CN107491830A - A kind for the treatment of method and apparatus of time-serial position - Google Patents
A kind for the treatment of method and apparatus of time-serial position Download PDFInfo
- Publication number
- CN107491830A CN107491830A CN201710534624.5A CN201710534624A CN107491830A CN 107491830 A CN107491830 A CN 107491830A CN 201710534624 A CN201710534624 A CN 201710534624A CN 107491830 A CN107491830 A CN 107491830A
- Authority
- CN
- China
- Prior art keywords
- time
- amplitude
- time point
- point
- serial position
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0277—Online advertisement
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Complex Calculations (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a kind for the treatment of method and apparatus of time-serial position, this method includes：The amplitude difference sequence of time-serial position is determined, the amplitude difference sequence includes multiple amplitude differences；The amplitude difference in the amplitude difference sequence is filtered according to preparatory condition；According to multiple amplitude differences after filtering, the amplitude peak and/or amplitude valley of the time-serial position are determined；The time point of amplitude peak and/or amplitude valley is determined, is designated as object time point；According in the number summation of the object time point and following scheduled time section apart from the predicted time point that today is nearest, the time-serial position is blocked using different modes of blocking, wherein, the data that the time-serial position after blocking is used to be pointed in the following scheduled time section on each time point are predicted.The present invention can lift data prediction effect.
Description
Technical field
The present invention relates to technical field of data prediction, more particularly to the processing method and dress of a kind of time-serial position
Put.
Background technology
Curve matching is the common method in time series analysis, and traditional curve-fitting method is applied to relatively smoothly bent
Line, it ignores the section of some precipitous changes that may be present in curve, for example, the shadow due to data acquisition or accident
Ring caused by (for example popular video is reached the standard grade, TV play end etc.) and jumbo shake be present in the history tendency of time-serial position
Swing section.
So, conventional curvature approximating method of the prior art in face of with it is jumbo concussion section curve when, simply
Simply several exceptional values in curve are removed, then the time-serial position for eliminating exceptional value is carried out it is following certain
Data prediction on a little time points, will so produce larger data prediction deviation, cause fitting effect bad, and then influence
The accuracy of prediction.
The content of the invention
The invention provides a kind for the treatment of method and apparatus of time-serial position, to solve in the prior art to time sequence
The problem of data prediction accuracy difference of row curve.
In order to solve the above problems, according to an aspect of the present invention, the invention discloses a kind of time-serial position
Processing method, including：
The amplitude difference sequence of time-serial position is determined, the amplitude difference sequence includes multiple amplitude differences；
The amplitude difference in the amplitude difference sequence is filtered according to preparatory condition；
According to multiple amplitude differences after filtering, the amplitude peak and/or amplitude valley of the time-serial position are determined；
The time point of amplitude peak and/or amplitude valley is determined, is designated as object time point；
According in the number summation of the object time point and following scheduled time section apart from the prediction that today is nearest
At time point, the time-serial position is blocked using different modes of blocking, wherein, the time series after blocking
The data that curve is used to be pointed in the following scheduled time section on each time point are predicted.
Alternatively, the amplitude difference sequence for determining time-serial position, including：
Each point in travel time sequence curve successively, determines at least two extreme values in the time-serial position
Point；
Successively by the amplitude of the extreme point of time rearward in each two extreme point adjacent on time coordinate and when
Between the amplitude of a forward extreme point make difference operation, obtain the amplitude difference sequence being made up of multiple amplitude differences.
Alternatively, multiple amplitude differences according to after filtering, determine the time-serial position amplitude peak and/
Or amplitude valley, including：
In multiple amplitude differences after filtration, superposition meter is made to jack per line and at least two temporally adjacent amplitude differences
Calculate, obtain amplitude peak and/or amplitude valley.
Alternatively, the time point for determining amplitude peak and/or amplitude valley, object time point is designated as, including：
In multiple amplitude differences after filtration, by the extreme point of time corresponding to each amplitude difference rearward when
Between point be defined as time point of each amplitude difference；
The object time point of the target amplitude difference of time point rearward at least two amplitude difference of superposition is remembered
Record the time point for the amplitude peak and/or the amplitude valley.
Alternatively, it is described according in the number summation of the object time point and following scheduled time section apart from today
Nearest predicted time point, using different the step of mode are blocked to the time-serial position of blocking, including：
If the number summation of the object time point is even number, the predicted time point and the object time point of record are calculated
Time interval between the time point of middle time the latest；
If the time interval is more than or equal to preset time threshold, the note will be located in the time-serial position
In the object time point of record on the time point of time the latest and curve before is blocked；
If the time interval is less than the preset time threshold, outlier processing is made to the time-serial position.
Alternatively, it is described according in the number summation of the object time point and following scheduled time section apart from today
Nearest predicted time point, using different the step of mode are blocked to the time-serial position of blocking, in addition to：
If the number summation of the object time point is odd number, the record will be located in the time-serial position
In object time point on a time point in evening time second and curve before is blocked；
Outlier processing is made to the time-serial position after blocking.
