CN111427752A - Regional anomaly monitoring method based on edge calculation - Google Patents
Regional anomaly monitoring method based on edge calculation Download PDFInfo
- Publication number
- CN111427752A CN111427752A CN202010515822.9A CN202010515822A CN111427752A CN 111427752 A CN111427752 A CN 111427752A CN 202010515822 A CN202010515822 A CN 202010515822A CN 111427752 A CN111427752 A CN 111427752A
- Authority
- CN
- China
- Prior art keywords
- data
- points
- distance
- abnormal
- edge
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3065—Monitoring arrangements determined by the means or processing involved in reporting the monitored data
- G06F11/3072—Monitoring arrangements determined by the means or processing involved in reporting the monitored data where the reporting involves data filtering, e.g. pattern matching, time or event triggered, adaptive or policy-based reporting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- 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
-
- 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
- H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Evolutionary Computation (AREA)
- Computer Networks & Wireless Communication (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a regional anomaly monitoring method and a regional anomaly monitoring device based on edge calculation, and relates to the technical field of regional anomaly monitoring; the monitoring method comprises the following steps: the method comprises the steps that monitoring characteristic data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device; presetting a used processing model, and inputting the collected characteristic data into the model; performing edge calculation; identity authentication; and the computing equipment comprises a data acquisition module, a computing processing module, an identity verification module and a result output module. The method mainly adopts a distance-based edge calculation method and a density-based edge calculation method to calculate and extract abnormal point data in the data, the distance-based edge calculation method relies on a multi-dimensional index structure, an integral set is divided into a plurality of dimensions, abnormal point data in each dimension are searched, and finally summarization is carried out, so that the accuracy and comprehensiveness of analysis and extraction of the abnormal point data are improved.
Description
Technical Field
The invention relates to the technical field of regional anomaly monitoring, in particular to a regional anomaly monitoring method based on edge calculation.
Background
With the development of the IOT (Internet of things), the equipment is accessed to the Internet of things in a large scale, massive data generated on the terminal equipment provides commercial value, and meanwhile, a new challenge is provided for big data wind control, namely, the abnormal data is monitored while the privacy is ensured.
Through retrieval, chinese publication No. CN109947079A discloses an edge calculation-based regional anomaly detection method and an edge, which includes: the method comprises the steps that monitoring feature data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device, wherein the plurality of Internet of things devices belong to the same designated area; and taking the monitoring characteristic data of the plurality of Internet of things devices as input of an anomaly detection random forest model to predict whether the specified area is abnormal, wherein the anomaly detection random forest model comprises a plurality of random forest decision trees which are respectively trained based on the monitoring characteristic data of the plurality of Internet of things devices.
The above patent uses the whole data set to carry out modeling analysis, and because the quantity of monitoring data is more, and whole database is great, directly utilizes set modeling analysis, may lead to the omission of data anomaly to can't guarantee the accuracy and the comprehensiveness of data.
Disclosure of Invention
The invention aims to solve the defects in the prior art, and provides a regional abnormity monitoring method based on edge calculation.
In order to achieve the purpose, the invention adopts the following technical scheme:
a regional anomaly monitoring method based on edge calculation comprises the following steps:
s1: the method comprises the steps that monitoring characteristic data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device;
s2: presetting a used processing model, and inputting the collected monitoring characteristic data into the model;
s3: performing edge calculation and analyzing abnormal data;
s4: identity authentication, which is to authenticate identity information of an operator; after the verification is successful, the abnormal data information can be read through the edge computing device.
Preferably: in S3, the edge calculation method includes an edge calculation method based on the direct distance of the feature data and an edge calculation method based on the direct density of the feature data.
Preferably: the edge calculation method based on the feature data direct distance comprises the following steps:
s11, inputting all the monitoring characteristic data into the data set S, and inputting a certain data point in the data set SIs called as,Representing the horizontal and vertical coordinate values of the data points;
s12: if the data pointSatisfying specific properties, using the abscissa and ordinate values of abnormal data points to replace,All statistically based outlier data points are represented.
Preferably: what is needed isIn the above-mentioned S12, the specific properties are: data setHas data points and data pointsIs greater thanThe algorithm is as follows: index-based methods rely on multidimensional indexing structures by querying nearest neighbors or data pointsThe complexity of the R-trees or kd-trees algorithm based on mostly index structures for finding all data points is realized by using answers of range queries as centersWhereinIs the dimension of the dimension number of the object,is the number of points of the data,representing a data point.
