CN112487048A - Correlation analysis method and device based on time series abnormal fluctuation - Google Patents
Correlation analysis method and device based on time series abnormal fluctuation Download PDFInfo
- Publication number
- CN112487048A CN112487048A CN202011205792.8A CN202011205792A CN112487048A CN 112487048 A CN112487048 A CN 112487048A CN 202011205792 A CN202011205792 A CN 202011205792A CN 112487048 A CN112487048 A CN 112487048A
- Authority
- CN
- China
- Prior art keywords
- time
- fluctuation
- abnormal fluctuation
- abnormal
- time series
- 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Fuzzy Systems (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a correlation analysis method and a device based on time series abnormal fluctuation, wherein the method comprises the following steps: extracting the characteristics of the time sequences through characteristic engineering to obtain the time sequence fluctuation characteristics corresponding to each time sequence; and performing correlation judgment according to the fluctuation characteristics of the time sequences to obtain the abnormal fluctuation incidence relation result between each time sequence. The method can efficiently perform abnormal fluctuation correlation analysis.
Description
Technical Field
The invention relates to the technical field of data analysis, in particular to a correlation analysis method and device based on time series abnormal fluctuation.
Background
Internet services have become an indispensable part of modern people's life, and in order to manage each service of an internet company, operation and maintenance personnel of the company usually monitor and collect thousands of key performance indexes, and the forms of the indexes (such as the number of service requests, the success rate of the service requests, etc.) are mostly time series. In actual operation and maintenance management work, service interruption of a company is inevitable due to various reasons (such as network interruption, malicious attacks and the like). When a service fails, a plurality of time sequence data related to a failure root cause also have abnormal fluctuation, and the abnormal fluctuation is also transmitted to other time sequences with business association or module calling relationship to form an alarm storm. Most of these alarms are redundant and only a few require attention and resolution by the operation and maintenance personnel. In addition, the time series with abnormal fluctuation are interlaced together, so that troubleshooting work is time-consuming and labor-consuming and is very difficult. By automatically mining the abnormal fluctuation relation among the time sequences, the operation and maintenance personnel can be helped to carry out troubleshooting more efficiently and intelligently.
There are many time series correlation analysis algorithms in academia and industry. The linear relation and the prediction Correlation among time sequences are concerned by Pearson Correlation Coefficient (Pearson Correlation Coefficient), Spearman Correlation Coefficient (Spearman Correlation Coefficient), Grangejen causal analysis algorithm (Granger Causality), and the like, and the abnormal fluctuation characteristics of the time sequences cannot be accurately described and captured. The J-measure algorithm needs to perform anomaly detection on the time sequence to obtain two classification values of the time sequence, but the accuracy of the anomaly detection is very challenging. The Structure-of-information Graphs algorithm only focuses on the outlier part of the time sequence and cannot adapt to more various time sequence types and abnormal types.
The problems concerned by the association analysis algorithms are different from the concept essence of the association analysis of the abnormal fluctuation, or only focus on the outlier part of the time series, so that the method cannot adapt to more various time series and abnormal fluctuation types. Therefore, these algorithms do not solve the problem of abnormal fluctuation correlation of time series well.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
To this end, an object of the present invention is to provide a correlation analysis method based on time series abnormal fluctuations, which can efficiently perform correlation analysis of abnormal fluctuations.
Another object of the present invention is to provide a correlation analysis apparatus based on time series abnormal fluctuation.
In order to achieve the above object, an embodiment of an aspect of the present invention provides a correlation analysis method based on time series abnormal fluctuation, including:
extracting the characteristics of the time sequences through characteristic engineering to obtain the time sequence fluctuation characteristics corresponding to each time sequence;
and performing correlation judgment according to the time sequence fluctuation characteristics to obtain an abnormal fluctuation incidence relation result between each time sequence.
In order to achieve the above object, another embodiment of the present invention provides a correlation analysis apparatus based on abnormal fluctuation of time series, including:
the characteristic extraction module is used for extracting the characteristics of the time sequences through characteristic engineering to obtain the time sequence fluctuation characteristics corresponding to each time sequence;
and the correlation analysis module is used for carrying out correlation judgment according to the time sequence fluctuation characteristics to obtain an abnormal fluctuation correlation relation result between each time sequence.