According to another aspect of the present invention, the invention also discloses a kind of processing unit of time-serial position, including：
First determining module, for determining the amplitude difference sequence of time-serial position, the amplitude difference sequence includes
Multiple amplitude differences；
Filtering module, for being filtered according to preparatory condition to the amplitude difference in the amplitude difference sequence；
Second determining module, for according to multiple amplitude differences after filtering, determining the amplitude of the time-serial position
Peak value and/or amplitude valley；
3rd determining module, for determining the time point of amplitude peak and/or amplitude valley, it is designated as object time point；
Truncation module, it is modern for distance in the number summation according to the object time point and following scheduled time section
Its nearest predicted time point, the time-serial position is blocked using different modes of blocking, wherein, after blocking
The data that the time-serial position is used to be pointed in the following scheduled time section on each time point are predicted.
Alternatively, first determining module includes：
First determination sub-module, for each point in travel time sequence curve successively, determine that the time series is bent
At least two extreme points in line；
Computing submodule, for a pole by the time in adjacent each two extreme point on time coordinate rearward successively
The amplitude of the amplitude of a value point extreme point forward with the time makees difference operation, obtains the difference of vibration being made up of multiple amplitude differences
Value sequence.
Alternatively, second determining module includes：
Superposition calculation submodule, in multiple amplitude differences after filtration, to jack per line and temporally adjacent at least two
Individual amplitude difference makees superposition calculation, obtains amplitude peak and/or amplitude valley.
Alternatively, the 3rd determining module includes：
Second determination sub-module, in multiple amplitude differences after filtration, by the time corresponding to each amplitude difference
The time point of an extreme point rearward is defined as the time point of each amplitude difference；
Record sub module, for the target amplitude difference of time point rearward at least two amplitude difference by superposition
Object time point be recorded as time point of the amplitude peak and/or the amplitude valley.
Alternatively, the truncation module includes：
Calculating sub module, if the number summation for the object time point is even number, calculate the predicted time point with
Time interval in the object time point of record between the time point of time the latest；
First blocks submodule, if being more than or equal to preset time threshold for the time interval, by the time
It is located in sequence curve in the object time point of the record on the time point of time the latest and curve before enters
Row blocks；
First abnormality processing submodule, if being less than the preset time threshold for the time interval, to it is described when
Between sequence curve make outlier processing.
Alternatively, the truncation module also includes：
Second blocks submodule, if the number summation for the object time point is odd number, by the time series
Curve in curve on a time point in evening time second in the object time point of the record and before is carried out
Block；
Second abnormality processing submodule, for making outlier processing to the time-serial position after blocking.
Compared with prior art, the present invention includes advantages below：
The present invention by time-serial position have violent amplitude object time point record, then further according to
Apart from the predicted time that today is nearest in the number summation of object time point with violent amplitude and following scheduled time section
The concrete condition of point come use it is different block mode to be blocked to the abnormal section of time-serial position, so, adopting
Data are carried out to being ensured that in the following scheduled time section during each time point prediction with the time-serial position after blocking
The degree of accuracy of the data of prediction, lift data prediction effect.
Brief description of the drawings
Fig. 1 is a kind of step flow chart of the processing method embodiment of time-serial position of the present invention；
Fig. 2 is the step flow chart of the processing method embodiment of another time-serial position of the present invention；
Fig. 3 is a kind of structured flowchart of the processing unit embodiment of time-serial position of the present invention.
Embodiment
In order to facilitate the understanding of the purposes, features and advantages of the present invention, it is below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is further detailed explanation.
Reference picture 1, a kind of step flow chart of the processing method embodiment of time-serial position of the present invention is shown,
Before the step of illustrating the embodiment of the present invention, first it is explained as follows：
The transverse axis of so-called time-serial position, i.e. curve is using the time as coordinate, and the coordinate of the longitudinal axis then can be according to reality
The type (such as advertisement stock) of data being predicted is needed in application scenarios, by historical time (such as apart from today it
The data of daily advertisement stock in first 300 days) in the concrete numerical values of these data formed on coordinate, during so as to history of forming
In certain categorical data time-serial position, each point on curve represents the data number having at time point corresponding to the point
Value, such as advertisement stock's numerical value.Wherein, so-called advertisement stock, i.e. some video website is at some time point (such as one day)
Advertisement playback volume.
So in order to using the time-serial position of historical time come to the ordinate in today or certain following a period of time
Data be predicted, in order to ensure prediction the degree of accuracy, the embodiment of the present invention can be by time-serial position in discriminating
The end points in abnormal section to carry out time-serial position the truncation of accommodation mode, so as to using after blocking abnormal section
Time-serial position carries out data prediction, ensures the degree of accuracy of prediction data.Below by step 101~step 104 to exception
The discriminating of the end points in section, which is made, to be illustrated：
Step 101, the amplitude difference sequence of time-serial position is determined, the amplitude difference sequence includes multiple difference of vibration
Value；
Wherein, amplitude difference is the difference between two amplitudes in time-serial position.