Preferably: the edge calculation method based on the feature data direct density comprises the following steps:
s21: of object PDistance is calledFor any natural numberDefinition of PDistance andis P and some objectDistance between, here data pointThe preset condition is required to be met;
s22: of object PDistance neighborhood calledOf a given PA distance ofOf PThe distance neighborhood contains all the distances to PNot exceedingObject of (2)I.e. by;
S23: the reachable distance of the object P with respect to O, given the natural number k,the reachable distance of the object P relative to the object O isThe local reachable density of an object P is the inverse of the average reachable distance of the object P from his MinPts-neighborhood, i.e. the;
s25, the local outlier factor OF the object P represents the degree OF the outlier OF P, the larger the local outlier factor is, the more likely it is to be the outlier, otherwise the probability is low, the object L OF close to the core store in the cluster is close to 1 and is not considered to be the local outlier, and the object L OF at the edge OF the cluster or outside the cluster is relatively large.
Preferably: the edge calculation method in the step S3 is anomaly detection of high-dimensional data, and includes the steps of:
s31: dividing each dimension of the data space intoEach equal depth interval; the equal depth intervals are spatially defined by the data being collected, and each interval includes equal valuesThe data points of (a);
s32: each dimension in a k-dimensional subspace of the datasetEach equal-depth interval is taken to form a k-dimensional cube, and the number of data mapping points in the cube is a random number,;
Wherein, thereinRepresentsDimension cubeThe number of points included in the pool,the number of the total points is represented,representing the sparse coefficients of the image data to be processed,when negative, the cube is describedWhere the data point is below the expected value,the smaller the cubeThe more sparse the data in (1).
Preferably: detecting abnormal points of the high-dimensional data; the abnormal point detection of high-dimensional data optimizes the problem solving process by using a genetic algorithm, wherein the solving process starts with the selected P legal modes as candidate solutions and repeatedly experiencesSelecting, crossing, mutating, etc. until obtaining more satisfactory solution, analyzing abnormal data also includes abnormal point detection algorithm based on clustering, in the detection process of abnormal points, firstly removing abnormal points from clustering analysis to make detection, if the number of removed points is less than that of actual abnormal points to be detected, then detecting residual abnormal points from each cluster, after clustering analysis of data set S, dividing data object S into n clusters, marking as,and get rid ofThe number of outliers, noted,whereinThen, thenEach cluster center is marked asEach abnormal pointThe distance to the cluster center is noted as:wherein, in the step (A),represents the dimensions of the data and is,represents an integer whenWhen is in use, theSolving the distance by using a Manhattan distance formula method; when in useWhen is in use, theSolving the distance by an Euclidean distance formula method, wherein the value of q is determined according to specific conditions; any object rejectedTo each cluster coreIs a distance ofCan be represented by a matrix R, in which the ith row elementIndicating the distance to each cluster centerLet us order,Is the sum of the ith row in the matrix R and represents the distance sum of the ith culled point to each cluster center,the larger the value, the farther it is from each centroid, whenWhen the temperature of the water is higher than the set temperature,before the maximum valueThe removed object is the abnormal point whenWhen it is, thisThe removed points are abnormal points, and the rest points are detected from each clusterAn abnormal point, set clusterAny of the data objectsTo the cluster core where it is locatedA distance ofSelecting from each clusterBefore the maximum valueA point of the object being located at the center, thusCommon elimination in clustersPoints because ofTherefore, it isThen from thisDetection in pointsAn anomaly point.
Preferably: the edge calculation method in S3 may be calculated by the following formula:
wherein the content of the first and second substances,in order to transfer the function values,the amount of the carbon dioxide is the intermediate amount,in order to be a function of the edges,as a function of the diffusion coefficient of the points in the region,in the case of the edge bevel angle,is the angle of inclination of the midpoint of the region;
for in the collected characteristic dataPoint, if the transfer function value obtainedInconsistent with the collected data, it is called an outlier, otherwise it is a normal point.
