
| CN105117602B (en) | A kind of metering device running status method for early warning | |
| CN107274105B (en) | Linear discriminant analysis-based multi-attribute decision tree power grid stability margin evaluation method | |
| CN106204330A (en) | A kind of power distribution network intelligent diagnosis system | |
| CN110428005B (en) | Umbrella algorithm-based dynamic security misclassification constraint method for power system | |
| CN111625991A (en) | Low-voltage distribution network topology verification method | |
| CN105471647B (en) | A kind of power communication network fault positioning method | |
| CN104881735A (en) | System and method of smart power grid big data mining for supporting smart city operation management | |
| CN107153845A (en) | A kind of isolated island detection method of the adaptive grid-connected photovoltaic system based on machine learning | |
| Pietrucha-Urbanik | Multidimensional comparative analysis of water infrastructures differentiation | |
| CN116432123A (en) | Electric energy meter fault early warning method based on CART decision tree algorithm | |
| CN111652478A (en) | Electric power system voltage stability evaluation misclassification constraint method based on umbrella algorithm | |
| CN113740666B (en) | Method for positioning root fault of storm alarm in power system of data center | |
| CN110348480A (en) | A kind of non-supervisory anomaly data detection algorithm | |
| CN103218664A (en) | Warning weight determination method based on wavelet neural network | |
| CN117787698A (en) | Micro-grid risk assessment method and system based on power supply range maximization | |
| CN117034149A (en) | Fault processing strategy determining method and device, electronic equipment and storage medium | |
| CN111965442A (en) | Energy internet fault diagnosis method and device under digital twin environment | |
| CN111404742A (en) | Industry interconnection framework towards wisdom supply chain | |
| Xu et al. | Federated traffic synthesizing and classification using generative adversarial networks | |
| CN108258802A (en) | The monitoring method and device of the operation conditions of controller switching equipment in a kind of power distribution network | |
| CN113569961B (en) | Power grid node classification method and computer readable medium | |
| CN115598459A (en) | Power failure prediction method for 10kV feeder line fault of power distribution network | |
| Shan et al. | Root Cause Analysis of Failures for Power Communication Network Based on CNN |
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: 20190430 |