CN103560925A - IT operation and maintenance index forecasting method utilizing relevance - Google Patents

IT operation and maintenance index forecasting method utilizing relevance Download PDF

Info

Publication number
CN103560925A
CN103560925A CN201310462893.7A CN201310462893A CN103560925A CN 103560925 A CN103560925 A CN 103560925A CN 201310462893 A CN201310462893 A CN 201310462893A CN 103560925 A CN103560925 A CN 103560925A
Authority
CN
China
Prior art keywords
data
index
amp
forecast
conv
Prior art date
Application number
CN201310462893.7A
Other languages
Chinese (zh)
Inventor
不公告发明人
Original Assignee
成都勤智数码科技股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 成都勤智数码科技股份有限公司 filed Critical 成都勤智数码科技股份有限公司
Priority to CN201310462893.7A priority Critical patent/CN103560925A/en
Publication of CN103560925A publication Critical patent/CN103560925A/en

Links

Abstract

The invention discloses an IT operation and maintenance index forecasting method utilizing relevance. According to the method, necessary preprocessing is carried out on historical data, the relevance relation between indexes is determined by calculating a covariance matrix, the fluctuation range of the covariance between the indexes is determined through eigenvalue decomposition, the test covariance is calculated, and the indexes are forecast according to the relevance and the fluctuation range. The IT operation and maintenance index forecasting method utilizing relevance has higher forecasting accuracy.

Description

Utilize the IT O&M indices prediction method of correlation

Technical field

The present invention relates to IT operation management field, especially the monitoring between the index of IT O&M and management domain, be specially a kind of IT O&M indices prediction method of utilizing correlation.

Background technology

IT operation management, IT enterprises or department adopt relevant method, means, technology, system, flow process and document etc., the integrated management that IT running environment (comprising physical environment, hardware environment etc.), IT business system and IT O&M personnel are carried out.Along with IT construction deepen continuously and perfect, the operation maintenance of computer hardware and software system has obtained attention, because this is a new problem producing along with the deep application of computer information technology, therefore how research carries out effective IT operation management, will have vast potential for future development and huge realistic meaning.

Briefly say, the organize content of IT O&M can manage and safeguard through being taken into index.Index, also describe the data of a certain characteristics of objects.The administration behaviour of IT O&M, can be abstracted into the change of data in essence.Therefore, the management of research IT O&M index highly significant.In the present invention, proposition is a kind of IT O&M index intelligent forecasting method, Forecasting Methodology, prediction unit, predictor that utilizes correlation.

Intelligent prediction and prediction, carry out the process of alarm or estimation by the mode of unartificial detection to certain desired value.Intelligentized example is a lot, as the clustering algorithm of pattern recognition being applied to the function of mobile phone or terminal hand-writing input method, can improve input efficiency; For another example some music software provides the function of automatic recommendation song, by recording audience's historical record, predicts, this didactic mode can further meet audience's wish; For another example 360 security guards are to the program updates of operating system with safeguard the function that automatic forecasting is provided, can optimization system, improve system useful life.

Intelligentized theoretical system has developed to obtain comparative maturity, intelligent theoretical method and the means of application mainly comprise at present: (1) adaptation theory system, and this theory is a kind of feedback theory in essence, comprises artificial neural net system, by learning training sample, predict future data; (2) area of pattern recognition, reaches the object of identification by structure different mode system; (3) Optimum Theory system, this theory comprises supporting vector machine model, ant group algorithm, genetic algorithm, linearity and nonlinear restriction model reach the object of optimization aim data by modeling; (4) modern signal processing Domain Theory and method, signal processing method is as moving average adaptive regression model, and filtering method is as Wiener filtering, Kalman filter model, by modeling to future time amount predict, level and smooth or estimate.

In the present invention, will directly not use above-described intelligent method, but utilize correlation.