The utility model provides an edge computing device, includes data acquisition module, calculation processing module, authentication module and result output module, the monitoring characteristic data that a plurality of thing networking devices that data acquisition module was used for accessing edge computing device gathered, calculation processing module can accept the monitoring characteristic data that the acquisition module obtained to according to belong to and predetermine the algorithm calculation and the extraction of unusual dot data, authentication module is used for verifying operating personnel's identity, if verify successful, allows to read unusual data result, if verify failure, refuse to read, result output module can show the unusual dot data after the calculation was extracted.
The invention has the beneficial effects that:
1. the invention mainly adopts a distance-based edge calculation method and a density-based edge calculation method to calculate and extract abnormal point data in the data, the distance-based edge calculation method relies on a multi-dimensional index structure based on an index method, and the nearest neighbor is inquired or the data points are used for searchingFinding all of them is accomplished by the answer of the range-query for the centerThe accuracy and comprehensiveness of analyzing and extracting the abnormal point data are improved by dividing a whole set into a plurality of dimensions, searching the abnormal point data in each dimension, and finally summarizing.
2. The invention optimizes the calculation amount of single solving by adopting the abnormal point detection means of high-dimensional data and utilizing the genetic algorithm to optimize the problem solving process, starts with P selected legal modes as candidate solutions, and repeatedly goes through several processes of selection, intersection, variation and the like until a more satisfactory solution is obtained, thereby avoiding the condition that the system load is increased due to overlarge single calculation amount of the calculation equipment while ensuring the precision.
3. According to the method, the rejected points are selected according to the distance from the cluster center, and the abnormal points are judged according to the distance from the rejected points to the cluster center.
Drawings
Fig. 1 is a process flow diagram of a regional anomaly monitoring method based on edge calculation according to the present invention.
Detailed Description
The technical solution of the present patent will be described in further detail with reference to the following embodiments.
In the description of this patent, it is noted that unless otherwise specifically stated or limited, the terms "mounted," "connected," and "disposed" are to be construed broadly and can include, for example, fixedly connected, disposed, detachably connected, disposed, or integrally connected and disposed. The specific meaning of the above terms in this patent may be understood by those of ordinary skill in the art as appropriate.
Example 1:
a regional anomaly monitoring method based on edge calculation comprises the following steps:
s1: the method comprises the steps that monitoring characteristic data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device;
s2: presetting a used processing model, and inputting the collected monitoring characteristic data into the model;
s3: performing edge calculation and analyzing abnormal data;
in S3, the edge calculation method includes an edge calculation method based on the direct distance of the feature data and an edge calculation method based on the direct density of the feature data.
The edge calculation method based on the feature data direct distance comprises the following steps:
s11, inputting all the monitoring characteristic data into the data set S, and inputting a certain data point in the data set SIs called as,Representing the horizontal and vertical coordinate values of the data points;
s12: if the data pointSatisfying specific properties, using the abscissa and ordinate values of abnormal data points to replace,All statistically based outlier data points are represented.
In S12, the specific properties are: data setHas data points and data pointsIs greater thanThe algorithm is as follows: index-based methods rely on multidimensional indexing structures by querying nearest neighbors or data pointsThe complexity of the R-trees or kd-trees algorithm based on mostly index structures for finding all data points is realized by using answers of range queries as centersWhereinIs the dimension of the dimension number of the object,is the number of points of the data,representing a data point.
The edge calculation method based on the feature data direct density comprises the following steps:
s21: of object PDistance is calledFor any natural numberDefinition of PDistance andis P and some objectDistance between, here data pointThe preset condition is required to be met;
s22: of object PDistance neighborhood calledOf a given PA distance ofOf PThe distance neighborhood contains all the distances to PNot exceedingObject of (2)I.e. by;
S23: the reachable distance of the object P relative to the object O is given by a natural number kThe local reachable density of an object P is the inverse of the average reachable distance of the object P from his MinPts-neighborhood, i.e. the;
s25, the local outlier factor OF the object P represents the degree OF the outlier OF P, the larger the local outlier factor is, the more likely it is to be the outlier, otherwise the probability is low, the object L OF close to the core store in the cluster is close to 1 and is not considered to be the local outlier, and the object L OF at the edge OF the cluster or outside the cluster is relatively large.