According to the correlation analysis method and device based on time series abnormal fluctuation, the characteristic extraction is carried out on the multiple time series through the characteristic engineering, and the time series fluctuation characteristic corresponding to each time series is obtained; and performing correlation judgment according to the fluctuation characteristics of the time sequences to obtain the abnormal fluctuation incidence relation result between each time sequence. This makes it possible to efficiently perform correlation analysis of abnormal fluctuations.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flow chart of a method for correlation analysis based on time series anomalous fluctuations in accordance with one embodiment of the present invention;
FIG. 2 is a diagram of a method architecture for correlation analysis based on time series anomalous fluctuations in accordance with one embodiment of the present invention;
FIG. 3 is a diagram illustrating an example of a time series, prediction series, and fluctuation signature series according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating an abnormal fluctuation characteristic curve corresponding to 6 time series and time series according to an embodiment of the present invention;
FIG. 5 is a heat map representation of an example of alarm compression according to one embodiment of the present invention, with a time series within a box being an alarm cluster;
figure 6 is a diagram illustrating an example of a possible reason for recommending TOPN according to one embodiment of the present invention;
FIG. 7 is a diagram of an example of automatically building a wave propagation chain on a database service, in accordance with one embodiment of the present invention;
fig. 8 is a schematic structural diagram of a correlation analysis apparatus based on time series abnormal fluctuation according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
The following describes a correlation analysis method and apparatus based on time series abnormal fluctuation according to an embodiment of the present invention with reference to the drawings.
First, a correlation analysis method based on time series abnormal fluctuations proposed according to an embodiment of the present invention will be described with reference to the drawings.
Fig. 1 is a flowchart of a correlation analysis method based on time series abnormal fluctuation according to an embodiment of the present invention.
As shown in fig. 1, the correlation analysis method based on abnormal fluctuation of time series includes the following steps:
and step S1, performing feature extraction on the plurality of time sequences through feature engineering to obtain time sequence fluctuation features corresponding to each time sequence.
And step S2, performing correlation judgment according to the fluctuation characteristics of the time sequences to obtain the abnormal fluctuation incidence relation result among the time sequences.
The invention provides an unsupervised time sequence abnormal fluctuation correlation analysis algorithm.
As shown in fig. 2, which is an overall architecture of the time-series abnormal fluctuation correlation analysis algorithm, in an embodiment of the present invention, the input is two time-series curves, and the output result is an abnormal fluctuation correlation relationship between the two time-series curves, that is: and if the two time sequences are related to each other, outputting the sequence of the fluctuation and the direction of the abnormal fluctuation at the same time. The algorithm architecture is integrally divided into two core parts: feature engineering and correlation measurements.
First, a problem definition is made, S ═ S for a time series1,s2,…,sm]，siIs the data of the time series S at time i, and m is the length of the time series. For a single time series, the data time interval is required to be the same during data acquisition and preprocessing. For two time series, if the time intervals are different, the least common multiple of the time intervals of the two series can be taken as the common time interval. Prediction sequence P ═ P for time series S1,p2,…,pm]，piIs siThe predicted value of (2). Thus, a time series of prediction error sequences F ═ F1,f2,…,fm]，fi＝si-pi. For a time series, the normal part is easily and accurately predicted, but the abnormal fluctuation part is usually caused by some unpredictable burst factors and is difficult to predict. Prediction error ofThe method can be well used for representing the abnormal fluctuation characteristics of the time series, and therefore, the method uses the prediction error sequence of the time series to represent the abnormal fluctuation characteristics of the time series. Fig. 3 shows an example of a time series, a prediction series, and a fluctuation feature series. In addition, the prediction algorithm and corresponding parameters used to obtain the prediction sequence are the feature extractor.
