CN110399910A - Fire abnormal point online test method based on sliding window and HWKS theoretical frame - Google Patents
Fire abnormal point online test method based on sliding window and HWKS theoretical frame Download PDFInfo
- Publication number
- CN110399910A CN110399910A CN201910610771.5A CN201910610771A CN110399910A CN 110399910 A CN110399910 A CN 110399910A CN 201910610771 A CN201910610771 A CN 201910610771A CN 110399910 A CN110399910 A CN 110399910A
- Authority
- CN
- China
- Prior art keywords
- binary tree
- fire
- sliding window
- data
- difference
- 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
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 25
- 238000010998 test method Methods 0.000 title claims abstract description 12
- 238000009826 distribution Methods 0.000 claims abstract description 16
- 238000012544 monitoring process Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 18
- 238000000034 method Methods 0.000 claims description 9
- 241001269238 Data Species 0.000 claims description 3
- 230000008520 organization Effects 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 12
- 230000006378 damage Effects 0.000 description 3
- 208000027418 Wounds and injury Diseases 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 208000014674 injury Diseases 0.000 description 2
- 238000013450 outlier detection Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 239000000443 aerosol Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 210000004725 window cell Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Alarm Systems (AREA)
- Fire Alarms (AREA)
Abstract
The fire abnormal point online test method based on sliding window and HWKS theoretical frame that the present invention relates to a kind of, comprising the following steps: obtain fire hazard environment and monitor time series data, establish data receiver buffer area；The data to be tested in data receiver buffer area are divided into several subsegments using sliding window theory, determine the distribution of fire abnormal point；It is realized using HWKS algorithm and mean value binary tree and difference binary tree is constructed to sliding window internal data field, and realize the fast search of fire monitoring data exception feature by introducing binary search strategy.The present invention realizes the quick detection of abnormal point in fire detection.
Description
Technical field
The present invention relates to fire outlier detection technical fields, theoretical based on sliding window and HWKS more particularly to one kind
The fire abnormal point online test method of frame.
Background technique
Important accident one of of the fire as human lives brings huge life threat to people's lives and property damages
It loses.The generation of fire each time all can be along with huge property loss, or even causes casualties.Why fire will cause
The intensity of a fire cannot be controlled effectively after such serious consequence, mainly fire occur, scene of fire personnel cannot in time from
It opens.How fire detection is accurately and effectively carried out, is the major issue that we study.Detection mainly allows people early
Point discovery fire makes reply in time, it would be possible to which the injury brought is preferably minimized.But the wrong report of fire equally can
Bring many unnecessary losses, such as caused by the dust etc. of non-fire hazard aerosol fog, the oil smoke in kitchen and non-fire
Wrong report.It also can may carry out unnecessary injury to the life zone of people during commercial building evacuation on fire, or step on
Step on event.There are also the change that the fire alarm in transport hub may cause suspension of service and traffic scheduling scheme, fire
Calamity wrong report may bring very big inconvenience to the trip of people.So the accurate property of fire alarm is urgently to be resolved at present
A problem；At the same time, the real-time and rapidity of fire alarm be at present with more realistic meaning, there is an urgent need to solve
A problem certainly.
Summary of the invention
Technical problem to be solved by the invention is to provide a kind of fire based on sliding window and HWKS theoretical frame is different
Often point online test method, realizes the quick detection of abnormal point in fire detection.
The technical solution adopted by the present invention to solve the technical problems is: providing a kind of theoretical based on sliding window and HWKS
The fire abnormal point online test method of frame, comprising the following steps:
(1) it obtains fire hazard environment and monitors time series data, establish data receiver buffer area；
(2) data to be tested in data receiver buffer area are divided into several subsegments using sliding window theory, really
Determine the distribution of fire abnormal point；
(3) it is realized using HWKS algorithm and mean value binary tree and difference binary tree is constructed to sliding window internal data field, and led to
It crosses and introduces the fast search that binary search strategy realizes fire monitoring data exception feature.
Data receiver buffer area uses queue storage organization in the step (1).