The utility model provides an edge computing device, includes data acquisition module, calculation processing module, authentication module and result output module, the monitoring characteristic data that a plurality of thing networking equipment collection that the data acquisition module was used for accessing edge computing device, calculation processing module can accept the data that the acquisition module obtained to according to belong to and predetermine the algorithm calculation and the extraction of abnormal point data, authentication module is used for verifying operating personnel's identity, if verify successful, then allow to read abnormal data result, if verify failure, refuse to read, result output module can show the abnormal point data after the calculation extraction.
In the embodiment, the distance-based edge calculation method and the density-based edge calculation method are mainly adopted to calculate the abnormal point data in the extracted data, and the distance-based edge calculation method relies on a multi-dimensional index structure through nearest neighbor query or data point queryFinding all of them is accomplished by the answer of the range-query for the centerThe accuracy and comprehensiveness of analyzing and extracting the abnormal point data are improved by dividing a whole set into a plurality of dimensions, searching the abnormal point data in each dimension, and finally summarizing.
Example 2:
a regional anomaly monitoring method based on edge calculation comprises the following steps:
s1: the method comprises the steps that monitoring characteristic data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device;
s2: presetting a used processing model, and inputting the collected characteristic data into the model;
s3: performing edge calculation and analyzing abnormal data;
s4: identity authentication, which is to authenticate identity information of an operator; after the verification is successful, the abnormal data information can be read through the edge computing device.
The edge calculation method in the step S3 is anomaly detection of high-dimensional data, and includes the steps of:
s31: dividing each dimension of the data space intoEach equal depth interval; the equal depth intervals are spatially defined by the data being collected, and each interval includes equal valuesThe data points of (a);
s32: an equal depth interval is respectively taken on each dimension in the k-dimensional subspace of the data set to form a k-dimensional cube, and the number of data mapping points in the cube is a random number,;
Wherein, thereinRepresentsDimension cubeThe number of points included in the pool,the number of the total points is represented,representing the sparse coefficients of the image data to be processed,when negative, the cube is describedWhere the data point is below the expected value,the smaller the cubeThe more sparse the data in (1).
In S3, the anomaly detection of the high-dimensional data optimizes a problem solving process using a genetic algorithm, where the solving process starts with the selected P legal modes as candidate solutions, and repeatedly goes through several processes such as selection, intersection, and mutation, until a more satisfactory solution is obtained.
The utility model provides an edge computing device, includes data acquisition module, calculation processing module, authentication module and result output module, the monitoring characteristic data that a plurality of thing networking equipment collection that the data acquisition module was used for accessing edge computing device, calculation processing module can accept the data that the acquisition module obtained to according to belong to and predetermine the algorithm calculation and the extraction of abnormal point data, authentication module is used for verifying operating personnel's identity, if verify successful, then allow to read abnormal data result, if verify failure, refuse to read, result output module can show the abnormal point data after the calculation extraction.
In this embodiment: in embodiment 1, the whole set is divided into a plurality of dimensions to be searched, a multidimensional index structure needs to be established, which is time-consuming, and simultaneously, the performance of all index structures rapidly decreases with the increase of the dimensions, so that the performance of the algorithm is not good.
Example 3:
a regional anomaly monitoring method based on edge calculation comprises the following steps:
s1: the method comprises the steps that monitoring characteristic data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device;
s2: presetting a used processing model, and inputting the collected characteristic data into the model;
s3: performing edge calculation and analyzing abnormal data;
s4: identity authentication, which is to authenticate identity information of an operator; after the verification is successful, the abnormal data information can be read through the edge computing device.
The edge calculation method in S3 is an abnormal point detection algorithm based on clustering, during the detection process of abnormal points, firstly, the abnormal points are removed from the clustering analysis for detection, if the number of the removed points is less than the number of actual abnormal points to be detected, then the remaining abnormal points are detected from each cluster, after the clustering analysis is carried out on the data set S, the data object S is divided into n clusters and recorded as n clusters,and get rid ofThe number of outliers, noted,whereinThen, thenEach cluster center is marked asEach abnormal pointThe distance to the cluster center is noted as:wherein, in the step (A),represents the dimensions of the data and is,represents an integer whenWhen is in use, theSolving the distance by using a Manhattan distance formula method; when in useWhen is in use, theSolving the distance by an Euclidean distance formula method, wherein the value of q is determined according to specific conditions; any object rejectedTo each cluster coreIs a distance ofCan be represented by a matrix R, in which the ith row elementIndicating the distance to each cluster centerLet us order,Is the sum of the ith row in the matrix R and represents the distance sum of the ith culled point to each cluster center,the larger the value, the farther it is from each centroid, whenWhen the temperature of the water is higher than the set temperature,before the maximum valueThe removed object is the abnormal point whenWhen it is, thisThe removed points are abnormal points, and the rest points are detected from each clusterAn abnormal point, set clusterAny of the data objectsTo the cluster core where it is locatedA distance ofSelecting from each clusterBefore the maximum valueA point of the object being located at the center, thusCommon elimination in clustersPoints because ofTherefore, it isThen from thisDetection in pointsAn anomaly point.