The invention aims to mine the abnormal fluctuation incidence relation of two time sequences, and the abnormal fluctuation incidence relation can help troubleshooting. Specifically, this relationship includes three specific problems: whether the abnormal fluctuations of the two time series are correlated; if so, how the abnormal fluctuation is in sequence (simultaneous occurrence or abnormal fluctuation of a certain time series curve occurs first); whether the directions of the abnormal undulations coincide (the abnormal undulation directions coincide or are opposite).
Further, in one embodiment of the present invention, whether the abnormal fluctuation between the plurality of time series is related or not is judged by the abnormal fluctuation characteristic curve.
The method and the device have the advantages that the original time series curve is difficult to judge the abnormal fluctuation incidence relation, so the method and the device rely on the abnormal fluctuation characteristic curve of the time series. If the characteristic curves of the abnormal fluctuation of the two time series are correlated (i.e. the abnormal fluctuation occurs at the same time or has a certain phase difference), the fluctuation of the original two time series is correlated. Fig. 4 shows 6 time series and their corresponding fluctuation characteristics. As can be seen in FIG. 4, S1And S2、S3Fluctuation is related, but with S4The fluctuation is irrelevant; s1Prior to S2And S3Occurs and the wave direction is opposite; s2And S3The wave motion of (2) occurs simultaneously and the wave motion direction is the same.
The purpose of feature engineering is to find a suitable abnormal fluctuation feature extractor for time series, and a single feature extractor cannot be applied to all time series due to different features of different time series. Thus, if enough excellent time series prediction algorithms are provided that are filtered, then for any one time series, there will always be one algorithm that can extract the appropriate fluctuation characteristics for that time series.
The feature engineering mainly comprises two steps of feature extraction and feature amplification.
1) Feature extraction: as shown in table 1, the feature extraction mainly includes the pre-selected prediction models and parameters, and a total of 86 feature extractors are provided, so that for a time series, 86 fluctuation features can be obtained.
TABLE 1 prediction algorithm used in feature engineering of the present invention and its feature extractor
2) Feature amplification: the time sequence is normal in most of time range, has no great abnormal fluctuation and only has random noise. Only when the online service is affected will an abnormal fluctuation occur. Therefore, the number of abnormal fluctuations is much smaller than normal data. In order to attenuate the effect of noise, the invention proposes an improved version of the excitation function: the larger the fluctuation degree of a time series, the larger the fluctuation characteristic is amplified, so that the abnormal fluctuation characteristic is more distinctive and the final correlation measurement is more helpful.
Wherein x is a characteristic sequence obtained after characteristic extraction, and a and b are hyperparameters of an improved version excitation function. Wherein a is the growth rate of the function f, the larger the value of a is, the faster the growth rate of the function f is, and a generally takes a value of 0.5. b is the truncated value of the function, and if | x | > b, the value of the function f does not grow.
After the amplified abnormal fluctuation features are extracted, the correlation calculation is carried out next. If the two time series are related to abnormal fluctuation, at least the abnormal fluctuation characteristics from the two time series are related. The purpose of the correlation measurement is to obtain a correlation result by calculating the abnormal fluctuation characteristics according to the characteristic engineering. Next, the feature engineering and correlation measurement will be specifically described. When the correlation judgment is performed on the abnormal fluctuation characteristics of the two time series, the slight deformation and the phase difference of the time series need to be considered. Therefore, Cross-Correlation (Cross-Correlation) was chosen to measure the Correlation results of two anomalous fluctuation features, since the algorithm can adapt well to slight deformations of the curve and phase changes.
And (3) performing total calculation on all the abnormal fluctuation characteristics (respectively expressed by G and H, wherein G and H are respectively one of the abnormal fluctuation characteristic sequences in G and H) of the two time series (X and Y), and selecting the correlation result with the largest absolute value as a final result. If the result is larger than a threshold (the threshold can be set through manual experience, and the value range is 0-1), the abnormal fluctuation is relevant, otherwise, the abnormal fluctuation is irrelevant.
GwDenotes G shifts to the right w (w)>0) Unit (if w)<0, G shifted to the left by-w units), l is the length of the anomalous fluctuation signature sequences G and H.
The abnormal fluctuation sequence can be judged by the cross-correlation displacement result, if the displacement w is more than 0, X is prior to Y fluctuation; if the displacement w is less than 0, then X fluctuates later than Y; if the displacement w is equal to 0, then X and Y fluctuations occur simultaneously.