Between some index of IT O&M, certainly exist correlation.It is example that the WLAN index of take detects, the field intensity signal to noise ratio intensity of WLAN signal directly affects network data bandwidth, even if the connectedness of network is as ping packet success rate, the Congestion Level SPCC of network may affect WEB Authentication target, because when offered load is overweight, the WEB authentication access delay time may increase.In actual application scenarios, because of Cost Problems, some WLAN index should not be monitored constantly, as field intensity signal to noise ratio, and some data can obtain constantly by the mode of software supervision, and between these two kinds of indexs or exist contact between more indexs, in this case, utilize the correlation between index just can overcome the problem that other Intelligent Plan is unpredictable or predictablity rate declines, because no matter whether data are known, correlation between index is constantly to exist, and only needs as adopted the method in the present invention just can reach the effect of prediction.In addition, correlation can also, when some index unknown data dynamic range, forecast whether it exceeds standard.

The Mathematics Proof of correlation is as follows:

For two vector x i=[X i(l), x i(N) I, x j=[x j(I) ..., x j(N)] covariance so between the two can be expressed as conv ( x i , x j ) = x i · x ′ j N - Σ i = 1 N x i N * Σ j = 1 N x ′ j N - - - ( 1 )

By the cross covariance between M index, formed the matrix of the capable M row of M,

Definition coefficient correlation ρ ( x i , x j ) = conv ( x i , x j ) / Dx i Dx j , According to the character of coefficient correlation, auto-correlation coefficient equals that 0, two vector is uncorrelated, and auto-correlation coefficient absolute value equals 1, and two SYSTEM OF LINEAR VECTOR that and if only if are relevant.Thus, we infer, covariance absolute value is more more uncorrelated close to 0, two index, otherwise more relevant.

Summary of the invention

The invention provides a kind of IT O&M indices prediction method of utilizing correlation, the feature of each step of the method is:

(1) upgrade Data Source, training data sample and test data sample data are provided, and wherein the training data of each index is multidimensional, and test sample book is one dimension, along with passage of time, after being incorporated to historical data base, test sample book make training sample huge gradually.

(2) training, comprises data preliminary treatment and data and calculates two steps, and the burr data such as minimax can be eliminated in training sample source after data preliminary treatment, reach smooth effect, thereby provides accurately reasonably Data Source for next step; During data process data calculation procedure after pretreatment, according to formula (1), (2), obtain a covariance matrix, then calculate covariance fluctuation range.

Preferably, first, matrix (2) is done to Eigenvalues Decomposition and obtain

conv=V*D*V -l?????????????(3)

V, D is respectively characteristic vector and characteristic value diagonal matrix, then, retains the larger characteristic value of absolute value, reject little order and equal zero, thereby obtain D', so,

conv'=V*D'*V -1???????????????(4)

Conv' is inevitable is also a symmetrical matrix, and differs from Conv, considers the element of triangular portions on it, defines fluctuation range and is: the Yi Ge circle conv (X of fluctuation range i, X j) so another boundary be

conv"(X i,X j)=2*conv(X i,X j)-conv'(X i,X j)????????????(5)

(3) test, comprises data forecast and two steps of data prediction.

In data forecast step,

Preferably, first, from data source, obtain test sample book, the fluctuation range of any two indexs that obtain according to training module and i and j index average the covariance defining between any two test sample book data is expressed as,

( y i , y j ) = y i * y j - Σ i * Σ j ( 6 )

Can judge conv (y i, y j) whether drop on conv'(X i, X j)~conv " (X i, X j) fluctuation range in, thereby forecast.

Preferably, if known a certain index but cannot forecast whether it exceeds standard, forecast that thought is: find in training module, draw with the maximally related several indexs of this index, if one of them index can be forecast sequentially, stop forecast.

Detecting under the prerequisite of achievement data, can predict index.

Preferably, according to formula (6), the algorithm of Accurate Prediction is: first find a maximally related index j with index i to be measured, then find the maximally related index k with j, can think conv ij-conv ik=conv ij-CONV ik, the equation left side is unknown test covariance, the right is known training covariance.Thereby three systems of linear equations of simultaneous, separate obtain to predict the outcome also separate.Also be that solving equations obtains X

conv ij = X * x j - Σ i Σ j conv ik = X * x k - Σ i Σ k conv ij - conv ik = CONV ij - CONV ik - - - ( 7 )

The present invention also provides a kind of intelligent forecasting and prediction unit that utilizes correlation simultaneously, comprises,

Data source module, the initialization data using existing historical data as training module, selects as far as possible large.Meanwhile, for the test data of constantly updating, be incorporated to tranining database after often testing one group of data, guaranteed upgrading in time of database.

Preferably, when data volume reaches certain scale, carry out packet training, to improve test accuracy.Referring to key diagram 1.

Training module, comprises data pretreatment unit and data computing unit,

Data pretreatment unit,

Preferably, in order to reach, eliminate burr object, to each index, under initial situation, first remove obviously extreme several sample values and retain remaining sample, calculate as several extremely large arithmetic mean M and several extremely little value arithmetic mean m, when at every turn more during new data, if find, data drop on outside M or m, are regarded as burr and reject, the data group of simultaneously rejecting forms new manifold, upgrades M and m.Go in such a manner, make data reach level and smooth effect as far as possible.Shown in key diagram 5.

Data computing unit,

Preferably, because data preprocessing part is eliminated burr processing to each index, may make between two achievement data vectors dimension different, the mode solving is, for burr of an every elimination of index, when shortage of data, use the arithmetic mean of all data acquisition systems above to replace, the error while calculating covariance matrix to reduce;

Preferably, the rule of rejecting less characteristic value is, by additions that take absolute value of all characteristic values, then calculates the ratio of each characteristic value, if this characteristic value ratio is less than as 0.05, claims that characteristic value contribution margin is too small, even it can be considered to rejecting, also it equals zero.Reject manyly, the fluctuation range of calculating is larger.This execution mode can be referring to shown in key diagram 6.

Test module, comprises data forecast unit and data prediction unit,

Data forecast unit, comprises discrimination module and forecast module,

Discrimination module, once some index is measured and just had with reference to scope in reality, therefore without forecast, and measures not with reference to scope for other index, and whether therefore first distinguish index needs to forecast;

Preferably, the algorithm principle of forecast module is first to see whether with maximally related that index of index x to be measured be index known and in known dynamic range, if not continue search, until search out front m, meet the demands, m maximum can reach all known dynamic range index numbers.First is made as i, and index i and x are calculated to covariance conv (x, i), if be less than fluctuation range, forecasting index x does not exceed standard; If fluctuation range is calculated again and the index j of index x correlations otherwise be greater than, if conv (j, i) is less than fluctuation range, forecast that x exceeds standard, otherwise, claim i prediction to lose efficacy, with j, replace i, the flow process of repetition i.So repeatedly, until all front m indexs are all predicted inefficacy, forecast that x does not exceed standard.

This unit specifically can be referring to shown in key diagram 7.

Data prediction unit, the data for predicting that some cannot direct-detection, are divided into discrimination module and prediction module, preferably, according to mentioning the thought of solving an equation in method, carry out.Referring to key diagram 8.

The flow chart of whole device is as shown in key diagram 4.

A kind of IT O&M index intelligent forecasting method, Forecasting Methodology, prediction unit, predictor that utilizes correlation provided by the invention, its intelligent being embodied in: in the time of cannot judging when given data source whether it exceeds standard, usage data test cell, alarm in actual IT O&M system; In the time cannot directly detecting index due to chance failure or additive method, utilize all the other associated desired values and data prediction unit, can predict more accurately this index.

A kind of IT O&M index intelligent forecasting method, Forecasting Methodology, prediction unit, predictor that utilizes correlation provided by the invention, its advantage and feature are: with traditional intelligent prediction or Forecasting Methodology comparison, all need training and testing two steps, but amount of calculation is much smaller, and can reach higher accuracy.

Accompanying drawing explanation

Examples of the present invention will be described by way of reference to the accompanying drawings, wherein

Fig. 1 is every group of number of station work and certain index success rate graph of a relation of prediction;

Fig. 2 is the magnitude relationship figure of a certain test index warning probability and this index;

Fig. 3 is that the predicted value deviation ratio of a certain test index is with the variation relation figure of index size;

Fig. 4 is the flow chart of device;

Fig. 5 is the flow chart of the data pretreatment unit of training module;

Fig. 6 is the flow chart of data computing unit in training module;

Fig. 7 is the flow chart of data forecast unit in test module;

Fig. 8 is the flow chart of data prediction unit in test module;

Fig. 9 is whole method and apparatus system principle schematic diagram.

Embodiment

For making the inventive method and device can reach result and the function of expectation, simultaneously for more clear and intuitive expression method of the present invention, will adopt the simulation result figure of MATLAB describe and show.

In specific embodiment 1, with reference to key diagram 1,

Suppose under real scene, receive altogether 20 achievement data sources, statistical history data, suppose that the initial sampled data of each index is fixed as 1000, and establishing index training data to be measured source is that average is 10, the just too distributed data that variance is 0.1.Consideration divides into groups to enter the processing of training module to it, in theory, for guarantee that fluctuation range calculates accurately, every group of number is unsuitable very few, simultaneously for smoothing processing, group number should not very little, therefore, have a compromise.This routine object be checking when data source is fixedly time, how to distribute these data can reach good performance.For embodiment 2 does foundation.

Shown in key diagram 1, under testing data known cases, set two kinds of situations:

Index test data to be measured equal 10, in scope, and presentation of results, under 1000 data are divided into the scope of every group 100~500, predicated error is lower than 0.1; Test data equals 14 outside scope, and presentation of results, when 1000 data are divided into every group 100~500, can reach better prediction effect relatively, and predicated error is minimum in 0.4 left and right.

By embodiment 1, obtain the allocation proportion of 1000 number grouping numbers and group number, can elect 100 every group as, totally 10 groups, as the foundation of next embodiment.

Meanwhile, this example also illustrated on duty exceeded scope after, its predicted value is very inaccurate, this explanation several indexs relevant to this index have all exceeded scope, because the satisfied condition of predicting, so this situation does not meet application category of the present invention.

In specific embodiment 2, refer to key diagram 2.

Suppose under real scene, 10 of index numbers, training data adds up to 10000, be divided into 100 groups, every group of 100 data, training data source is the random number between 0~1, preset desired value to be measured often increases progressively 0.5 until approach 20 from 0, puts 1 for reporting to the police (exceeding standard), and 0 does not report to the police.In theory, when data are got over away from this scope of 0~1, report to the police and should be 1, otherwise be 0.The algorithm robustness providing due to method exists, so, after smoothing processing, by warning probability, reflect forecast performance.

Shown in key diagram 2, when initialize data (wait forecasting test data) is gradually away from 1 time, warning probability rises gradually, until approach 1.In reality, the mode of solution is, sets up a threshold values, reports to the police, otherwise do not report to the police when certain test data obtains warning probability higher than threshold values.

This embodiment has verified the validity of inventive method data forecasts, and a solution is provided.

In specific embodiment 3, refer to key diagram 3.

Suppose under real scene, index number is 20, and there are 1000 data in every group of index training data source, and achievement data to be measured source is to take 10 as average, 0.1 random number that is variance, and preset index test data to be measured are incremented to 15 from 5 with 0.5, calculate prediction deviation rate.

Shown in key diagram 3, when presetting range is during in 10 scope, the I of predicated error is lower than 0.1, otherwise predicated error is increasing.This key diagram, the same manner as in Example 1, illustrated that the Forecasting Methodology that the present invention provides has higher precision.