The utility model provides an edge computing device, includes data acquisition module, calculation processing module, authentication module and result output module, the monitoring characteristic data that a plurality of thing networking equipment collection that the data acquisition module was used for accessing edge computing device, calculation processing module can accept the data that the acquisition module obtained to according to belong to and predetermine the algorithm calculation and the extraction of abnormal point data, authentication module is used for verifying operating personnel's identity, if verify successful, then allow to read abnormal data result, if verify failure, refuse to read, result output module can show the abnormal point data after the calculation extraction.
In this embodiment: in the embodiment 1 and the embodiment 2, the whole data set is used as analysis, abnormal points may be removed as noise, the integrity of the data is damaged, and the accuracy of the data is reduced.
Example 4:
a regional anomaly monitoring method based on edge calculation comprises the following steps:
s1: the method comprises the steps that monitoring characteristic data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device;
s2: presetting a used processing model, and inputting the collected characteristic data into the model;
s3: performing edge calculation and analyzing abnormal data;
s4: identity authentication, which is to authenticate identity information of an operator; after the verification is successful, the abnormal data information can be read through the edge computing equipment
The edge calculation method in S3 may be calculated by the following formula:
wherein the content of the first and second substances,in order to transfer the function values,the amount of the carbon dioxide is the intermediate amount,in order to be a function of the edges,as a function of the diffusion coefficient of the points in the region,in the case of the edge bevel angle,is the angle of inclination of the midpoint of the region;
for points in the collected characteristic data, transfer function values obtainedInconsistent with the collected data, it is called an outlier, otherwise it is a normal point.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (10)
1. A regional anomaly monitoring method based on edge calculation is characterized by comprising the following steps:
s1: the method comprises the steps that monitoring characteristic data collected by a plurality of Internet of things devices accessed to the edge computing device are obtained by the edge computing device;
s2: presetting a used processing model, and inputting the collected monitoring characteristic data into the model;
s3: performing edge calculation and analyzing abnormal data;
s4: identity authentication, which is to authenticate identity information of an operator; after the verification is successful, the abnormal data information can be read through the edge computing device.
2. The method for monitoring regional abnormalities based on edge calculation as claimed in claim 1, wherein in said S3, the edge calculation method includes an edge calculation method based on direct distance of characteristic data and an edge calculation method based on direct density of characteristic data.
3. The method for monitoring regional abnormality based on edge calculation according to claim 2, wherein the edge calculation method based on feature data direct distance comprises the following steps:
s11, inputting all the monitoring characteristic data into the data set S, and inputting a certain data point in the data set SIs called as,Representing the horizontal and vertical coordinate values of the data points;
4. The method for monitoring regional anomalies based on edge calculation as claimed in claim 3, wherein in the step S12, the specific properties are: data setHas data points and data pointsIs greater thanThe algorithm is as follows: index-based methods rely on multidimensional indexing structures by querying nearest neighbors or data pointsThe complexity of the R-trees or kd-trees algorithm based on mostly index structures for finding all data points is realized by using answers of range queries as centersWhereinIs the dimension of the dimension number of the object,is the number of points of the data,representing a data point.
5. The method for monitoring regional anomalies based on edge computation according to claim 2, wherein the method for edge computation based on direct density of feature data includes:
s21: of object PDistance is calledFor any natural numberDefinition of PDistance andis P and some objectDistance between, here data pointThe preset condition is required to be met;
s22: of object PDistance neighborhood calledOf a given PA distance ofOf PThe distance neighborhood contains all the distances to PNot exceedingObject of (2)I.e. by;
S23: the reachable distance of the object P relative to the object O is given by a natural number kThe local reachable density of an object P is the inverse of the average reachable distance of the object P from his MinPts-neighborhood, i.e. the;
S24: local outlier factor of object P:
s25, the local outlier factor OF the object P represents the degree OF the outlier OF P, the larger the local outlier factor is, the more likely it is to be the outlier, otherwise the probability is low, the object L OF close to the core store in the cluster is close to 1 and is not considered to be the local outlier, and the object L OF at the edge OF the cluster or outside the cluster is relatively large.
7. The method for monitoring regional anomalies based on edge computation of claim 1, wherein the edge computation method in S3 is anomaly detection of high-dimensional data, and includes the steps of:
s31: dividing each dimension of the data space intoEach equal depth interval; the equal depth intervals are spatially defined by the data being collected, and each interval includes equal valuesThe data points of (a);
s32: an equal depth interval is respectively taken on each dimension in the k-dimensional subspace of the data set to form a k-dimensional cube, and the number of data mapping points in the cube is a random number,;
Wherein, thereinRepresentsDimension cubeThe number of points included in the pool,the number of the total points is represented,representing the sparse coefficients of the image data to be processed,when negative, the cube is describedWhere the data point is below the expected value,the smaller the cubeThe more sparse the data in (1).
8. The method for monitoring regional anomalies based on edge computation of claim 7, wherein the anomaly points of the high-dimensional data are detected; the abnormal point detection of high dimensional data optimizes the problem solving process by using genetic algorithm, the solving process starts with P selected legal modes as candidate solutions, and repeatedly goes through several processes of selection, intersection, variation and the like until a more satisfactory solution is obtained, the abnormal data analysis also comprises cluster-based abnormal point detection algorithm, in the abnormal point detection process, firstly, abnormal points are removed from the cluster analysis for detection, if the number of the removed points is less than the number of actual abnormal points to be detected, then, the remaining abnormal points are detected from each cluster, after the data set S is subjected to cluster analysis, the data object S is divided into n clusters and recorded,and get rid ofThe number of outliers, noted,whereinThen, thenEach cluster center is marked asEach abnormal pointThe distance to the cluster center is noted as:wherein, in the step (A),represents the dimensions of the data and is,represents an integer whenWhen is in use, theSolving the distance by using a Manhattan distance formula method; when in useWhen is in use, theSolving the distance by an Euclidean distance formula method, wherein the value of q is determined according to specific conditions; any object rejectedTo each cluster coreIs a distance ofCan be represented by a matrix R, in which the ith row elementIndicating the distance to each cluster centerLet us order,Is the sum of the ith row in the matrix R and represents the distance sum of the ith culled point to each cluster center,the larger the value, the farther it is from each centroid, whenWhen the temperature of the water is higher than the set temperature,before the maximum valueThe removed object is the abnormal point whenWhen it is, thisThe removed points are abnormal points, and the rest points are detected from each clusterAn abnormal point, set clusterAny of the data objectsTo the cluster core where it is locatedA distance ofSelecting from each clusterBefore the maximum valueA point of the object being located at the center, thusCommon elimination in clustersPoints because ofTherefore, it isThen from thisDetection in pointsAn anomaly point.