Step 102, the amplitude difference in the amplitude difference sequence is filtered according to preparatory condition；
For example, the amplitude difference that the preparatory condition can represent for amplitude difference to be less than threshold value filters, and certainly, the default bar
Part can also be the condition that can arbitrarily filter the amplitude difference for influenceing data prediction.
Step 103, according to multiple amplitude differences after filtering, determine the time-serial position amplitude peak and/or
Amplitude valley；
Wherein, amplitude peak, amplitude valley are selected from the stack result of amplitude difference.
Step 104, the time point of amplitude peak and/or amplitude valley is determined, is designated as object time point；
Wherein, the object time point is amplitude peak, amplitude valley corresponding time point in time-serial position.
Wherein, what amplitude peak, the time point of amplitude valley reflected is the time end points that violent amplitude terminates, can be by this
A little time end points (i.e. above-mentioned object time point) are recorded.
Step 105, according in the number summation of the object time point and following scheduled time section apart from today most
Near predicted time point, the time-serial position is blocked using different modes of blocking, wherein, it is described after blocking
The data that time-serial position is used to be pointed in the following scheduled time section on each time point are predicted.
Wherein, after the time point that violent amplitude terminates in have found time-serial position, it is possible to according to institute in curve
The number summation at the time point for the violent amplitude having and the currently time to be predicted (such as today or following a certain section
Time) in apart from nearest time point (i.e. predicted time point) today, different block mode come time-serial position to use
Truncation is carried out, then, using the time-serial position after blocking come to data (the i.e. number of ordinate in predicted time point
Value) it is predicted.
By means of the technical scheme of the above embodiment of the present invention, the present invention is by determining have acutely in time-serial position
The object time point of amplitude, then in the number summation further according to the time point with violent amplitude and following scheduled time section
The concrete condition of the predicted time point nearest apart from today different blocks mode come the exception to time-serial position to use
Section is blocked, and so, data is being carried out to each in following scheduled time section using the time-serial position after blocking
The degree of accuracy of the data of prediction is ensured that during time point prediction, lifts data prediction effect.
Reference picture 2, the step flow chart of the processing method embodiment of another time-serial position of the present invention is shown,
Specifically comprise the following steps：
Step 201, each point in travel time sequence curve successively, determines at least two in the time-serial position
Individual extreme point；
Wherein it is possible to each point in time-serial position is traveled through, to determine the extreme point in curve, wherein,
Here curve at least two extreme points.So, extreme point all in curve can be found.
Step 202, an extreme point by the time in adjacent each two extreme point on time coordinate rearward successively
The amplitude of an amplitude extreme point forward with the time makees difference operation, obtains the amplitude difference sequence being made up of multiple amplitude differences
Row；
After all extreme points are so have found in curve, it is possible to by the number of the ordinate of two adjacent extreme points
Value (i.e. amplitude) makees difference operation, wherein, in order to find the variation tendency between these extreme points, it is necessary to extreme value by the time rearward
For the amplitude of point as minuend, the amplitude of time forward extreme point is used as subtrahend, it is possible to obtains the adjacent extreme value of each two
The amplitude difference of point.The amplitude difference describes the change severe degree between two adjacent extreme points, so can be obtained by
Multiple amplitude differences, so as to form amplitude difference sequence.
Step 203, the absolute value of amplitude difference in the amplitude difference sequence is less than to the difference of vibration of default amplitude threshold
Value filters out；
So in order to find the section having in history in time-serial position compared with acute variation, the area of small change is avoided
Between to the influence for the end points searched, need exist for going the fainter amplitude difference of the intensity of variation between extreme point in curve
Fall, i.e. the absolute value of amplitude difference in the above-mentioned amplitude difference sequence being calculated is less than to the difference of vibration of default amplitude threshold
Value filters out.
Step 204, in multiple amplitude differences after filtration, jack per line and at least two temporally adjacent amplitude differences are made
Superposition calculation, obtain amplitude peak and/or amplitude valley；
Wherein, after being filtered to amplitude difference, it is possible to be overlapped calculating to remaining amplitude difference, have to find
The point of violent amplitude.Wherein, when being overlapped, still follow temporally adjacent principle and come at least two adjacent jack per lines
Amplitude difference makees superposition, such as has 5 amplitude differences according to time vertical order sequence after filtration, respectively
For B1, B2, B3 ,-B4 ,-B5, then be exactly here that amplitude difference B1, amplitude difference B2 and amplitude difference B3 are entered in superposition
Row superposition, find that the amplitude difference-B4 of contrary sign then stops being superimposed after amplitude difference B3 is superimposed to, and-the B4 of jack per line ,-B5
Stacked system similarly, wherein, amplitude difference and (B1+B2+B3) be amplitude peak, and amplitude difference and-(B4+B5) then
For amplitude valley.