The direction of the abnormal fluctuation can be passed through the original correlation result CC (G)wAnd H) judging the positive or negative of the product. If original result CC (G)wH) is positive, the direction of the abnormal fluctuation is consistent, if the result CC (G)wH) is negative, and the direction of abnormal fluctuation is opposite.
The analysis method of the invention is tested in the actual production environment of the Internet company, and the experimental part comprises two data sets: data set 1: the shapes of the original curve data of the two time sequences related to the abnormal fluctuation are different; data set 2: the shape of the raw data curves of the two time series associated with the abnormal fluctuation is similar. To prove the effect of the invention, four kinds of correlation analysis work are selected as comparison algorithms, and table 2 shows the results of the invention and the comparison algorithms on two data sets. The results show that the present invention works better than other comparative algorithms on both datasets and that F1-score on the three anomalous fluctuation correlation problems is greater than 0.84, 0.92, and 0.95, respectively.
TABLE 2 Effect of the invention and comparison algorithms on two data sets, N/A indicates no result
Specifically, the abnormal fluctuation correlation result of the time series can help troubleshooting in the following three aspects: alarm compression, recommending possible causes of TOPN, and constructing an abnormal fluctuation propagation chain. Next, an application example of the present invention in troubleshooting work will be described.
And (3) alarm compression: for a large number of time series, the invention calculates the degree of abnormal fluctuation correlation between every two of all the time series, and then uses K-means to perform clustering (K can be selected by using a contour coefficient method). The time series within each class can be treated as a cluster of alarms, with alarms occurring as a whole at the time of alarm. FIG. 5 shows an example of the present invention using the fluctuation correlation results of 24 time series to cluster 3 alarms. According to the invention, the alarms are compressed on 699 time sequences, and 208 alarms are finally formed, wherein the compression efficiency is 70.24% on the premise of compression accuracy of 0.9748.
Possible reasons for recommending TOPN: as shown in fig. 6, for any time series X, the invention can find the TOPN-related time series curve of the time series in a massive time series data set. When troubleshooting, the operation and maintenance personnel can preferentially check the TOPN curve to perform rapid anomaly analysis on the time series X. The invention makes TOP5 recommendation on 117 time series, and the recommendation accuracy is 0.8051.
Constructing an abnormal wave propagation chain: the invention constructs an abnormal fluctuation propagation chain by analyzing the fluctuation propagation relation among the time sequences, and the abnormal fluctuation propagation chain can reflect how the fluctuations among different time sequences are related together. Fig. 7 shows an example of automatically constructing a wave propagation chain on a database service according to the present invention, and compared with a manual method, the present invention can automatically and accurately construct an abnormal wave propagation chain without expert knowledge. FIG. 7 (left) shows an artificially constructed chain of abnormal wave propagation; fig. 7 (right) shows an abnormal wave propagation chain automatically constructed by the present invention. (arrows indicate the order of the abnormal fluctuation,; "+" indicates that the abnormal fluctuation direction is consistent, and "-" indicates that the abnormal fluctuation direction is opposite).
The correlation analysis method based on the abnormal fluctuation of the time series provided by the embodiment of the invention comprises the following steps of
Next, a correlation analysis apparatus based on time-series abnormal fluctuations proposed according to an embodiment of the present invention is described with reference to the drawings.
Fig. 8 is a schematic structural diagram of a correlation analysis apparatus based on time series abnormal fluctuation according to an embodiment of the present invention.
As shown in fig. 8, the correlation analysis device based on time-series abnormal fluctuation includes: a feature extraction module 801 and an association analysis module 802.
The feature extraction module 801 is configured to perform feature extraction on the multiple time sequences through feature engineering to obtain a time sequence fluctuation feature corresponding to each time sequence.
And the association analysis module 802 is configured to perform correlation judgment according to the time sequence fluctuation characteristics to obtain an abnormal fluctuation association relationship result between each time sequence.
Further, in an embodiment of the present invention, the method further includes: and the output module is used for simultaneously outputting the sequence of the fluctuation and the direction of the abnormal fluctuation if the abnormal fluctuation of the two time sequences is related.
Further, in one embodiment of the present invention, the abnormal fluctuation feature of the time series is represented by a prediction error series of the time series.
Further, in one embodiment of the present invention, whether the abnormal fluctuation between the plurality of time series is related or not is judged by the abnormal fluctuation characteristic curve.
It should be noted that the foregoing explanation of the method embodiment is also applicable to the apparatus of this embodiment, and is not repeated herein.
According to the correlation analysis device based on the abnormal fluctuation of the time series, which is provided by the embodiment of the invention, the characteristic extraction is carried out on the time series through the characteristic engineering, so that the fluctuation characteristic of the time series corresponding to each time series is obtained; and performing correlation judgment according to the fluctuation characteristics of the time sequences to obtain the abnormal fluctuation incidence relation result between each time sequence. This makes it possible to efficiently perform correlation analysis of abnormal fluctuations.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A correlation analysis method based on time series abnormal fluctuation is characterized by comprising the following steps:
extracting the characteristics of the time sequences through characteristic engineering to obtain the time sequence fluctuation characteristics corresponding to each time sequence;
and performing correlation judgment according to the time sequence fluctuation characteristics to obtain an abnormal fluctuation incidence relation result between each time sequence.
2. The method of claim 1, further comprising: if the abnormal fluctuation of the two time sequences is related, outputting the sequence of the fluctuation and the direction of the abnormal fluctuation at the same time.
3. The method of claim 1, wherein the abnormal fluctuation characteristics of the time series are represented by a prediction error sequence of the time series.
4. The method according to claim 1, wherein whether the abnormal fluctuation between the plurality of time series is related is judged by an abnormal fluctuation characteristic curve.
5. The method of claim 1, further comprising: and performing feature amplification on the extracted features.
6. The method of claim 1, wherein the correlation result of two abnormal fluctuation characteristics is measured by cross-correlation.
7. A correlation analysis device based on abnormal fluctuation of time series is characterized by comprising:
the characteristic extraction module is used for extracting the characteristics of the time sequences through characteristic engineering to obtain the time sequence fluctuation characteristics corresponding to each time sequence;
and the correlation analysis module is used for carrying out correlation judgment according to the time sequence fluctuation characteristics to obtain an abnormal fluctuation correlation relation result between each time sequence.
8. The correlation analysis device based on time-series abnormal fluctuations according to claim 7, characterized by further comprising: and the output module is used for simultaneously outputting the sequence of the fluctuation and the direction of the abnormal fluctuation if the abnormal fluctuation of the two time sequences is related.
9. The correlation analysis device according to claim 7, wherein the time-series abnormal fluctuation characteristics are represented by a time-series prediction error series.
10. The apparatus according to claim 7, wherein the abnormal fluctuation characteristics curve is used to determine whether the abnormal fluctuation between the plurality of time series is correlated.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2020111736921 | 2020-10-28 | ||
| CN202011173692 | 2020-10-28 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112487048A true CN112487048A (en) | 2021-03-12 |
Family
ID=74926362
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011205792.8A Pending CN112487048A (en) | 2020-10-28 | 2020-11-02 | Correlation analysis method and device based on time series abnormal fluctuation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112487048A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113422763A (en) * | 2021-06-04 | 2021-09-21 | 桂林电子科技大学 | Alarm correlation analysis method constructed based on attack scene |
| CN113421020A (en) * | 2021-07-13 | 2021-09-21 | 神策网络科技（北京）有限公司 | Multi-index abnormal point contact ratio analysis method |