9. The method for monitoring regional abnormality based on edge calculation according to claim 8, wherein the edge calculation method in S3 is calculated by the following formula:
wherein the content of the first and second substances,in order to transfer the function values,the amount of the carbon dioxide is the intermediate amount,in order to be a function of the edges,as a function of the diffusion coefficient of the points in the region,in the case of the edge bevel angle,is the angle of inclination of the midpoint of the region;
10. The edge computing device is characterized by comprising a data acquisition module, a computing processing module, an identity verification module and a result output module, wherein the data acquisition module is used for accessing monitoring feature data acquired by a plurality of Internet of things devices of the edge computing device, the computing processing module can receive the monitoring feature data acquired by the acquisition module and calculate and extract abnormal point data according to a preset algorithm, the identity verification module is used for verifying the identity of an operator, if the verification is successful, the result of reading the abnormal data is allowed, if the verification is failed, the result of reading is refused, and the result output module can display the calculated and extracted abnormal point data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010515822.9A CN111427752A (en) | 2020-06-09 | 2020-06-09 | Regional anomaly monitoring method based on edge calculation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010515822.9A CN111427752A (en) | 2020-06-09 | 2020-06-09 | Regional anomaly monitoring method based on edge calculation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111427752A true CN111427752A (en) | 2020-07-17 |
Family
ID=71558915
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010515822.9A Pending CN111427752A (en) | 2020-06-09 | 2020-06-09 | Regional anomaly monitoring method based on edge calculation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111427752A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150286707A1 (en) * | 2014-04-08 | 2015-10-08 | International Business Machines Corporation | Distributed clustering with outlier detection |
CN109765863A (en) * | 2019-01-21 | 2019-05-17 | 苏州首拓信息科技有限公司 | A kind of device parameter edge calculations method based on cloud platform |
CN109947079A (en) * | 2019-03-20 | 2019-06-28 | 阿里巴巴集团控股有限公司 | Region method for detecting abnormality and edge calculations equipment based on edge calculations |
CN110717942A (en) * | 2018-07-11 | 2020-01-21 | Oppo广东移动通信有限公司 | Image processing method and device, electronic equipment and computer readable storage medium |
-
2020
- 2020-06-09 CN CN202010515822.9A patent/CN111427752A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150286707A1 (en) * | 2014-04-08 | 2015-10-08 | International Business Machines Corporation | Distributed clustering with outlier detection |
CN110717942A (en) * | 2018-07-11 | 2020-01-21 | Oppo广东移动通信有限公司 | Image processing method and device, electronic equipment and computer readable storage medium |
CN109765863A (en) * | 2019-01-21 | 2019-05-17 | 苏州首拓信息科技有限公司 | A kind of device parameter edge calculations method based on cloud platform |
CN109947079A (en) * | 2019-03-20 | 2019-06-28 | 阿里巴巴集团控股有限公司 | Region method for detecting abnormality and edge calculations equipment based on edge calculations |
Non-Patent Citations (1)
Title |
---|
王元明等: "异常数据的检测方法", 《重庆工学院学报(自然科学)》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2022110557A1 (en) | Method and device for diagnosing user-transformer relationship anomaly in transformer area | |
CN110070121B (en) | Rapid approximate K nearest neighbor method based on tree strategy and balanced K mean clustering | |
CN107682319B (en) | Enhanced angle anomaly factor-based data flow anomaly detection and multi-verification method | |
Puzicha et al. | Non-parametric similarity measures for unsupervised texture segmentation and image retrieval | |
Lian et al. | Monochromatic and bichromatic reverse skyline search over uncertain databases | |
CN104462184B (en) | A kind of large-scale data abnormality recognition method based on two-way sampling combination | |
CN105930862A (en) | Density peak clustering algorithm based on density adaptive distance | |
CN109522926A (en) | Method for detecting abnormality based on comentropy cluster | |
CN110795690A (en) | Wind power plant operation abnormal data detection method | |
JP4937395B2 (en) | Feature vector generation apparatus, feature vector generation method and program | |
WO2019200739A1 (en) | Data fraud identification method, apparatus, computer device, and storage medium | |
CN111709668A (en) | Power grid equipment parameter risk identification method and device based on data mining technology | |
JP5014479B2 (en) | Image search apparatus, image search method and program | |
CN115021679A (en) | Photovoltaic equipment fault detection method based on multi-dimensional outlier detection | |
CN115600194A (en) | Intrusion detection method, storage medium and device based on XGboost and LGBM | |
CN110083731B (en) | Image retrieval method, device, computer equipment and storage medium | |
CN117556369B (en) | Power theft detection method and system for dynamically generated residual error graph convolution neural network | |
CN115033591A (en) | Intelligent detection method and system for electricity charge data abnormity, storage medium and computer equipment | |
CN106611016A (en) | Image retrieval method based on decomposable word pack model | |
CN109783586B (en) | Water army comment detection method based on clustering resampling | |
CN111427752A (en) | Regional anomaly monitoring method based on edge calculation | |
CN116450710A (en) | Data analysis tracing method and system based on big data | |
CN112418313B (en) | Big data online noise filtering system and method | |
CN111863124B (en) | Copy number variation detection method, system, storage medium and computer equipment | |
CN115756919A (en) | Root cause positioning method and system for multidimensional data |
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: 20200717 |
|
RJ01 | Rejection of invention patent application after publication |