Step 205, in multiple amplitude differences after filtration, by the pole of time corresponding to each amplitude difference rearward
The time point of value point is defined as the time point of each amplitude difference；
Specifically, such as multiple amplitude differences after filtering are formed by following amplitude：Time point t1 extreme point is corresponding
Amplitude A 1, time point t2 extreme point corresponding amplitude A2, time point t3 extreme point corresponding amplitude A3, time point t4 extreme point
Corresponding amplitude A4, t1<t2<t3<T4, amplitude difference B1=A2-A1, amplitude difference B2=A3-A2, amplitude difference B3=A4-A3.
Wherein, by taking amplitude difference B1 as an example, it has corresponded to time point t1 extreme point and time point t2 extreme point, then here can be with
The time point t2 of the extreme point of time point rearward is defined as to amplitude difference B1 time point, similarly, amplitude difference B2 when
Between point be defined as t3, amplitude difference B3 time point is defined as t4.
The time point of so-called amplitude difference, it can be understood as the time point that every section of amplitude terminates.
Step 206, by the target of the target amplitude difference of time point rearward at least two amplitude difference of superposition
Time point is recorded as the time point of the amplitude peak and/or the amplitude valley；
Here, the embodiment of the present invention just have found violent amplitude (such as the amplitude obtained above in time-serial position
Difference and (B1+B2+B3) ,-(B4+B5), namely amplitude peak, amplitude valley, and it is determined that amplitude peak, amplitude valley
During time point, the target amplitude difference B3 of time point in amplitude difference B1, B2, B3 rearward time point t4 can be recorded as this
The time point of amplitude peak (B1+B2+B3)；And by the target amplitude difference of time point in amplitude difference-B4 ,-B5 rearward for example-
B5 time point is recorded as the amplitude valley-(B4+B5) time point.Wherein, the time point reflection of amplitude peak, amplitude valley
Be time end points that violent amplitude terminates, these time end points (i.e. above-mentioned object time point) can be recorded.
Step 207, judge the number summation of the object time point for odd number or even number；
If the number summation of the object time point is even number, step 208, the predicted time point and record are calculated
Time interval in object time point between the time point of time the latest；
Wherein, the number summation of the object time point represents for even number：Before above-mentioned predicted time point, the play on curve
The number of strong amplitude is even number.The time point of time the latest i.e. near the time point of the predicted time point, such as
210th day.
Step 209, judge whether the time interval is more than or equal to preset time threshold；
If the time interval is more than or equal to preset time threshold, step 210, then by the time-serial position
Curve on the time point of time the latest in the object time point of the record and before is blocked；
If the time interval is more than or equal to preset time threshold (such as 70 days), illustrate with violent amplitude when
Between point apart from current predictive time point time interval farther out, then it is assumed that from range prediction time point it is nearest there is violent amplitude
Time point start after time point, curve has been enter into plateau, will not be because of being predicted current predicted time point
Data impact, that is, the data to be predicted on current predictive time point will not be influenceed by the violent amplitude, therefore,
Can will in the time-serial position in the object time point of the record on the time point of time the latest with
And curve before is (before the time point of the curve point at a time point i.e. positioned at the time the latest and time the latest
The curve at time point) blocked (that is, the curve progress by time coordinate in curve before the 210th day and the 210th day
Block, wherein, when carrying out data prediction, it is possible to using the stable data collection blocked in rear remaining curve come to currently needing
The data of the following a period of time to be predicted are predicted).
If the time interval is less than the preset time threshold, step 211 is performed, the time-serial position is made
Outlier processing.
Wherein, if the time interval is less than the preset time threshold, illustrate the time point with violent amplitude away from
Nearer from predicted time point time interval, the data that the violent amplitude at the time point can be predicted predicted time point cause shadow
Ring, it is thus impossible to there is the curve of violent amplitude to block this, but outlier processing is made to the time-serial position
(wherein, outlier processing flow will not be repeated here to carry out the flow of outlier processing to curve in the prior art).
If the number summation of the object time point is odd number, step 212, institute will be located in the time-serial position
State on a time point in evening time in the object time point of record second and curve before is blocked；
If the number summation of the object time point is odd number, that is, before above-mentioned predicted time point, the play on curve
The number of strong amplitude is odd number, that is, includes not paired a crest or trough, then is needed the time series is bent
Curve point in line before second curve and time point on the time point of the predicted time point is blocked.
Step 213, outlier processing is made to the time-serial position after blocking.
Wherein, because any one curve may all have exceptional value, then here in order to ensure the point in curve is
Normally, outlier processing flow of the prior art can be carried out；And if the detection of artificial or machine is without in discovery curve
Exceptional value, then it can omit the outlier processing flow in above-described embodiment.This method equally the application protection domain it
It is interior.