The step (2) specifically: the fire time series data Z for being N for length sets the width of sliding window as W, makes
The number for obtaining data in each window is W, then it is W's that fire time series data Z is divided into several width by sliding window model
In window.
Multistage Haar small wave converting method is utilized in the step (3), to each subsegment determined by sliding window when
Sequence sequence Z={ z1,z2,...,znDifferent frequency domain components are parsed into, it indicates are as follows:Wherein, cA and cD is respectively Mean Parameters and difference parameter, by time series
Z resolves into the Mean Parameters matrix and difference parameter matrix of multidimensional, and Mean Parameters matrix and difference parameter matrix are mapped to pair
Each level of child nodes of the mean value binary tree and difference binary tree that answer is constructed by Mean Parameters matrix and difference parameter matrix
It is worth the non-leaf nodes of binary tree and difference binary tree；Using improved KS statistical theory creation mean value binary tree search strategy and
Difference binary tree search strategy, in conjunction with mean value binary tree search strategy and difference binary tree search strategy, most from mean value binary tree
The root of upper end starts, and preferentially catastrophe point path is successively found using mean value binary tree search strategy, if there is mean value binary tree
The case where search strategy fails, then just finding the node in searching route at this time on corresponding difference binary tree, then makes again
With difference binary tree search decision search path, until leaf node.
The mean value binary tree search strategy specifically: the non-leaf nodes in assumed average binary tree is cAk,j, it
The child node on the left side and the child node on the right side are cA respectivelyk-1,2j-1And cAk-1,2j；
If D'mn(k-1,2j-1) < D'mn(k-1,2j) and D'mn(k-1,2j) > C (α), then select the child node on the left side
cAk-1,2jAs searching route current in mean value binary tree；
If D'mn(k-1,2j-1) > D'mn(k-1,2j) and D'mn(k-1,2j) > C (α), then select the child node on the right side
cAk-1,2jAs searching route current in mean value binary tree；Wherein, statistic D'mn(k, j) indicates node cAk.jIt is corresponding
Time series data in two segment datas experience distribution difference size；C (α) is the estimates of parameters that the level of signifiance is α.
The difference binary tree search strategy specifically: for the n omicronn-leaf child node cD on difference binary treek,j, left and right
Child's node is respectively cDk-1,2j-1And cDk-1,2j；
(1) if | cDk-1,2j-1| > | cDk-1,2j|, then selecting the left child of node is searching route；
(2) if | cDk-1,2j-1| < | cDk-1,2j|, then selecting the right child of node is searching route.
Beneficial effect
Due to the adoption of the above technical solution, compared with prior art, the present invention having the following advantages that and actively imitating
Fruit: the present invention can obtain the time series data of fire hazard environment monitoring by establishing data buffer zone in real time.It is managed in conjunction with sliding window
By being several subsegments by fire time series data cutting, on-line checking is carried out to each subsegment respectively, can quickly determine that fire is different
The distribution of regular data.Compared to traditional abnormal point detecting method, the present invention combine multistage Harr wavelet transformation and improvement
KS statistical theory constructs mean value binary tree, difference binary tree and two kinds of binary tree search strategies, can quickly determine optimal search
Pathfinding line finds the time series data where abnormal point.
Detailed description of the invention
Fig. 1 is the fire outlier detection process schematic of the invention based on sliding window theory and HWKS frame；
Fig. 2 is the sliding window model A based on element number and the sliding window model B schematic diagram based on time interval；
Fig. 3 is the sliding window model schematic diagram in the present invention；
Fig. 4 is of the invention based on multistage Harr wavelet transformation building mean value binary tree and the signal of difference binary tree process
Figure.
Specific embodiment
Present invention will be further explained below with reference to specific examples.It should be understood that these embodiments are merely to illustrate the present invention
Rather than it limits the scope of the invention.In addition, it should also be understood that, after reading the content taught by the present invention, those skilled in the art
Member can make various changes or modifications the present invention, and such equivalent forms equally fall within the application the appended claims and limited
Range.