Claims (4)

1. the IT O&M indices prediction method of utilizing correlation, is characterized in that, described method comprises three key steps: upgrade Data Source, training and forecast;
Upgrade Data Source, specifically comprise with initialized historical data as training data, be constantly incorporated to test data simultaneously, renewal historical data base;
Training, specifically comprises data preliminary treatment and data calculation procedure, data preliminary treatment, and input training data, eliminates burr data and carries out smoothing processing; Described data calculation procedure, calculates covariance matrix, determines the correlative relationship between index, it is carried out to Eigenvalues Decomposition and process definite fluctuation range;
Covariance matrix utilizes covariance formula to obtain, and the computing formula of covariance is: if x i=[x 1..., x n], x j=[x' 1..., x' n] representing i and j achievement data, both covariances can be expressed as so conv ( x i , x j ) = x i · x ′ j N - Σ i = 1 N x i N * Σ j = 1 N x ′ j N , This value is more close to zero, and two indexs are more uncorrelated, otherwise more relevant, thereby determines index related relation; After calculating covariance matrix, it is carried out to Eigenvalues Decomposition, by the larger reservation of absolute value in all characteristic values, rejecting is close to zero part, again revert to new covariance matrix, establishing the covariance size of i and j index in the new covariance matrix recovering after the past, characteristic value was processed is conv'(x again i, x j), establishing it for fluctuation range Yi Ge circle, another boundary is conv''(x so i, x j)=2*conv (x i, x j)-conv'(x i, x j), thereby obtain any two indexes covariance fluctuation range, be conv'(x i, x j)~conv''(x i, x j).
Data forecast, by the correlative relationship and the fluctuation range that obtain after training, utilizes index of correlation to forecast whether it exceeds standard.
2. utilize according to claim 1 the IT O&M indices prediction method of correlation, it is characterized in that, the algorithm of the elimination burr of described data pre-treatment step is: for some indexs, initialization, from its historical data base, select three data of minimax, form burr collection, with the average of other data, fill its position; After upgrading historical data, the data that newly add, see whether it is greater than the average of burr collection, if be greater than, add burr collection, otherwise, as normal data, enter data calculation procedure.
3. utilize according to claim 1 the IT O&M indices prediction method of correlation, it is characterized in that, condition and the implication of data forecast are: the data of index to be measured have recorded and have been under the situation of non-burr data, the standard of there is no goes to define its excursion, therefore utilizes correlation indirectly to forecast whether it exceeds standard.
4. utilize according to claim 3 the IT O&M indices prediction method of correlation, be further characterized in that, the algorithm steps of described data forecast is: known a certain test index value X, first find maximally related other several indexs of desired value X that obtain with this test that draw in training module, according to sequencing, needing only some indexs can forecast, stops forecast.
CN201310462893.7A 2011-05-04 2011-05-04 IT operation and maintenance index forecasting method utilizing relevance CN103560925A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310462893.7A CN103560925A (en) 2011-05-04 2011-05-04 IT operation and maintenance index forecasting method utilizing relevance

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310462893.7A CN103560925A (en) 2011-05-04 2011-05-04 IT operation and maintenance index forecasting method utilizing relevance

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
CN2011101141501A Division CN102195814B (en) 2011-05-04 2011-05-04 Method and device for forecasting and predicting by using relevant IT (Information Technology) operation and maintenance indexes

Publications (1)

Publication Number Publication Date
CN103560925A true CN103560925A (en) 2014-02-05

Family

ID=50015079

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310462893.7A CN103560925A (en) 2011-05-04 2011-05-04 IT operation and maintenance index forecasting method utilizing relevance

Country Status (1)

Country Link
CN (1) CN103560925A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103824130A (en) * 2014-02-27 2014-05-28 武汉理工大学 Grain condition forecasting and early warning method and system based on SVM
CN104143128A (en) * 2014-07-21 2014-11-12 快威科技集团有限公司 Information system security evaluation index development method and device
CN106487570A (en) * 2015-09-02 2017-03-08 中国移动通信集团公司 A kind of method and apparatus of assessment network performance index variation tendency

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020169562A1 (en) * 2001-01-29 2002-11-14 Gregory Stephanopoulos Defining biological states and related genes, proteins and patterns
CN101267362A (en) * 2008-05-16 2008-09-17 亿阳信通股份有限公司 A dynamic identification method and its device for normal fluctuation range of performance normal value
CN101771758A (en) * 2008-12-31 2010-07-07 北京亿阳信通软件研究院有限公司 Dynamic determine method for normal fluctuation range of performance index value and device thereof
CN101964997A (en) * 2009-07-21 2011-02-02 中国移动通信集团黑龙江有限公司 Method and device for carrying out early warning on network performance

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020169562A1 (en) * 2001-01-29 2002-11-14 Gregory Stephanopoulos Defining biological states and related genes, proteins and patterns
CN101267362A (en) * 2008-05-16 2008-09-17 亿阳信通股份有限公司 A dynamic identification method and its device for normal fluctuation range of performance normal value
CN101771758A (en) * 2008-12-31 2010-07-07 北京亿阳信通软件研究院有限公司 Dynamic determine method for normal fluctuation range of performance index value and device thereof
CN101964997A (en) * 2009-07-21 2011-02-02 中国移动通信集团黑龙江有限公司 Method and device for carrying out early warning on network performance

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103824130A (en) * 2014-02-27 2014-05-28 武汉理工大学 Grain condition forecasting and early warning method and system based on SVM
CN104143128A (en) * 2014-07-21 2014-11-12 快威科技集团有限公司 Information system security evaluation index development method and device
CN106487570A (en) * 2015-09-02 2017-03-08 中国移动通信集团公司 A kind of method and apparatus of assessment network performance index variation tendency
CN106487570B (en) * 2015-09-02 2019-06-25 中国移动通信集团公司 A kind of method and apparatus for assessing network performance index variation tendency

Similar Documents

Publication Publication Date Title
Chen et al. Road traffic congestion monitoring in social media with hinge-loss Markov random fields
Flury et al. Bayesian inference based only on simulated likelihood: particle filter analysis of dynamic economic models
Dray et al. Community ecology in the age of multivariate multiscale spatial analysis
Capozzoli et al. Fault detection analysis using data mining techniques for a cluster of smart office buildings
Huang et al. Spatiotemporal analysis of rural–urban land conversion
Zuber et al. High-dimensional regression and variable selection using CAR scores
Bifet et al. Pitfalls in benchmarking data stream classification and how to avoid them
CN102110365B (en) Road condition prediction method and road condition prediction system based on space-time relationship
Aghdam et al. Feature selection using particle swarm optimization in text categorization
Papacharalampous et al. Comparison of stochastic and machine learning methods for multi-step ahead forecasting of hydrological processes
CN103077402B (en) Partial discharge of transformer mode identification method based on singular value decomposition algorithm
CN103020978B (en) SAR (synthetic aperture radar) image change detection method combining multi-threshold segmentation with fuzzy clustering
Giam et al. Quantifying variable importance in a multimodel inference framework
Liu A method of SVM with normalization in intrusion detection
US7769561B2 (en) Robust sensor correlation analysis for machine condition monitoring
CN104572391B (en) Monitoring alarm tactics configuring method and device, monitoring alarm method and device
CN103033362B (en) Gear fault diagnosis method based on improving multivariable predictive models
Damle et al. Flood prediction using time series data mining
CN103176981B (en) A kind of event information excavates and the method for early warning
KR20050007306A (en) Processing mixed numeric and/or non-numeric data
Li et al. A comparison of fuzzy clustering algorithms for bearing fault diagnosis
CN103730006A (en) Short-time traffic flow combined forecasting method
Farsadnia et al. Identification of homogeneous regions for regionalization of watersheds by two-level self-organizing feature maps
CN102025531B (en) Filling method and device thereof for performance data
EP3112960B1 (en) Combined method for detecting anomalies in a water distribution system

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20140205

WD01 Invention patent application deemed withdrawn after publication