By means of the above-mentioned technical proposal of the embodiment of the present invention, the present invention to having in time-serial position by acutely shaking
The object time point of width is recorded, and is then made a reservation for further according to the number summation of the object time point with violent amplitude with future
Used apart from the concrete condition of nearest predicted time point today in time interval and different block mode come to time series
The abnormal section of curve is blocked, and so, data is being carried out to the following scheduled time using the time-serial position after blocking
The degree of accuracy of the data of prediction is ensured that in section during each time point prediction, lifts data prediction effect.
Alternatively, before above-mentioned steps 201 are performed, the method for the embodiment of the present invention can also include：Calculated using smooth
Method makees smoothing processing to time-serial position.
Wherein, because time-serial position is not smooth curve, but the line of sawtooth pattern, then in order to find curve
In maximum and minimum, therefore, it is necessary first to use such as moving average, exponential smoothing
Make smoothing processing to time-serial position Deng smoothing algorithm, i.e., be modified the small sawtooth in curve, and formed smoothly
Curve, the trend of curve is detected beneficial in subsequent step, finds maximum and/or minimum.
So, by making smoothing processing to curve in advance so that the small sawtooth in curve is modified, and is formed smoothly
Curve, the trend of curve is detected beneficial in subsequent step, finds maximum and/or minimum；It is and default by setting one
Amplitude threshold, minor fluctuations in curve can be filtered out, avoid them from influenceing the search to violent amplitude section.
Wherein, in one embodiment, the default amplitude threshold can in the time-serial position institute a little
Have the 50% of the average value of amplitude.
So, by setting such a default amplitude threshold, minor fluctuations in curve can be filtered out, avoids it
Influence search to violent amplitude section.
It should be noted that for embodiment of the method, in order to be briefly described, therefore it is all expressed as to a series of action group
Close, but those skilled in the art should know, the embodiment of the present invention is not limited by described sequence of movement, because according to
According to the embodiment of the present invention, some steps can use other orders or carry out simultaneously.Secondly, those skilled in the art also should
Know, embodiment described in this description belongs to preferred embodiment, and the involved action not necessarily present invention is implemented
Necessary to example.
It is corresponding with the method that the embodiments of the present invention are provided, reference picture 3, show a kind of time series of the present invention
The structured flowchart of the processing unit embodiment of curve, it can specifically include following module：
First determining module 31, for determining the amplitude difference sequence of time-serial position, the amplitude difference sequence bag
Include multiple amplitude differences；
Filtering module 32, for being filtered according to preparatory condition to the amplitude difference in the amplitude difference sequence；
Second determining module 33, for according to multiple amplitude differences after filtering, determining shaking for the time-serial position
Width peak value and/or amplitude valley；
3rd determining module 34, for determining the time point of amplitude peak and/or amplitude valley, it is designated as object time point；
Truncation module 35, for distance in the number summation according to the object time point and following scheduled time section
Today, nearest predicted time point, was blocked using different modes of blocking to the time-serial position, wherein, after blocking
The time-serial position data that are used to be pointed in the following scheduled time section on each time point be predicted.
Alternatively, first determining module 31 includes：
First determination sub-module, for each point in travel time sequence curve successively, determine that the time series is bent
At least two extreme points in line；
Computing submodule, for a pole by the time in adjacent each two extreme point on time coordinate rearward successively
The amplitude of the amplitude of a value point extreme point forward with the time makees difference operation, obtains the difference of vibration being made up of multiple amplitude differences
Value sequence.
Alternatively, second determining module 33 includes：
Superposition calculation submodule, in multiple amplitude differences after filtration, to jack per line and temporally adjacent at least two
Individual amplitude difference makees superposition calculation, obtains amplitude peak and/or amplitude valley.
Alternatively, the 3rd determining module 34 includes：
Second determination sub-module, in multiple amplitude differences after filtration, by the time corresponding to each amplitude difference
The time point of an extreme point rearward is defined as the time point of each amplitude difference；
Record sub module, for the target amplitude difference of time point rearward at least two amplitude difference by superposition
Object time point be recorded as time point of the amplitude peak and/or the amplitude valley.
Alternatively, the truncation module 35 includes：
Calculating sub module, if the number summation for the object time point is even number, calculate predicted time point and record
Object time point in time interval between the time point of time the latest；
First blocks submodule, if being more than or equal to preset time threshold for the time interval, by the time
It is located in sequence curve in the object time point of the record on the time point of time the latest and curve before enters
Row blocks；
First abnormality processing submodule, if being less than the preset time threshold for the time interval, to it is described when
Between sequence curve make outlier processing.
Alternatively, the truncation module 35 also includes：
Second blocks submodule, if the number summation for the object time point is odd number, by the time series
Curve in curve on a time point in evening time second in the object time point of the record and before is carried out
Block；
Second abnormality processing submodule, for making outlier processing to the time-serial position after blocking.
Alternatively, described device also includes：
Smoothing module, for making smoothing processing to time-serial position using smoothing algorithm.