| CN115186007A (en) * | 2022-07-08 | 2022-10-14 | 北京普利永华科技发展有限公司 | Airborne data identification real-time display method and system for monitoring and reminding |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015176565A1 (en) * | 2014-05-22 | 2015-11-26 | 袁志贤 | Method for predicting faults in electrical equipment based on multi-dimension time series |
| CN110457184A (en) * | 2018-05-07 | 2019-11-15 | 中国石油化工股份有限公司 | Associated chemical industry exception causality analysis and figure methods of exhibiting are fluctuated based on timing |
-
2020
- 2020-11-02 CN CN202011205792.8A patent/CN112487048A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015176565A1 (en) * | 2014-05-22 | 2015-11-26 | 袁志贤 | Method for predicting faults in electrical equipment based on multi-dimension time series |
| CN110457184A (en) * | 2018-05-07 | 2019-11-15 | 中国石油化工股份有限公司 | Associated chemical industry exception causality analysis and figure methods of exhibiting are fluctuated based on timing |
Non-Patent Citations (1)
| Title |
|---|
| YA SU 等: "CoFlux: Robustly Correlating KPIs by Fluctuations for Service Troubleshooting", 《2019 IEEE/ACM 27TH INTERNATIONAL SYMPOSIUM ON QUALITY OF SERVICE (IWQOS)》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113422763A (en) * | 2021-06-04 | 2021-09-21 | 桂林电子科技大学 | Alarm correlation analysis method constructed based on attack scene |
| CN113422763B (en) * | 2021-06-04 | 2022-10-25 | 桂林电子科技大学 | Alarm correlation analysis method constructed based on attack scene |
| CN113421020A (en) * | 2021-07-13 | 2021-09-21 | 神策网络科技（北京）有限公司 | Multi-index abnormal point contact ratio analysis method |
| CN115186007A (en) * | 2022-07-08 | 2022-10-14 | 北京普利永华科技发展有限公司 | Airborne data identification real-time display method and system for monitoring and reminding |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6658540B2 (en) | System analysis device, system analysis method and program | |
| EP3125057B1 (en) | System-analyzing device, analysis-model generation method, system analysis method, and system-analyzing program | |
| EP2442288A1 (en) | Device abnormality monitoring method and system | |
| CN112487048A (en) | Correlation analysis method and device based on time series abnormal fluctuation | |
| CN110457184B (en) | Chemical engineering abnormal cause and effect analysis and graph display method based on time sequence fluctuation correlation | |
| TWI662424B (en) | Selection method of leading auxiliary parameters and method for pre-diagnosis of equipment maintenance by combining key parameters and leading auxiliary parameters | |
| KR102440335B1 (en) | A method and apparatus for detecting and managing a fault | |
| Han et al. | Combining FAP, MAP and correlation analysis for multivariate alarm thresholds optimization in industrial process | |
| JP2017010263A (en) | Preprocessor of abnormality sign diagnosis device and processing method of the preprocessor | |
| CN111209934B (en) | Fan fault pre-alarm method and system | |
| CN111931834B (en) | Method, equipment and storage medium for detecting abnormal flow data in extrusion process of aluminum profile based on isolated forest algorithm | |
| CN117131110B (en) | Method and system for monitoring dielectric loss of capacitive equipment based on correlation analysis | |
| US11120350B2 (en) | Multilevel pattern monitoring method for industry processes | |
| CN107490964A (en) | A kind of rotating machinery fault feature reduction method of feature based evidence discretization | |
| CN117076258A (en) | Remote monitoring method and system based on Internet cloud | |
| CN105607631A (en) | Batch process weak fault model control limit establishment method and weak fault monitoring method | |
| Janka et al. | Detecting and locating patterns in time series using machine learning | |
| CN117278591A (en) | Park abnormity alarm system based on cloud platform | |
| Miao et al. | Singularity detection in machinery health monitoring using Lipschitz exponent function | |
| CN115169650B (en) | Equipment health prediction method for big data analysis | |
| Yu et al. | A hybrid learning-based model for on-line monitoring and diagnosis of out-of-control signals in multivariate manufacturing processes | |
| Laumonier et al. | Towards alarm flood reduction | |
| KR102486463B1 (en) | Method and Apparatus for Real Time Fault Detection Using Time series data According to Degradation | |
| CN115358137A (en) | Bearing metal temperature prediction method based on CNN-GRU algorithm | |
| Yang et al. | Information theory based probabilistic approach to blade damage detection of turbomachine using sensor 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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210312 |
|
| WD01 | Invention patent application deemed withdrawn after publication |