The content of the invention
It is an object of the invention to provide a kind of monitoring method, system and equipment, false alarm is reducedProbability, improves the accuracy of alarm.
To achieve the above object, the present invention proposes a kind of monitoring method, applied to cloud computing platform,Including：
Obtain the monitoring data for specifying the period；
Based on the monitoring data, alarm monitoring is carried out according to default alarm trigger condition.
Further, the above method can also have the characteristics that, described to be based on the monitoring data, rootAlarm monitoring is carried out according to default alarm trigger condition, including：Monitoring is extracted from the monitoring dataSampled data set of the target in period T, is designated as first set { xi, the period T'sTermination time point is current time；
Doubtful noise x is filtered out from the first setj, the doubtful noise, which refers to, is more than default thresholdValue th sampled data, second set { x is designated as by doubtful noise setj}；
The irrelevance stdr of doubtful noise in the period T is calculated,
Wherein, std is the first set { xiStandard deviation,Std ' for instituteState second set { xjRelative standard deviation,S is the second set { xj}In data amount check,For the first set { xiAverage value,
Judge whether the irrelevance stdr meets default first alarm trigger condition, first reportWarning trigger condition is：The irrelevance stdr is more than or equal to default irrelevance threshold value；
Alarm monitoring is carried out according to the judged result of the described first alarm trigger condition.
Further, the above method can also have the characteristics that, described to be triggered according to the described first alarmThe judged result of condition carries out alarm monitoring, including：
In the case where the irrelevance stdr meets the first alarm trigger condition, triggering alarm.
Further, the above method can also have the characteristics that, described to be based on the monitoring data, rootAlarm monitoring is carried out according to default alarm trigger condition, including：
In the case where the irrelevance stdr does not meet the first alarm trigger condition, when will be describedBetween section T be divided into n period Ti, n, i are natural number, and n >=i >=0；
Period T is calculated respectivelyiInterior sampled data removes the average value after doubtful noise, is designated as denoisingAverage value；
Determine the regression straight line of n denoising average valueSlope a, wherein it is described return it is straightThe line span straight line minimum from the fore-and-aft distance sum of the n denoising average value；
Judge whether the slope a meets default second alarm trigger condition, second alarm is touchedClockwork spring part is：The slope a is more than or equal to the slope threshold value；
Alarm monitoring is carried out according to the judged result of the described second alarm trigger condition.
Further, the above method can also have the characteristics that, described to be triggered according to the described second alarmThe judged result of condition carries out alarm monitoring, including：
In the case where the slope a meets the second alarm trigger condition, triggering alarm.
Further, the above method can also have the characteristics that, according to the described second alarm trigger conditionJudged result carry out alarm monitoring, including：
In the case where the slope a does not meet the second alarm trigger condition, triggering is forbidden to alarm.
The monitoring method of the embodiment of the present invention, by the skew for reflecting doubtful noise set in a period of timeWith the irrelevance stdr of discrete case judge whether to need alarm, it is to avoid mistake caused by data noiseAlarm, so as to effectively reduce false alarm, improves the accuracy of alarm.
To achieve the above object, the invention also provides a kind of monitoring system, applied to cloud computing platform,Including：
Acquisition device, the monitoring data for obtaining the specified period；
Supervising device, for the monitoring data obtained based on the acquisition device, according to default alarmTrigger condition carries out alarm monitoring.
Further, said apparatus can also have the characteristics that, the supervising device includes：
Extraction module, for extracting sampling of the monitoring objective in period T from the monitoring dataData acquisition system, is designated as first set { xi, the termination time point of the period T is current time；
Screening module, for from the first set { xiIn filter out doubtful noise xj, it is described doubtfulNoise refers to the sampled data more than predetermined threshold value th, and doubtful noise set is designated as into second set { xj}；
Computing module, the irrelevance stdr for calculating doubtful noise in the period T,Wherein, std is the first set { xiStandard deviation,Std ' is the second set { xjRelative standard deviation,S is described theTwo set { xjIn data amount check,For the first set { xiAverage value,
First judge module, is touched for judging whether the irrelevance stdr meets default first alarmClockwork spring part, it is described first alarm trigger condition be：The irrelevance stdr is more than or equal to default inclinedFrom degree threshold value；
First monitoring module, for being alarmed according to the judged result of the described first alarm trigger conditionMonitoring.
Further, said apparatus can also have the characteristics that, first monitoring module includes：
First trigger element, for meeting the first alarm trigger condition in the irrelevance stdrIn the case of, triggering alarm.