Preferably, the default amplitude threshold is 50% of the average value of all amplitudes in the time-serial position.
For device embodiment, because it is substantially similar to embodiment of the method, so description is fairly simple, it is related
Part illustrates referring to the part of embodiment of the method.
Each embodiment in this specification is described by the way of progressive, what each embodiment stressed be with
The difference of other embodiment, between each embodiment identical similar part mutually referring to.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can be provided as method, apparatus or calculate
Machine program product.Therefore, the embodiment of the present invention can use complete hardware embodiment, complete software embodiment or combine software and
The form of the embodiment of hardware aspect.Moreover, the embodiment of the present invention can use one or more wherein include computer can
With in the computer-usable storage medium (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form of the computer program product of implementation.
The embodiment of the present invention is with reference to method according to embodiments of the present invention, terminal device (system) and computer program
The flow chart and/or block diagram of product describes.It should be understood that can be by computer program instructions implementation process figure and/or block diagram
In each flow and/or square frame and the flow in flow chart and/or block diagram and/or the combination of square frame.These can be provided
Computer program instructions are set to all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing terminals
Standby processor is to produce a machine so that is held by the processor of computer or other programmable data processing terminal equipments
Capable instruction is produced for realizing in one flow of flow chart or multiple flows and/or one square frame of block diagram or multiple square frames
The device for the function of specifying.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing terminal equipments
In the computer-readable memory to work in a specific way so that the instruction being stored in the computer-readable memory produces bag
The manufacture of command device is included, the command device is realized in one flow of flow chart or multiple flows and/or one side of block diagram
The function of being specified in frame or multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing terminal equipments so that
Series of operation steps is performed on computer or other programmable terminal equipments to produce computer implemented processing, so that
The instruction performed on computer or other programmable terminal equipments is provided for realizing in one flow of flow chart or multiple flows
And/or specified in one square frame of block diagram or multiple square frames function the step of.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art once know base
This creative concept, then other change and modification can be made to these embodiments.So appended claims are intended to be construed to
Including preferred embodiment and fall into having altered and changing for range of embodiment of the invention.
Finally, it is to be noted that, herein, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between any this actual relation or order be present.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or terminal device including a series of elements are not only wrapped
Those key elements, but also the other element including being not expressly set out are included, or is also included for this process, method, article
Or the key element that terminal device is intrinsic.In the absence of more restrictions, wanted by what sentence "including a ..." limited
Element, it is not excluded that other identical element in the process including the key element, method, article or terminal device also be present.
Processing method to a kind of time-serial position provided by the present invention and a kind of place of time-serial position above
Device to be managed, is described in detail, specific case used herein is set forth to the principle and embodiment of the present invention,
The explanation of above example is only intended to help the method and its core concept for understanding the present invention；Meanwhile for the one of this area
As technical staff, according to the thought of the present invention, there will be changes in specific embodiments and applications, to sum up institute
State, this specification content should not be construed as limiting the invention.
Claims (12)
- A kind of 1. processing method of time-serial position, it is characterised in that including：The amplitude difference sequence of time-serial position is determined, the amplitude difference sequence includes multiple amplitude differences；The amplitude difference in the amplitude difference sequence is filtered according to preparatory condition；According to multiple amplitude differences after filtering, the amplitude peak and/or amplitude valley of the time-serial position are determined；The time point of amplitude peak and/or amplitude valley is determined, is designated as object time point；According in the number summation of the object time point and following scheduled time section apart from the predicted time that today is nearest Point, the time-serial position is blocked using different modes of blocking, wherein, the time-serial position after blocking It is predicted for being pointed to the data in the following scheduled time section on each time point.
- 2. according to the method for claim 1, it is characterised in that the amplitude difference sequence for determining time-serial position, Including：Each point in travel time sequence curve successively, determines at least two extreme points in the time-serial position；The amplitude of the extreme point of time rearward in adjacent each two extreme point on time coordinate is leaned on the time successively The amplitude of a preceding extreme point makees difference operation, obtains the amplitude difference sequence being made up of multiple amplitude differences.
- 3. according to the method for claim 1, it is characterised in that multiple amplitude differences according to after filtering, determine institute The amplitude peak and/or amplitude valley of time-serial position are stated, including：In multiple amplitude differences after filtration, superposition calculation is made to jack per line and at least two temporally adjacent amplitude differences, obtained To amplitude peak and/or amplitude valley.
- 4. according to the method for claim 1, it is characterised in that the time for determining amplitude peak and/or amplitude valley Point, object time point is designated as, including：In multiple amplitude differences after filtration, by the time point of the extreme point of time corresponding to each amplitude difference rearward It is defined as the time point of each amplitude difference；The object time point of the target amplitude difference of time point rearward at least two amplitude difference of superposition is recorded as The time point of the amplitude peak and/or the amplitude valley.