Embodiments of the present invention are related to a kind of fire abnormal point based on sliding window and HWKS theoretical frame and examine online
Survey method, as shown in Figure 1, comprising the following steps:
(1) it obtains fire hazard environment and monitors time series data, establish data receiver buffer area.Wherein, data receiver buffer area is adopted
With queue structure, newly generated fire time series data will successively fall in lines from tail of the queue sequentially in time, when data buffering complete, into
When entering data handling procedure, buffered data is successively fallen out from head of the queue in chronological order again.
(2) data to be tested in data receiver buffer area are divided into several subsegments using sliding window theory, really
Determine the distribution of fire abnormal point.
Sliding window model is that the time of window or the quantity of data element is enabled to keep constant, by dividing data flow
For several subsegments, subsequent one subsegment of another subsegment carries out Singularity detection.Sliding window cell list shows in data flow
The width in some subinterval, this section is determined according to time or space, is divided into two classes, and window elements number is kept constant
Be known as the sliding window based on element number, widow time interval remain unchanged be known as the sliding window based on time interval
Mouthful, as shown in Figure 2.It is N discrete data sequences Z to be detected for length, sets the size of window as W, then just by sequence
Column Z is divided into m=N/W subsequence, then for this m subsequence, carries out catastrophe point respectively with Singularity detection method
Detection.
The appearance of sliding window provides thinking for the on-line checking of fire abnormal point.Pass through the sequential data stream that will be cached
Segmentation, is divided into several segments for time series data, then detect respectively to each subsegment.Spy when being occurred according to practical fire
Sign, such as temperature is got higher, smokescope is got higher, CO and CO2 concentration is got higher, and determines the distribution of fire abnormal point.This implementation
The sliding window model of mode sets the width of sliding window as W as shown in figure 3, the fire time series data Z for being N for length,
The number of data is W in i.e. each window, then time series data Z is divided into several wide child windows by sliding window model
It is interior, quickly determine the distribution of fire abnormal point.
(3) it is realized using HWKS algorithm and mean value binary tree and difference binary tree is constructed to sliding window internal data field, and led to
It crosses and introduces the fast search that binary search strategy realizes fire monitoring data exception feature.
Using multistage Haar small wave converting method, to the time series Z={ z of each subsegment determined by sliding window1,
z2,...,znDifferent frequency domain components are parsed into, it indicates are as follows:Wherein, cA and
CD is respectively Mean Parameters and difference parameter, and time series Z is resolved into the Mean Parameters matrix of multidimensionalWith difference parameter matrixWherein, 0≤k≤
M=log2N,1≤j≤N/2k, Mean Parameters matrix and difference parameter matrix are mapped to corresponding mean value binary tree and difference two
Each level of child nodes for pitching tree constructs mean value binary tree and difference binary tree by Mean Parameters matrix and difference parameter matrix
Non-leaf nodes realizes building (see Fig. 4) of mean value binary tree and difference binary tree
Mean value binary tree search strategy and difference binary tree search strategy are created using improved KS statistical theory, in conjunction with equal
It is worth binary tree search strategy and difference binary tree search strategy, since the root of mean value binary tree the top, preferentially uses mean value
Binary tree search strategy successively finds catastrophe point path, the case where if there is mean value binary tree search strategy fails, then just
The node in searching route at this time on corresponding difference binary tree is found, difference binary tree search decision search road is then reused
Diameter, until leaf node.
Wherein, the mean value binary tree search strategy specifically: it is assumed that the non-leaf nodes in TcA is cAk,j, its left side
The child node in face and the child node on the right side are cA respectivelyk-1,2j-1And cAk-1,2j。
If (a) D'mn(k-1,2j-1) < D'mn(k-1,2j) and D'mn(k-1,2j) > C (α) then selects the son section on the left side
Point cAk-1,2jAs searching route current in TcA.
If (b) D'mn(k-1,2j-1) > D'mn(k-1,2j) and D'mn(k-1,2j) > C (α) then selects the son section on the right side
Point cAk-1,2jAs searching route current in TcA.Wherein, statistic D'mn(k, j) indicates node cAk.jCorresponding timing
The difference size of the experience distribution of two segment datas in data.
As (D 'mn(k-1,2j-1) < D 'mnWhen (k-1,2j), illustrate node cAk,jThe corresponding time series data of left subtree in
Distribution distance it is bigger, in other words mean that catastrophe point more likely appears in this partial data.Because being mutated
When point, the time series data after catastrophe point can bring different data distributions, and generate a distribution distance.And if data do not have
It mutates, then it accumulates experience, distribution is just very close to distribution distance is naturally also with regard to very little.If D ' at this timemn(k-1,2j)
> C (α)) also set up simultaneously, then just illustrating that this distance has had reached the decision threshold of setting, i.e. the node corresponding data
In there are catastrophe points.If there are catastrophe points in the corresponding time series data of current node, according to search rule one
By distribution distance in its left and right subtree it is biggish that elect, and select the node as catastrophe point path, then give up another
One paths.
The difference binary tree search strategy specifically: work as D'mn(k-1,2j-1)==D'mn(k-1,2j) or max { D'mn
(k-1,2j-1),D'mn(k-1,2j) } < C (α) when, it may appear that strategy one fail the case where.It is, therefore, desirable to provide search strategy
Two supplements as strategy one.Its thinking is to compare data using the details coefficients of wavelet decomposition based on difference binary tree TcD
Fluctuation situation, the fluctuation of data is more severe, then the absolute value of details coefficients is bigger.It is therefore contemplated that data catastrophe point is more
It possibly is present at the biggish part of fluctuation, and in this, as the standard in selection path, finds data catastrophe point.
For the n omicronn-leaf child node cD on difference binary treek,j, left and right child's node is respectively cDk-1,2j-1And cDk-1,2j。
(1) if | cDk-1,2j-1| > | cDk-1,2j|, then selecting the left child of node is searching route.
(2) if | cDk-1,2j-1| < | cDk-1,2j|, then selecting the right child of node is searching route.
Finally, being realized with visual pattern to fire time series data by carrying out emulation experiment to the sample data of simulation
The real-time analysis of abnormal point and on-line monitoring.Pass through the emulation experiment of analogue data, it was demonstrated that present embodiment and other algorithm phases
Than the HWKS algorithm for introducing sliding window is successfully realized the quick detection of fire catastrophe point, and the used time is shorter, relative error rate
It is minimum, hit rate highest.
It is not difficult to find that present invention introduces slide window implementations to the on-line checking of time series data catastrophe point, i.e. real-time
Detection, the used time is short, and accuracy is high；Mean value binary tree (TcA) and difference binary tree (TcD) are constructed using HWKS algorithm, is created
Two kinds of binary tree search strategies, algorithm is simple and effective, can fast and accurately find optimal path, realize in fire detection
The quick detection of abnormal point.
Claims (6)
1. a kind of fire abnormal point online test method based on sliding window and HWKS theoretical frame, which is characterized in that including
Following steps:
(1) it obtains fire hazard environment and monitors time series data, establish data receiver buffer area；
(2) data to be tested in data receiver buffer area are divided into several subsegments using sliding window theory, determine fire
The distribution of calamity abnormal point；
(3) it is realized using HWKS algorithm and mean value binary tree and difference binary tree is constructed to sliding window internal data field, and by drawing
Enter the fast search that binary search strategy realizes fire monitoring data exception feature.
2. the fire abnormal point online test method according to claim 1 based on sliding window and HWKS theoretical frame,
It is characterized in that, data receiver buffer area uses queue storage organization in the step (1).
3. the fire abnormal point online test method according to claim 1 based on sliding window and HWKS theoretical frame,
It is characterized in that, the step (2) specifically: the fire time series data Z for being N for length, set the width of sliding window as
W, so that the number of data is W in each window, then fire time series data Z is divided into several width and is by sliding window model
In the window of W.
4. the fire abnormal point online test method according to claim 1 based on sliding window and HWKS theoretical frame,
It is characterized in that, multistage Haar small wave converting method is utilized in the step (3), to each subsegment determined by sliding window
Time series Z={ z1,z2,...,znDifferent frequency domain components are parsed into, it indicates are as follows:Wherein, cA and cD is respectively Mean Parameters and difference parameter, by time series
Z resolves into the Mean Parameters matrix and difference parameter matrix of multidimensional, and Mean Parameters matrix and difference parameter matrix are mapped to pair
Each level of child nodes of the mean value binary tree and difference binary tree that answer is constructed by Mean Parameters matrix and difference parameter matrix
It is worth the non-leaf nodes of binary tree and difference binary tree；Using improved KS statistical theory creation mean value binary tree search strategy and
Difference binary tree search strategy, in conjunction with mean value binary tree search strategy and difference binary tree search strategy, most from mean value binary tree
The root of upper end starts, and preferentially catastrophe point path is successively found using mean value binary tree search strategy, if there is mean value binary tree
The case where search strategy fails, then just finding the node in searching route at this time on corresponding difference binary tree, then makes again
With difference binary tree search decision search path, until leaf node.
5. the fire abnormal point online test method according to claim 4 based on sliding window and HWKS theoretical frame,
It is characterized in that, the mean value binary tree search strategy specifically: the non-leaf nodes in assumed average binary tree is cAk,j, it
The child node on the left side and the child node on the right side be cA respectivelyk-1,2j-1And cAk-1,2j；
If (a) D'mn(k-1,2j-1) < D'mn(k-1,2j) and D'mn(k-1,2j) > C (α), then select the child node on the left side
cAk-1,2jAs searching route current in mean value binary tree；
If (b) D'mn(k-1,2j-1) > D'mn(k-1,2j) and D'mn(k-1,2j) > C (α), then select the child node on the right side
cAk-1,2jAs searching route current in mean value binary tree；Wherein, statistic D'mn(k, j) indicates node cAk.jIt is corresponding
Time series data in two segment datas experience distribution difference size；C (α) is the estimates of parameters that the level of signifiance is α.
6. the fire abnormal point online test method according to claim 4 based on sliding window and HWKS theoretical frame,
It is characterized in that, the difference binary tree search strategy specifically: for the n omicronn-leaf child node cD on difference binary treek,j, left
Right child's node is respectively cDk-1,2j-1And cDk-1,2j；
(1) if | cDk-1,2j-1| > | cDk-1,2j|, then selecting the left child of node is searching route；
(2) if | cDk-1,2j-1| < | cDk-1,2j|, then selecting the right child of node is searching route.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910610771.5A CN110399910A (en) | 2019-07-08 | 2019-07-08 | Fire abnormal point online test method based on sliding window and HWKS theoretical frame |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910610771.5A CN110399910A (en) | 2019-07-08 | 2019-07-08 | Fire abnormal point online test method based on sliding window and HWKS theoretical frame |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110399910A true CN110399910A (en) | 2019-11-01 |
Family
ID=68322826
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910610771.5A Pending CN110399910A (en) | 2019-07-08 | 2019-07-08 | Fire abnormal point online test method based on sliding window and HWKS theoretical frame |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110399910A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111488496A (en) * | 2020-04-30 | 2020-08-04 | 湖北师范大学 | Sliding window based Tango tree construction method and system |
| CN114912852A (en) * | 2022-07-18 | 2022-08-16 | 山东鑫亚装备制造有限公司 | Abnormity monitoring and early warning system for storage tank of refrigeration bin |
| CN116304957A (en) * | 2023-05-17 | 2023-06-23 | 成都交大光芒科技股份有限公司 | On-line identification method for monitoring state mutation of power supply and transformation equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106485211A (en) * | 2016-09-26 | 2017-03-08 | 西安交通大学 | A kind of line of text accurate positioning method based on binary tree |
| CN108335467A (en) * | 2018-01-05 | 2018-07-27 | 东华大学 | A kind of fire on-line early warning and rapid analysis method based on catastrophe point detection |
| CN108921202A (en) * | 2018-06-12 | 2018-11-30 | 成都信息工程大学 | A kind of abnormal point detecting method based on data structure |
-
2019
- 2019-07-08 CN CN201910610771.5A patent/CN110399910A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106485211A (en) * | 2016-09-26 | 2017-03-08 | 西安交通大学 | A kind of line of text accurate positioning method based on binary tree |
| CN108335467A (en) * | 2018-01-05 | 2018-07-27 | 东华大学 | A kind of fire on-line early warning and rapid analysis method based on catastrophe point detection |
| CN108921202A (en) * | 2018-06-12 | 2018-11-30 | 成都信息工程大学 | A kind of abnormal point detecting method based on data structure |
Non-Patent Citations (3)
| Title |
|---|
| 宋巧红: "一种数据流突变点快速探测算法的研究与实现", 《中国优秀硕士学位论文全文数据库（信息科技辑）》 * |
| 宋巧红等: "基于多级Haar小波变换与KS统计的突变点快速探测方法", 《计算机工程》 * |
| 戴慧等: "多变量时间序列滑动窗口异常点的检测", 《南京信息工程大学学报(自然科学版)》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111488496A (en) * | 2020-04-30 | 2020-08-04 | 湖北师范大学 | Sliding window based Tango tree construction method and system |
| CN114912852A (en) * | 2022-07-18 | 2022-08-16 | 山东鑫亚装备制造有限公司 | Abnormity monitoring and early warning system for storage tank of refrigeration bin |
| CN114912852B (en) * | 2022-07-18 | 2022-09-23 | 山东鑫亚装备制造有限公司 | Abnormity monitoring and early warning system for storage tank of refrigeration bin |
| CN116304957A (en) * | 2023-05-17 | 2023-06-23 | 成都交大光芒科技股份有限公司 | On-line identification method for monitoring state mutation of power supply and transformation equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110399910A (en) | Fire abnormal point online test method based on sliding window and HWKS theoretical frame | |
| US10966113B2 (en) | Multisensory change detection for internet of things domain | |
| CN112398779B (en) | Network traffic data analysis method and system | |
| CN105306475B (en) | A kind of network inbreak detection method based on Classification of Association Rules | |
| CN109714343B (en) | Method and device for judging network traffic abnormity | |
| CN105824798A (en) | Examination question de-duplicating method of examination question base based on examination question key word likeness | |
| CN105336342B (en) | Speech recognition result evaluation method and system | |
| CN108595655B (en) | Abnormal user detection method based on session feature similarity fuzzy clustering | |
| CN114124482B (en) | Access flow anomaly detection method and equipment based on LOF and isolated forest | |
| CN110087207B (en) | Method for reconstructing missing data of wireless sensor network | |
| CN111008337B (en) | Deep attention rumor identification method and device based on ternary characteristics | |
| CN110019074A (en) | Analysis method, device, equipment and the medium of access path | |
| CN107943704A (en) | A kind of Software Defects Predict Methods based on depth forest network | |
| CN108335467B (en) | A kind of fire on-line early warning and rapid analysis method based on catastrophe point detection | |
| CN110012009B (en) | Internet of things intrusion detection method based on combination of decision tree and self-similarity model | |
| CN111600876B (en) | Slow denial of service attack detection method based on MFOPA algorithm | |
| CN110475246B (en) | Malicious anchor node detection method based on isolated forest and sequential probability ratio detection | |
| CN116938683B (en) | Network path analysis system and method based on network security anomaly detection | |
| CN112199670A (en) | Log monitoring method for improving IFOREST (entry face detection sequence) to conduct abnormity detection based on deep learning | |
| KR102407834B1 (en) | Method and apparatus for property-based classification of long-pulse radar signals | |
| Akram et al. | Automatic event-detection and time-picking algorithms for downhole microseismic data processing | |
| Zuech et al. | Feature popularity between different web attacks with supervised feature selection rankers | |
| Milner et al. | An analysis of packet loss models for distributed speech recognition. | |
| CN106209868A (en) | A kind of large-scale network traffic exception detecting method and system | |
| Ramos et al. | Anomalies detection in wireless sensor networks using bayesian changepoints |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191101 |