Further, said apparatus can also have the characteristics that, the supervising device includes：
Division module, for being that the irrelevance stdr is less than the irrelevance threshold in the comparative resultIn the case of value, the period T is divided into n period Ti, n, i are natural number, and n≥i≥0；
Computing module, for calculating period T respectivelyiInterior sampled data removes flat after doubtful noiseAverage, is designated as denoising average value；
Determining module, the regression straight line for determining n denoising average valueSlope a,Wherein described regression straight line span is straight from the fore-and-aft distance sum minimum of the n denoising average valueLine；
Second judge module, for judging whether the slope a meets default second alarm triggering barPart, it is described second alarm trigger condition be：The slope a is more than or equal to the slope threshold value；
Second monitoring module, for being alarmed according to the judged result of the described second alarm trigger conditionMonitoring.
Further, said apparatus can also have the characteristics that, second monitoring module includes：
Second trigger element, the situation for meeting the second alarm trigger condition in the slope aUnder, triggering alarm.
Further, said apparatus can also have the characteristics that, second monitoring module includes：
Forbid unit, in the case of not meeting the second alarm trigger condition in the slope a,Triggering is forbidden to alarm.
The monitoring system of the embodiment of the present invention, by the skew for reflecting doubtful noise set in a period of timeWith the irrelevance stdr of discrete case judge whether to need alarm, it is to avoid mistake caused by data noiseAlarm, so as to effectively reduce false alarm, improves the accuracy of alarm.
To achieve the above object, the invention also provides a kind of monitoring device, including foregoing any one instituteThe monitoring system stated.
The monitoring device of the embodiment of the present invention includes monitoring system, doubtful in a period of time by reflectingThe skew of noise set and the irrelevance stdr of discrete case judge whether to need alarm, it is to avoid dataFalse alarm caused by noise, so as to effectively reduce false alarm, improves the accuracy of alarm.
Embodiment
The principle and feature of the present invention are described below in conjunction with accompanying drawing, illustrated embodiment is served only for solutionThe present invention is released, the scope of the present invention is not intended to limit.For those of ordinary skill in the art,On the premise of not paying creative work, all embodiments obtained according to spirit of the invention, allBelong to protection scope of the present invention.
Monitoring method and monitoring system in the embodiment of the present invention can be applied to cloud computing platform.Fig. 2For the flow chart of monitoring method in the embodiment of the present invention.As shown in Fig. 2 in the present embodiment, monitoring sideMethod may comprise steps of：
Step S201, obtains the monitoring data for specifying the period；
Here, monitoring data is the monitoring data in a period, rather than some time pointMonitoring data.What the monitoring data at some time point reflected is the data cases of moment, with very bigContingency, the alarm monitoring error rate that this allows for the monitoring data based on time point is very high.And oneWhat the monitoring data in the individual period reflected is the overall data situation of the period, can be more objectiveGround reflects the truth of data, such as development trend etc., so as to be entered according to overall data casesRow alarm analysis, and then determine the need for alarm.This way it is possible to avoid wrong caused by data noiseFalse alarm.Therefore, the alarm monitoring of the monitoring data based on the period can substantially reduce warning errorRate, so that monitoring alarm result is more accurate.
Wherein, the termination time point for specifying the period can be current time.The specified period can beFrom a period of time of current time forward.
Step S202, based on the monitoring data, alarm monitoring is carried out according to default alarm trigger condition.
The monitoring method of the embodiment of the present invention, is judged whether based on the monitoring data of certain time periodNeed alarm, it is to avoid false alarm caused by data noises, so that false alarm is effectively reduced,Improve the accuracy of alarm.
The monitoring method provided based on above-described embodiment, the embodiment of the present invention is to foregoing step S202It is specifically described.
A kind of flow chart that Fig. 3 is step S202 in the embodiment of the present invention.As shown in figure 3, this realityApply in example, step S202 can specifically include：Step S301, extracts monitoring from monitoring dataSampled data set of the target in period T, is designated as first set { xi, wherein, period TTermination time point be current time；
In embodiments of the present invention, monitoring is carried out in real time.Here, judge whether to need what is alarmedData basis be using current time as terminate time point period T in data, rather than someData on time point.Thus can be from monitoring data be macroscopically investigated, so as to be prevented effectively from factorAccording to false alarm caused by noise.
Step S302, filters out doubtful noise x from first setj, wherein, doubtful noise refers to greatlyIn predetermined threshold value th sampled data, doubtful noise set is designated as second set { xj}；
Step S303, calculates the irrelevance stdr of doubtful noise in period T,
Wherein, std is first set { xiStandard deviation,Std ' is describedSecond set { xjRelative standard deviation,S is second set { xjInData amount check,For first set { xiAverage value,
Fig. 9 is the interval data distribution schematic diagram of T time section.As shown in figure 9, have two it is intervalData value has exceeded threshold value th, then is doubtful noise in two interval sampled datas.Doubtful noiseThe collection of composition is combined into second set.
The more low then irrelevance stdr of frequency that noise occurs is smaller, and the frequency that noise occurs is higher, then partiallyIt is bigger from degree stdr.Skew and the discrete case of doubtful noise set are may determine that by irrelevance stdr,So as to ignore small probability, low-density doubtful noise.So, false alarm can be effectively reduced,Improve the accuracy of alarm.
Step S304, judges whether irrelevance stdr meets default first alarm trigger condition, wherein,First alarm trigger condition be：Irrelevance stdr is more than or equal to default irrelevance threshold value；
Step S305, alarm monitoring is carried out according to the judged result of the first alarm trigger condition.
That is, determining whether alarm according to irrelevance stdr and default irrelevance threshold value comparative result.
In embodiments of the present invention, alarm monitoring is carried out according to the judged result of the first alarm trigger conditionIt can include：In the case where irrelevance stdr meets the first alarm trigger condition, triggering alarm.
Irrelevance threshold value represent can received doubtful noise set skew and discrete case, beyond inclinedProgress is accomplished by from degree threshold value to alarm.
The monitoring method of the embodiment of the present invention, by the skew for reflecting doubtful noise set in a period of timeWith the irrelevance stdr of discrete case judge whether to need alarm, it is to avoid mistake caused by data noiseAlarm, so as to effectively reduce false alarm, improves the accuracy of alarm.
Another flow chart that Fig. 4 is step S202 in the embodiment of the present invention.As shown in figure 4, thisIn embodiment, step S202 can specifically include：
Step S401, extracts sampled data set of the monitoring objective in period T from monitoring dataClose, be designated as first set { xi, wherein, period T termination time point is current time；
Step S402, filters out doubtful noise x from first setj, wherein, doubtful noise refers to greatlyIn predetermined threshold value th sampled data, doubtful noise set is designated as second set { xj}；
Step S403, calculates the irrelevance stdr of doubtful noise in period T,
Wherein, std is first set { xiStandard deviation,Std ' is describedSecond set { xjRelative standard deviation,S is second set { xjInData amount check,For first set { xiAverage value,
Step S404, judges whether irrelevance stdr meets default first alarm trigger condition, ifIrrelevance stdr meets the first alarm trigger condition, then step S410 is performed, otherwise, if the deviation from degreeStdr does not meet the first alarm trigger condition, then performs step S405；
Step S405, n period T is divided into by period Ti, n, i are natural number, and n≥i≥0；
Period TiLength choose and can be set according to the interval size adaptations of overall time section T.
Step S406, calculates period T respectivelyiInterior sampled data removes being averaged after doubtful noiseValue, is designated as denoising average value；
The calculating process of denoising average value can be：First by period TiInterior doubtful noise is excluded, soAveraged afterwards to excluding remaining data after doubtful noise.
Step S407, determines the regression straight line of n denoising average valueSlope a, whereinThe minimum straight line of the fore-and-aft distance sum of from n denoising average value of regression straight line span；
Wherein, the schematic diagram of denoising average value and regression straight line is as shown in Figure 5 and Figure 6.Fig. 5 and figureThe bullet of regression straight line annex represents the denoising average value of each period in 6.
The regression straight line of denoising average value reflects the average growth trend after denoising, can reflect prisonSurvey target data macro development trend, for judge whether need alarm provide it is more structurally sound according toAccording to.
Step S408, judges whether slope a meets default second alarm trigger condition, wherein, theTwo alarm trigger conditions be：Slope a is more than or equal to slope threshold value, if slope a meets defaultSecond alarm trigger condition, then perform step S410, otherwise, if slope a does not meet default theTwo alarm trigger conditions, then perform step S409；
If the slope a of regression straight line is approximately 0 (closely 0 or equal to 0), illustrate macroscopic viewOn see, the data value of monitoring objective is generally stable, it is not necessary to alarmed.If regression straight line is obliqueRate a is more than 0, from the point of view of illustrating macroscopically, and the data value of monitoring objective is generally elevated, it is necessary to reportIt is alert.
Wherein, slope threshold value can be one be approximately 0 very small value.Can be according to monitoringIt is specific to need to set slope threshold value.
Step S409, forbids triggering to alarm, and terminates；
Step S410, triggering alarm, terminates.
The monitoring method of the embodiment of the present invention, by the skew for reflecting doubtful noise set in a period of timeWith the irrelevance stdr of discrete case judge whether to need alarm, it is to avoid mistake caused by data noiseAlarm, so as to effectively reduce false alarm, improves the accuracy of alarm.Also, it is of the invention realThe monitoring method of example is applied, when irrelevance stdr is less than irrelevance threshold value, further by goingAverage growth trend after making an uproar judges whether to need alarm, so that more intelligent, more accurately triggering alarm.
In order to implement the monitoring method in the embodiment of the present invention, the invention also provides following monitoring systemSystem embodiment.Above to monitoring method embodiment of the present invention explanation be applied to the present invention below prisonControl system embodiment.
Fig. 7 is the structured flowchart of monitoring system in the embodiment of the present invention.As shown in fig. 7, the present embodimentIn, monitoring system 700 can include acquisition device 710 and supervising device 720.
Wherein, acquisition device 710 is used for the monitoring data for obtaining the specified period.Supervising device 720For the monitoring data obtained based on acquisition device 710, reported according to default alarm trigger conditionPolice commissioner is controlled.
In embodiments of the present invention, supervising device 720 can include extraction module, screening module, meterCalculate module, the first judge module and the first monitoring module.Extraction module, screening module, computing module,First judge module and the first monitoring module can be sequentially connected.
Wherein, extraction module is used to obtain sampled data set of the monitoring objective in period T, remembersFor first set { xi, period T termination time point is current time.
Wherein, screening module is used for from first set { xiIn filter out doubtful noise xj, it is described doubtfulNoise refers to the sampled data more than predetermined threshold value th, and doubtful noise set is designated as into second set { xj}。
Wherein, computing module is used for the irrelevance stdr for calculating doubtful noise in period T,Wherein, std is first set { xiStandard deviation,Std ' isSecond set { the xjRelative standard deviation,S is second set { xj}In data amount check,For first set { xiAverage value,
Wherein, the first judge module is used to judge whether irrelevance stdr meets default first alarm and touchClockwork spring part, wherein, the first alarm trigger condition is：Irrelevance stdr is more than or equal to default deviateSpend threshold value.
Wherein, the first monitoring module is used to be alarmed according to the judged result of the first alarm trigger conditionMonitoring.
In embodiments of the present invention, the first monitoring module can include the first trigger element.First triggeringUnit is used in the case where irrelevance stdr meets the first alarm trigger condition, triggering alarm.
In embodiments of the present invention supervising device 720 can also include division module, computing module, reallyOrder member, the second judge module and the second monitoring module.Wherein, division module is used in comparative resultIn the case of being less than irrelevance threshold value for irrelevance stdr, period T is divided into n periodTi, n, i are natural number, and n >=i >=0.Computing module is used to calculate period T respectivelyiInterior adoptsSample data remove the average value after doubtful noise, are designated as denoising average value.Determining module is used to determine nThe regression straight line of individual denoising average valueSlope a, wherein regression straight line span is from the nThe minimum straight line of the fore-and-aft distance sum of individual denoising average value.Second judge module is used to judge slope aWhether default second alarm trigger condition is met, and the second alarm trigger condition is：Slope a be more than orEqual to the slope threshold value.Second monitoring module is used for the judged result according to the second alarm trigger conditionCarry out alarm monitoring.
In embodiments of the present invention, the second monitoring module can include the second trigger element.Second triggeringUnit is used in the case where slope a meets the second alarm trigger condition, triggering alarm.
In embodiments of the present invention, the second monitoring module can include forbidding unit.Unit is forbidden to be used forIn the case where slope a does not meet the second alarm trigger condition, triggering is forbidden to alarm.
The monitoring system of the embodiment of the present invention, by the skew for reflecting doubtful noise set in a period of timeWith the irrelevance stdr of discrete case judge whether to need alarm, it is to avoid mistake caused by data noiseAlarm, so as to effectively reduce false alarm, improves the accuracy of alarm.Also, it is of the invention realThe monitoring system of example is applied, when irrelevance stdr is less than irrelevance threshold value, moreover it is possible to further lead toThe average growth trend crossed after denoising judges whether to need alarm, thus it is more intelligent, more accurately triggerAlarm.
Fig. 8 is the structured flowchart of monitoring device in the embodiment of the present invention.As shown in figure 8, the present embodimentIn, monitoring device 800 can include monitoring system 700.Wherein, monitoring system 700 can be thisAny of invention previous embodiment monitoring system.
Wherein, monitoring system 700 can be used for the monitoring data for obtaining the specified period；And, baseIn the monitoring data of acquisition, alarm monitoring is carried out according to default alarm trigger condition.
The monitoring device of the embodiment of the present invention includes monitoring system, doubtful in a period of time by reflectingThe skew of noise set and the irrelevance stdr of discrete case judge whether to need alarm, it is to avoid dataFalse alarm caused by noise, so as to effectively reduce false alarm, improves the accuracy of alarm.
Presently preferred embodiments of the present invention is the foregoing is only, is not intended to limit the invention, it is all in this hairWithin bright spirit and principle, any modification, equivalent substitution and improvements made etc. should be included inWithin protection scope of the present invention.