- 5. according to the method for claim 1, it is characterised in that the number summation according to the object time point and Apart from the predicted time point that today is nearest in following scheduled time section, mode is blocked to time series song using different The step of line is blocked, including：If the number summation of the object time point is even number, when calculating the predicted time point with the object time point of record Between time interval between time point the latest；If the time interval is more than or equal to preset time threshold, the record will be located in the time-serial position In object time point on the time point of time the latest and curve before is blocked；If the time interval is less than the preset time threshold, outlier processing is made to the time-serial position.
- 6. according to the method for claim 1, it is characterised in that the number summation according to the object time point and Apart from the predicted time point that today is nearest in following scheduled time section, mode is blocked to time series song using different The step of line is blocked, in addition to：If the number summation of the object time point is odd number, the target of the record will be located in the time-serial position In time point on a time point in evening time second and curve before is blocked；Outlier processing is made to the time-serial position after blocking.
- A kind of 7. processing unit of time-serial position, it is characterised in that including：First determining module, for determining the amplitude difference sequence of time-serial position, the amplitude difference sequence includes multiple Amplitude difference；Filtering module, for being filtered according to preparatory condition to the amplitude difference in the amplitude difference sequence；Second determining module, for according to multiple amplitude differences after filtering, determining the amplitude peak of the time-serial position And/or amplitude valley；3rd determining module, for determining the time point of amplitude peak and/or amplitude valley, it is designated as object time point；Truncation module, in the number summation according to the object time point and following scheduled time section apart from today most Near predicted time point, the time-serial position is blocked using different modes of blocking, wherein, it is described after blocking The data that time-serial position is used to be pointed in the following scheduled time section on each time point are predicted.
- 8. device according to claim 7, it is characterised in that first determining module includes：First determination sub-module, for each point in travel time sequence curve successively, determine in the time-serial position At least two extreme points；Computing submodule, for an extreme point by the time in adjacent each two extreme point on time coordinate rearward successively The amplitude of amplitude and time a forward extreme point make difference operation, obtain the amplitude difference sequence being made up of multiple amplitude differences Row.
- 9. device according to claim 7, it is characterised in that second determining module includes：Superposition calculation submodule, in multiple amplitude differences after filtration, being shaken to jack per line and temporally adjacent at least two Width difference makees superposition calculation, obtains amplitude peak and/or amplitude valley.
- 10. device according to claim 7, it is characterised in that the 3rd determining module includes：Second determination sub-module, in multiple amplitude differences after filtration, by the time corresponding to each amplitude difference rearward Time point of an extreme point be defined as time point of each amplitude difference；Record sub module, the mesh for the target amplitude difference of time point rearward at least two amplitude difference by superposition Mark time point is recorded as the time point of the amplitude peak and/or the amplitude valley.
- 11. device according to claim 7, it is characterised in that the truncation module includes：Calculating sub module, if the number summation for the object time point is even number, calculate the predicted time point and record Object time point in time interval between the time point of time the latest；First blocks submodule, if being more than or equal to preset time threshold for the time interval, by the time series It is located in curve in the object time point of the record on the time point of time the latest and curve before is cut It is disconnected；First abnormality processing submodule, if being less than the preset time threshold for the time interval, to the time sequence Row curve makees outlier processing.
- 12. device according to claim 7, it is characterised in that the truncation module also includes：Second blocks submodule, if the number summation for the object time point is odd number, by the time-serial position In on a time point in evening time second in the object time point of the record and curve before is blocked；Second abnormality processing submodule, for making outlier processing to the time-serial position after blocking.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710534624.5A CN107491830B (en) | 2017-07-03 | 2017-07-03 | Method and device for processing time series curve |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710534624.5A CN107491830B (en) | 2017-07-03 | 2017-07-03 | Method and device for processing time series curve |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107491830A true CN107491830A (en) | 2017-12-19 |
| CN107491830B CN107491830B (en) | 2021-03-26 |
Family
ID=60644516
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710534624.5A Active CN107491830B (en) | 2017-07-03 | 2017-07-03 | Method and device for processing time series curve |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107491830B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111937012A (en) * | 2018-03-30 | 2020-11-13 | 日本电气方案创新株式会社 | Index calculation device, prediction system, progress prediction evaluation method, and program |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010001966A1 (en) * | 2008-07-03 | 2010-01-07 | 日本電気株式会社 | Time-series data processing device and method and program thereof |
| CN102346745A (en) * | 2010-08-02 | 2012-02-08 | 阿里巴巴集团控股有限公司 | Method and device for predicting user behavior number for words |
| US20130116992A1 (en) * | 2011-11-08 | 2013-05-09 | International Business Machines Corporation | Time series data analysis method, system and computer program |
| CN103888315A (en) * | 2014-03-24 | 2014-06-25 | 北京邮电大学 | Self-adaptation burst flow detection device and detection method thereof |
| CN104462215A (en) * | 2014-11-05 | 2015-03-25 | 大连理工大学 | Scientific and technical literature quoting number predicting method based on time sequence |
| CN105303167A (en) * | 2008-01-23 | 2016-02-03 | 加州大学评议会 | Systems and methods for behavioral monitoring and calibration |
| CN105320845A (en) * | 2015-11-26 | 2016-02-10 | 电子科技大学 | Time sequence forecast method based on quantum gravity algorithm |
| US20160155334A1 (en) * | 2014-12-02 | 2016-06-02 | Robert Bosch Gmbh | Collision avoidance in traffic crossings using radar sensors |
-
2017
- 2017-07-03 CN CN201710534624.5A patent/CN107491830B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105303167A (en) * | 2008-01-23 | 2016-02-03 | 加州大学评议会 | Systems and methods for behavioral monitoring and calibration |
| WO2010001966A1 (en) * | 2008-07-03 | 2010-01-07 | 日本電気株式会社 | Time-series data processing device and method and program thereof |
| CN102346745A (en) * | 2010-08-02 | 2012-02-08 | 阿里巴巴集团控股有限公司 | Method and device for predicting user behavior number for words |
| US20130116992A1 (en) * | 2011-11-08 | 2013-05-09 | International Business Machines Corporation | Time series data analysis method, system and computer program |
| CN103888315A (en) * | 2014-03-24 | 2014-06-25 | 北京邮电大学 | Self-adaptation burst flow detection device and detection method thereof |
| CN104462215A (en) * | 2014-11-05 | 2015-03-25 | 大连理工大学 | Scientific and technical literature quoting number predicting method based on time sequence |
| US20160155334A1 (en) * | 2014-12-02 | 2016-06-02 | Robert Bosch Gmbh | Collision avoidance in traffic crossings using radar sensors |
| CN105320845A (en) * | 2015-11-26 | 2016-02-10 | 电子科技大学 | Time sequence forecast method based on quantum gravity algorithm |
Non-Patent Citations (5)
| Title |
|---|
| TAN, ZHIYI: "A Novel Time Series Approach for Predicting the Long-Term Popularity of Online Videos", 《IEEE TRANSACTIONS ON BROADCASTING》 * |
| ZHANG, JIE: "Collaborated Online Change-point Detection in Sparse Time Series for Online Advertising", 《IEEE INTERNATIONAL CONFERENCE ON DATA MINING》 * |
| 何永沛: "时间序列数据挖掘在股市预测分析中的应用研究", 《中国优秀硕士学位论文全文数据库（信息科技辑）》 * |
| 张俊池: "面向在线视频服务的播放量预测算法研究", 《中国优秀硕士学位论文全文数据库（信息科技辑）》 * |
| 朱开成: "视频分享网站热门视频快速挖掘预测模型", 《中国优秀硕士学位论文全文数据库（信息科技辑）》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111937012A (en) * | 2018-03-30 | 2020-11-13 | 日本电气方案创新株式会社 | Index calculation device, prediction system, progress prediction evaluation method, and program |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107491830B (en) | 2021-03-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104504120B (en) | A kind of similar pictures detection method and device | |
| CN108234961B (en) | Multi-path camera coding and video stream guiding method and system | |
| CN109862323A (en) | Playback method, device and the processing equipment of multi-channel video | |
| CN111160187B (en) | Method, device and system for detecting left-behind object | |
| CN113469075A (en) | Method, device and equipment for determining traffic flow index and storage medium | |
| CN107491830A (en) | A kind for the treatment of method and apparatus of time-serial position | |
| CN116957300A (en) | Logistics data management method and system based on unmanned warehouse shuttle | |
| CN109359760B (en) | Logistics path optimization method and device and server | |
| CN114627163A (en) | Global image target tracking method and system based on rapid scene splicing | |
| CN104484457A (en) | Method and system for extracting and searching moving object in parallel video | |
| CN116452616A (en) | Intersecting contour point co-edge method based on image segmentation | |
| CN106529528A (en) | Method and equipment for identifying effective moving target | |
| CN116753967A (en) | Virtual lane line generation method, device, equipment, medium and program product | |
| CN116067359A (en) | Low-precision track data processing method and system based on delaunay triangle network | |
| CN110211161B (en) | Method and device for multi-camera track coincidence analysis | |
| CN108229071A (en) | Cutting performance degradation assessment method and system based on AR models Yu SVDD algorithms | |
| CN104160629B (en) | Data compression device and data compression method | |
| CN104299206B (en) | Image processing method and image processing device | |
| CN111143589A (en) | Image processing method and device and storage medium | |
| CN113626711B (en) | Live video recommendation method and device for mobile phone bank | |
| CN114638712A (en) | Financial market form division method and device | |
| CN108182341B (en) | A kind of method and device of the data pattern of recognition time sequence | |
| CN110221202B (en) | Current curve processing method and device for working current of circuit breaker | |
| CN112597152B (en) | Indexing method and indexing device for characteristic time sequence data based on skip list | |
| CN111161548B (en) | Vehicle-mounted binocular camera and relative speed calculating method thereof |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |