CN107516114A - A kind of time Series Processing method and device - Google Patents
A kind of time Series Processing method and device Download PDFInfo
- Publication number
- CN107516114A CN107516114A CN201710753494.4A CN201710753494A CN107516114A CN 107516114 A CN107516114 A CN 107516114A CN 201710753494 A CN201710753494 A CN 201710753494A CN 107516114 A CN107516114 A CN 107516114A
- Authority
- CN
- China
- Prior art keywords
- target component
- dynamic similarity
- similarity degree
- time series
- parameter
- 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
- 238000003672 processing method Methods 0.000 title claims abstract description 13
- 238000000034 method Methods 0.000 claims abstract description 33
- 238000012545 processing Methods 0.000 claims description 30
- 238000013481 data capture Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 abstract description 18
- 238000000714 time series forecasting Methods 0.000 abstract description 6
- 230000015572 biosynthetic process Effects 0.000 abstract description 5
- 238000003786 synthesis reaction Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000009825 accumulation Methods 0.000 abstract description 2
- 238000007418 data mining Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 241000208340 Araliaceae Species 0.000 description 4
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 4
- 235000003140 Panax quinquefolius Nutrition 0.000 description 4
- 235000008434 ginseng Nutrition 0.000 description 4
- 230000002093 peripheral effect Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000004549 pulsed laser deposition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- General Engineering & Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of time Series Processing method and device provided in an embodiment of the present invention, data mining technology field.This method by first obtain in predetermined time period comprising parameter to be measured target component target component time series, the influence factor time series of the target component in the predetermined time period is obtained again, the target component time series and the influence factor time series are finally based on, the dynamic similarity sequence of the parameter to be measured is chosen from the M target component subsequence.So as to which the change procedure of target component and the change procedure of each influence factor of target component synthesis are included in the selection rule of Similar Time Series Based on Markov Chain, so that in the time series forecasting by accumulation effects, selection accuracy for Similar Time Series Based on Markov Chain is higher and also of a relatively high to the accuracy of time series forecasting, and then significantly reduces the prediction error to time series in the prior art.
Description
Technical field
The present invention relates to data mining technology field, in particular to a kind of time Series Processing method and device.
Background technology
In the existing research that time series forecasting is carried out using Similar Time Series Based on Markov Chain method, more by the change of target component
Rule is deduced and studied as the changing rule of time series, and thinks that this rule is changeless.It is but actual
On, this rule is time-varying.Thus, the still larger technical problem of prediction error be present in prior art.
The content of the invention
The present invention provides a kind of time Series Processing method and device, it is intended to improves above-mentioned technical problem.
A kind of time Series Processing method provided by the invention, including：Obtain in predetermined time period and include parameter to be measured
Target component target component time series, wherein, the target component time series includes M target component subsequence,
Wherein, M is the integer more than or equal to 2；Obtain the influence factor time series of the target component in the predetermined time period；
Based on the target component time series and the influence factor time series, chosen from the M target component subsequence
The dynamic similarity sequence of the parameter to be measured.
Preferably, it is described to be based on the target component time series and the influence factor time series, from the M mesh
The dynamic similarity sequence of parameter to be measured described in being chosen in parameter subsequence is marked, including：Obtain the M target component subsequence
Length of time series；The parameter subsequence to be measured of the parameter to be measured is obtained based on the length of time series；Obtain each
Multiple first dynamic similarity degree of the target component subsequence and the parameter subsequence to be measured；When obtaining the influence factor
Between sequence and the parameter subsequence to be measured multiple second dynamic similarity degree；Based on the multiple first dynamic similarity degree and institute
State the dynamic similarity sequence of parameter to be measured described in multiple second dynamic similarity retrievals.
Preferably, it is described based on described in the multiple first dynamic similarity degree and the acquisition of the multiple second dynamic similarity degree
The dynamic similarity sequence of parameter to be measured, including：Obtain the target component dynamic similarity degree corresponding to the first dynamic similarity degree
Influence factor dynamic similarity degree weight corresponding to weight and the second dynamic similarity degree, wherein, the target component dynamic
Similarity weight is 1 with the influence factor dynamic similarity degree weight sum；Obtain each first dynamic similarity degree and institute
State the first product of target component dynamic similarity degree weight；Obtain each two dynamic similarities degree and influence factor dynamic
Second product of similarity weight；The dynamic similarity of the parameter to be measured is obtained based on first product and second product
Sequence.
Preferably, the dynamic similarity sequence that the parameter to be measured is obtained based on first product and second product
Row, including：Obtain first product and second sum of products；First product and second sum of products are made
For comprehensive dynamic similarity degree；The sequence number of corresponding subsequence during numerical value maximum based on the comprehensive dynamic similarity degree；Will
Dynamic similarity sequence of the target component subsequence corresponding to the sequence number as the parameter to be measured.
Preferably, the comprehensive dynamic similarity degree meets：
Wherein, i=1,2 ..., r-1, β0For the target component dynamic similarity degree weight, βc(c=1,2 ..., C) is
The influence factor dynamic similarity degree weight.
A kind of time Series Processing device provided by the invention, including：First data capture unit, during for obtaining default
Between in length the target component comprising parameter to be measured target component time series, wherein, the target component time series bag
Containing M target component subsequence, wherein, M is the integer more than or equal to 2；Second data capture unit, it is described default for obtaining
The influence factor time series of the target component in time span；Data processing unit, during for based on the target component
Between sequence and the influence factor time series, the dynamic of the parameter to be measured is chosen from the M target component subsequence
Similar sequences.
Preferably, the data processing unit includes：First data acquisition subelement, for obtaining the M target ginseng
The length of time series of number subsequence；Second data acquisition subelement, for based on the length of time series obtain described in treat
Survey the parameter subsequence to be measured of parameter；3rd data acquisition subelement, for obtaining each target component subsequence and institute
State multiple first dynamic similarity degree of parameter subsequence to be measured；4th data acquisition subelement, for obtaining the influence factor
Time series and multiple second dynamic similarity degree of the parameter subsequence to be measured；5th data acquisition subelement, for based on
The multiple first dynamic similarity degree and the multiple second dynamic similarity degree obtain the dynamic similarity sequence of the parameter to be measured.
Preferably, the 5th data acquisition subelement includes：First data acquisition module, it is dynamic for obtaining described first
The influence factor corresponding to target component dynamic similarity degree weight and the second dynamic similarity degree corresponding to state similarity is moved
State similarity weight, wherein, the target component dynamic similarity degree weight and the influence factor dynamic similarity degree weight sum
For 1；Second data acquisition module, weighed for obtaining each first dynamic similarity degree with the target component dynamic similarity degree
First product of weight；3rd data acquisition module, moved for obtaining each second dynamic similarity degree with the influence factor
Second product of state similarity weight；Data processing module, for obtaining institute based on first product and second product
State the dynamic similarity sequence of parameter to be measured.
Preferably, the data processing module includes：First data acquisition submodule, for obtain first product with
Second sum of products；Data processing submodule, for using first product and second sum of products as synthesis
Dynamic similarity degree；Second data acquisition submodule, it is corresponding during for numerical value maximum based on the comprehensive dynamic similarity degree
The sequence number of subsequence；3rd data acquisition submodule, for using target component subsequence corresponding to the sequence number as institute
State the dynamic similarity sequence of parameter to be measured.
Preferably, the comprehensive dynamic similarity degree meets：
Wherein, i=1,2 ..., r-1, β0For the target component dynamic similarity degree weight, βc(c=1,2 ..., C) is
The influence factor dynamic similarity degree weight.
A kind of time Series Processing method and device that the invention described above provides, this method are grown by first obtaining preset time
The target component time series of target component comprising parameter to be measured in degree, then obtain the target in the predetermined time period
The influence factor time series of parameter, the target component time series and the influence factor time series are finally based on, from
The dynamic similarity sequence of the parameter to be measured is chosen in the M target component subsequence.So as to by the change of target component
The change procedure synthesis of each influence factor of journey and target component is included in the selection rule of Similar Time Series Based on Markov Chain so that is being involved
In the time series forecasting of product effects, the selection accuracy for Similar Time Series Based on Markov Chain is higher and to time series forecasting
Accuracy it is also of a relatively high, and then significantly reduce the prediction error to time series in the prior art.
Brief description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by embodiment it is required use it is attached
Figure is briefly described, it will be appreciated that the following drawings illustrate only certain embodiments of the present invention, therefore be not construed as pair
The restriction of scope, for those of ordinary skill in the art, on the premise of not paying creative work, can also be according to this
A little accompanying drawings obtain other related accompanying drawings.
Fig. 1 is the structured flowchart of electronic equipment provided in an embodiment of the present invention；
Fig. 2 is a kind of flow chart for time Series Processing method that first embodiment of the invention provides；
Fig. 3 is a kind of flow chart for time Series Processing method that second embodiment of the invention provides；
Fig. 4 is a kind of high-level schematic functional block diagram for time Series Processing device that third embodiment of the invention provides.
Embodiment
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
Part of the embodiment of the present invention, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.Therefore,
The detailed description of the embodiments of the invention to providing in the accompanying drawings is not intended to limit the model of claimed invention below
Enclose, but be merely representative of the selected embodiment of the present invention.Based on the embodiment in the present invention, those of ordinary skill in the art are not having
There is the every other embodiment made and obtained under the premise of creative work, belong to the scope of protection of the invention.
As shown in figure 1, the structured flowchart for a kind of electronic equipment provided in an embodiment of the present invention.The electronic equipment 300
Including time Series Processing device 400, memory 302, storage control 303, processor 304 and Peripheral Interface 305.
The memory 302, storage control 303, processor 304 and 305 each element of Peripheral Interface are direct between each other
Or be electrically connected with indirectly, to realize the transmission of data or interaction.For example, these elements can pass through one or more between each other
Communication bus or signal wire, which are realized, to be electrically connected with.The time Series Processing device 400 include it is at least one can be with software or solid
The form of part (firmware) is stored in the memory 302 or is solidificated in the operating system of the electronic equipment 300
Software function module in (operating system, OS).The processor 304 is used to perform what is stored in memory 302
Executable module, such as the software function module or computer program that the time Series Processing device 400 includes.
Wherein, memory 302 may be, but not limited to, random access memory (Random Access Memory,
RAM), read-only storage (Read Only Memory, ROM), programmable read only memory (Programmable Read-
Only Memory, PROM), erasable read-only memory (Erasable Programmable Read-Only Memory,
EPROM), electricallyerasable ROM (EEROM) (Electric Erasable Programmable Read-Only Memory,
EEPROM) etc..Wherein, memory 302 is used for storage program, and the processor 304 is after execute instruction is received, described in execution
Program, the method performed by server 100 that the stream process that foregoing any embodiment of the embodiment of the present invention discloses defines can answer
Realized in processor 304, or by processor 304.
Processor 304 is probably a kind of IC chip, has the disposal ability of signal.Above-mentioned processor 304 can
To be general processor, including central processing unit (Central Processing Unit, abbreviation CPU), network processing unit
(Network Processor, abbreviation NP) etc.；Can also be digital signal processor (DSP), application specific integrated circuit (ASIC),
Ready-made programmable gate array (FPGA) either other PLDs, discrete gate or transistor logic, discrete hard
Part component.It can realize or perform disclosed each method, step and the logic diagram in the embodiment of the present invention.General processor
Can be microprocessor or the processor can also be any conventional processor etc..
Various input/output devices are coupled to processor 304 and memory 302 by the Peripheral Interface 305.At some
In embodiment, Peripheral Interface 305, processor 304 and storage control 303 can be realized in one single chip.Other one
In a little examples, they can be realized by independent chip respectively.
Referring to Fig. 2, it is a kind of flow chart for time Series Processing method that first embodiment of the invention provides.Below will
Idiographic flow shown in Fig. 2 is described in detail.
Step S101, the target component time series of the target component comprising parameter to be measured in predetermined time period is obtained,
Wherein, the target component time series includes M target component subsequence, wherein, M is the integer more than or equal to 2.
Wherein, the selection of the predetermined time period can enter choose according to specific implementation process, for example, when described default
Between length can be one week or 3 days, can also be one month.Here, it is not especially limited.
As a kind of embodiment, the target ginseng of the target component of period and its preceding some periods where obtaining parameter to be measured
Number time series.
In the present embodiment, the target component subsequence can be obtained with passage time sliding window, according to set
Sliding window size can from target component time series obtain M target component subsequence.
In the present embodiment, the parameter to be measured refers to the parameter for being ready for detection.
Step S102, obtain the influence factor time series of the target component in the predetermined time period.
In the present embodiment, each target component corresponds to one or more influence factor time serieses.
The influence factor refers to the factor for influenceing target component, i.e., the factor that can be had an impact to target component.
For example, the target component is per day load, the influence factor can be max. daily temperature/mean daily temperature etc..
The influence factor time series refers to the influence factor in the predetermined time period.
Step S103, based on the target component time series and the influence factor time series, from the M target
The dynamic similarity sequence of the parameter to be measured is chosen in parameter subsequence.
The dynamic similarity sequence refers to the son with the parameter to be measured in the change procedure of the subsequence of target component
The maximum sequence of " comprehensive similarity " of sequence.It is rule searching pair by the change procedure of target component and each influence factor
As, from parameter temporal node to be measured before target component and each influence factor time series in search for and choose and parameter to be measured
Place target component and its all similar target component of each influence factor and the change of each influence factor period and its preceding some periods
Change process.Parameter time series to be measured are predicted with the dynamic effects that take into account cumulative effect, and then is significantly reduced
Error, so as to get the changing rule of target component time series more conform to actual conditions, when having existing for cumulative effect
Between in sequence prediction, its precision of prediction obtains further improving relatively.
In the present embodiment, by first determining the length of target component and each influence factor time subsequence, recycle dynamic
The similar thought of state determines target component dynamic similarity degree and each influence factor dynamic similarity degree description indexes respectively, then comprehensive again
Close and consider target component dynamic similarity degree and each influence factor dynamic similarity degree, it is determined that comprehensive dynamic similarity degree description indexes,
The maximum target component subsequence of comprehensive dynamic similarity degree is finally chosen, as target component subsequence dynamic where variable to be measured
Similar sequences.The foundation chosen as Similar Time Series Based on Markov Chain, embedding data excavate prediction algorithm, further realize time sequence
The prediction of row.
Referring to Fig. 3, it is a kind of flow chart for time Series Processing method that second embodiment of the invention provides.Below will
Idiographic flow shown in Fig. 3 is described in detail.
Step S201, the target component time series of the target component comprising parameter to be measured in predetermined time period is obtained,
Wherein, the target component time series includes M target component subsequence, wherein, M is the integer more than or equal to 2.
Step S202, obtain the influence factor time series of the target component in the predetermined time period.
Step S201 and step S202 embodiment refer to step corresponding in first embodiment, here,
Repeat no more.
Step S203, obtain the length of time series of the M target component subsequence.
As a kind of embodiment, passage time sliding window produces M target in the target component sequence successively
Parameter subsequence.Wherein, M value is the integer more than or equal to 2.And by set the size of the time slide window from
And obtain the length of time series of each target component subsequence.I.e. length of time series is the size of time slide window.Example
Such as, the size of the time slide window can be for the sliding window of t-1 (the general values 3,4,5 of t).For example, set target ginseng
The length of number time series is n.If the length of sliding time window be t-1 (the general values 3,4,5 of t) sliding window successively
Produce target component subsequence.R target component subsequence, r=n-t+1 can be produced altogether.Its i-th section of target component subsequence can
To be expressed as Li={ Li(1), Li..., L (2)i(t-1)}.Wherein, i=1,2 ..., r.
Step S204, the parameter subsequence to be measured of the parameter to be measured is obtained based on the length of time series.
In the present embodiment, the time span of the parameter subsequence to be measured and the time of the target component subsequence are grown
Spend equal.Therefore, by obtaining the time span of target component subsequence so as to getting the parameter subsequence to be measured.
Step S205, obtain multiple first dynamics of each target component subsequence and the parameter subsequence to be measured
Similarity.
In the present embodiment, all it is t-1 target component subsequence L provided with time spani={ Li(1), Li(2) ...,
Li} and parameter subsequence L to be measured (t-1)r={ Lr(1), Lr..., L (2)r(t-1) }, i=1,2 ..., r-1.Then target component
Dynamic similarity degree describes the similarity degree of t-1 target component changing rule before the two subsequences.For example, use dL.i(i
=1,2 ..., r-1) " relative change rate's square distance and " is represented, it can be described as LiAnd LrThe rate of change of middle target component
Difference degree.Calculation formula is：
In formula, LiAnd L (u)r(u) it is respectively target component subsequence LiWith parameter subsequence L to be measuredrIn u (u=1,
2 ..., t-1) individual target component.The first dynamic similarity degree SL.iMeet：
In formula, max (dL.i) and min (dL.i) it is respectively dL.iMaximum and minimum value, i=1,2 ..., r-1.
Step S206, obtain the influence factor time series and multiple second dynamic phases of the parameter subsequence to be measured
Like degree.
As a kind of embodiment, obtain in each influence factor subsequence and the parameter subsequence to be measured
Multiple second dynamic similarity degree of the subsequence of C influence factors, for example, being provided with and LiAnd LrCorresponding two sections are long
Degree is all the subsequence of c-th of influence factor of t target componentWith parameter to be measured
Influence factor subsequence corresponding to subsequence
Wherein, what influence factor dynamic similarity degree characterized is the similarity degree of influence factor change, and reaction is to target component size
With the influence degree of changing rule.For example, use dc.i(c=1,2 ..., C, i=1,2 ..., r-1) represent " relative change rate it
With ", it can be described asWithRate of change difference degree.Calculation formula is：
In formula, xc.iAnd x (v)c.r(v) it is respectively influence factor subsequenceWithIn v (v=1,2 ..., t)
Individual influence factor value.The second dynamic similarity degree Sc.iMeet：
In formula, max (dc.i) and min (dc.i) it is respectively dc.iMaximum and minimum value, i=1,2 ..., r-1.
Step S207, based on described in the multiple first dynamic similarity degree and the multiple second dynamic similarity retrieval
The dynamic similarity sequence of parameter to be measured.
As a kind of embodiment, according to target component dynamic similarity degree weight set in advance and influence factor dynamic phase
Like degree weight, the first dynamic similarity degree and the product and described two dynamics of the target component dynamic similarity degree weight are obtained respectively
The sum of products of similar sequences and the influence factor dynamic similarity degree weight, will be being obtained and as integrating dynamic similarity
Degree, according to the sequence number of corresponding subsequence when finding out comprehensive dynamic similarity degree maximum to comprehensive dynamic similarity degree, by the sequence
Similar Time Series Based on Markov Chain of the target component subsequence corresponding to row number as subsequence where parameter to be measured, i.e. dynamic similarity sequence.
The target component dynamic similarity degree weight is 1 with the influence factor dynamic similarity degree weight sum.And the target component
Dynamic similarity degree weight is all higher than zero with the influence factor dynamic similarity degree weight.
Wherein, the comprehensive dynamic similarity degree OiMeet：
Wherein, i=1,2 ..., r-1, β0For target component dynamic similarity degree weight, βc(c=1,2 ..., C) is target
Each influence factor dynamic similarity degree weight of parameter.
In the present embodiment, for β0And βc(c=1,2 ..., C's) asks for, and can pass through known mesh in historical data
Dynamic state of parameters similarity, influence factor dynamic similarity degree and comprehensive dynamic similarity degree are marked, is tried to achieve using least square method to optimize.
In the present embodiment, according to comprehensive dynamic similarity degree OiThe calculating of (i=1,2 ..., r-1), is sorted by size
And the sequence number of corresponding subsequence during comprehensive dynamic similarity degree maximum is found out, be designated as I, i.e. I=i | max (Oi)}.Choose mesh
Mark parameter subsequence LI={ LI(1), LI..., L (2)I(t-1) } the Similar Time Series Based on Markov Chain as subsequence where parameter to be measured,
That is, dynamic similarity sequence.And LI(t) the then dynamic similarity parameter as parameter to be measured.
By accessed dynamic similarity sequence, the foundation chosen as Similar Time Series Based on Markov Chain, embedding data excavates pre-
Method of determining and calculating, to carry out the prediction of time series.
For example, in August, the 2014 per day load of 3 to 8 is predicted in the per day load prediction of power system.Its
In, target component is the per day load from January 1st, 2012 to prediction day proxima luce (prox. luc), and each influence factor data are from 2012
The meteorological data on January 1, in prediction same day day.Retrievable meteorological data includes：Daily maximum temperature, daily minimal tcmperature, day
Temperature on average, wind speed, precipitation and relative humidity.It is similar with experience by the Relative Error that dynamic similarity method is calculated
The Relative Error of method is shown in Table 1.By table 1 it can be found that dynamic similarity prediction error is integrally missed significantly lower than experience comparability prediction
Difference.
Table 1
Referring to Fig. 4, it is a kind of time Series Processing apparatus function module diagram that third embodiment of the invention provides.
The time Series Processing device 400 includes the first data capture unit 410, the second data capture unit 420 and data processing
Unit 430.
First data capture unit 410, for obtaining the mesh of the target component comprising parameter to be measured in predetermined time period
Mark parameter time series, wherein, the target component time series include M target component subsequence, wherein, M for more than etc.
In 2 integer.
Second data capture unit 420, for obtaining the influence factor of the target component in the predetermined time period
Time series.
Data processing unit 430, for based on the target component time series and the influence factor time series, from
The dynamic similarity sequence of the parameter to be measured is chosen in the M target component subsequence
Wherein, the data processing unit 430 includes：First data acquisition subelement, the second data acquisition subelement,
Three data acquisition subelements, the 4th data acquisition subelement and the 5th data acquisition subelement.
First data acquisition subelement, for obtaining the length of time series of the M target component subsequence.
Second data acquisition subelement, for obtaining the parameter to be measured of the parameter to be measured based on the length of time series
Subsequence.
3rd data acquisition subelement, for obtaining each target component subsequence and the parameter subsequence to be measured
Multiple first dynamic similarity degree.
4th data acquisition subelement, for obtaining the influence factor time series and the parameter subsequence to be measured
Multiple second dynamic similarity degree.
5th data acquisition subelement, for based on the multiple first dynamic similarity degree and the multiple second dynamic phase
The dynamic similarity sequence of the parameter to be measured is obtained like degree.
Wherein, the 5th data acquisition subelement includes：First data acquisition module, the second data acquisition module,
Three data acquisition modules and data processing module.
First data acquisition module, for obtaining the target component dynamic similarity degree corresponding to the first dynamic similarity degree
Influence factor dynamic similarity degree weight corresponding to weight and the second dynamic similarity degree, wherein, the target component dynamic
Similarity weight is 1 with the influence factor dynamic similarity degree weight sum.
Second data acquisition module, for obtaining each first dynamic similarity degree and the target component dynamic similarity
Spend the first product of weight.
3rd data acquisition module, for obtaining each second dynamic similarity degree and the influence factor dynamic similarity
Spend the second product of weight.
Data processing module, for obtaining the dynamic of the parameter to be measured based on first product and second product
Similar sequences.
Wherein, the data processing module includes：First data acquisition submodule, data processing submodule and the second data
Acquisition submodule.
First data acquisition submodule, for obtaining first product and second sum of products；
Data processing submodule, for using first product and second sum of products as comprehensive dynamic similarity
Degree.
Second data acquisition submodule, corresponding sub- sequence during for numerical value maximum based on the comprehensive dynamic similarity degree
The sequence number of row.
3rd data acquisition submodule, for using target component subsequence corresponding to the sequence number as the ginseng to be measured
Several dynamic similarity sequences.
Wherein, the data processing submodule is specifically used for：Based on the first dynamic similarity degree and second dynamic
Similarity obtains multiple comprehensive dynamic similarity degree；The synthesis that numerical value is maximum in the multiple comprehensive dynamic similarity degree is obtained to move
Subsequence corresponding to state similarity；The target subsequences corresponding to the subsequence are obtained, by the target subsequences
Dynamic similarity sequence as parameter to be measured.Wherein, the comprehensive dynamic similarity degree meets：
Wherein, i=1,2 ..., r-1, β0For the target component dynamic similarity degree weight, βc(c=1,2 ..., C) is
The influence factor dynamic similarity degree weight.
In summary, the present invention provides a kind of time Series Processing method and device, when this method is by first obtaining default
Between in length the target component comprising parameter to be measured target component time series, then obtain described in the predetermined time period
The influence factor time series of target component, it is finally based on the target component time series and the influence factor time sequence
Row, the dynamic similarity sequence of the parameter to be measured is chosen from the M target component subsequence.So as to by the change of target component
The change procedure synthesis of each influence factor of change process and target component is included in the selection rule of Similar Time Series Based on Markov Chain so that
In time series forecasting by accumulation effects, the selection accuracy for Similar Time Series Based on Markov Chain is higher and to time series
The accuracy of prediction is also of a relatively high, and then significantly reduces the prediction error to time series in the prior art.
In several embodiments provided herein, it should be understood that disclosed apparatus and method, can also pass through
Other modes are realized.Device embodiment described above is only schematical, for example, flow chart and block diagram in accompanying drawing
Show the device of multiple embodiments according to the present invention, method and computer program product architectural framework in the cards,
Function and operation.At this point, each square frame in flow chart or block diagram can represent the one of a module, program segment or code
Part, a part for the module, program segment or code include one or more and are used to realize holding for defined logic function
Row instruction.It should also be noted that at some as in the implementation replaced, the function that is marked in square frame can also with different from
The order marked in accompanying drawing occurs.For example, two continuous square frames can essentially perform substantially in parallel, they are sometimes
It can perform in the opposite order, this is depending on involved function.It is it is also noted that every in block diagram and/or flow chart
The combination of individual square frame and block diagram and/or the square frame in flow chart, function or the special base of action as defined in performing can be used
Realize, or can be realized with the combination of specialized hardware and computer instruction in the system of hardware.
In addition, each functional module in each embodiment of the present invention can integrate to form an independent portion
Point or modules individualism, can also two or more modules be integrated to form an independent part.
If the function is realized in the form of software function module and is used as independent production marketing or in use, can be with
It is stored in a computer read/write memory medium.Based on such understanding, technical scheme is substantially in other words
The part to be contributed to prior art or the part of the technical scheme can be embodied in the form of software product, the meter
Calculation machine software product is stored in a storage medium, including some instructions are causing a computer equipment (can be
People's computer, server, or network equipment etc.) perform all or part of step of each embodiment methods described of the present invention.
And foregoing storage medium includes：USB flash disk, mobile hard disk, read-only storage (ROM, Read-Only Memory), arbitrary access
Memory (RAM, Random Access Memory), magnetic disc or CD etc. are various can be with the medium of store program codes.Need
It is noted that herein, such as first and second or the like relational terms are used merely to an entity or operation
Made a distinction with another entity or operation, and not necessarily require or imply these entities or exist between operating any this
Actual relation or order.Moreover, term " comprising ", "comprising" or its any other variant are intended to nonexcludability
Comprising so that process, method, article or equipment including a series of elements not only include those key elements, but also wrapping
Include the other element being not expressly set out, or also include for this process, method, article or equipment intrinsic want
Element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that wanted including described
Other identical element also be present in the process of element, method, article or equipment.
The preferred embodiments of the present invention are the foregoing is only, are not intended to limit the invention, for the skill of this area
For art personnel, the present invention can have various modifications and variations.Within the spirit and principles of the invention, that is made any repaiies
Change, equivalent substitution, improvement etc., should be included in the scope of the protection.It should be noted that：Similar label and letter exists
Similar terms is represented in following accompanying drawing, therefore, once being defined in a certain Xiang Yi accompanying drawing, is then not required in subsequent accompanying drawing
It is further defined and explained.
Claims (10)
- A kind of 1. time Series Processing method, it is characterised in that including：The target component time series of the target component comprising parameter to be measured in predetermined time period is obtained, wherein, the target Parameter time series include M target component subsequence, wherein, M is the integer more than or equal to 2；Obtain the influence factor time series of the target component in the predetermined time period；Based on the target component time series and the influence factor time series, from the M target component subsequence Choose the dynamic similarity sequence of the parameter to be measured.
- 2. according to the method for claim 1, it is characterised in that described to be based on the target component time series and the shadow Factors time sequence is rung, the dynamic similarity sequence of the parameter to be measured is chosen from the M target component subsequence, including：Obtain the length of time series of the M target component subsequence；The parameter subsequence to be measured of the parameter to be measured is obtained based on the length of time series；Obtain multiple first dynamic similarity degree of each target component subsequence and the parameter subsequence to be measured；Obtain multiple second dynamic similarity degree of the influence factor time series and the parameter subsequence to be measured；The dynamic of the parameter to be measured is obtained based on the multiple first dynamic similarity degree and the multiple second dynamic similarity degree Similar sequences.
- 3. according to the method for claim 2, it is characterised in that described based on the multiple first dynamic similarity degree and described Multiple second dynamic similarity degree obtain the dynamic similarity sequence of the parameter to be measured, including：Obtain the target component dynamic similarity degree weight corresponding to the first dynamic similarity degree and the second dynamic similarity degree Corresponding influence factor dynamic similarity degree weight, wherein, the target component dynamic similarity degree weight and the influence factor Dynamic similarity degree weight sum is 1；Obtain each first dynamic similarity degree and the first product of the target component dynamic similarity degree weight；Obtain each second dynamic similarity degree and the second product of the influence factor dynamic similarity degree weight；The dynamic similarity sequence of the parameter to be measured is obtained based on first product and second product.
- 4. according to the method for claim 3, it is characterised in that described to be obtained based on first product with second product The dynamic similarity sequence of the parameter to be measured is taken, including：Obtain first product and second sum of products；Using first product and second sum of products as comprehensive dynamic similarity degree；The sequence number of corresponding subsequence during numerical value maximum based on the comprehensive dynamic similarity degree；Dynamic similarity sequence using target component subsequence corresponding to the sequence number as the parameter to be measured.
- 5. according to the method for claim 4, it is characterised in that the comprehensive dynamic similarity degree meets：Wherein, i=1,2 ..., r-1, β0For the target component dynamic similarity degree weight, βc(c=1,2 ..., C) is described Influence factor dynamic similarity degree weight.
- A kind of 6. time Series Processing device, it is characterised in that including：First data capture unit, for obtain in predetermined time period comprising parameter to be measured target component target component when Between sequence, wherein, the target component time series includes M target component subsequence, wherein, M is whole more than or equal to 2 Number；Second data capture unit, for obtaining the influence factor time sequence of the target component in the predetermined time period Row；Data processing unit, for based on the target component time series and the influence factor time series, from the M The dynamic similarity sequence of the parameter to be measured is chosen in target component subsequence.
- 7. device according to claim 6, it is characterised in that the data processing unit includes：First data acquisition subelement, for obtaining the length of time series of the M target component subsequence；Second data acquisition subelement, for obtaining the sub- sequence of parameter to be measured of the parameter to be measured based on the length of time series Row；3rd data acquisition subelement, for obtaining the more of each target component subsequence and the parameter subsequence to be measured Individual first dynamic similarity degree；4th data acquisition subelement, for obtaining the multiple of the influence factor time series and the parameter subsequence to be measured Second dynamic similarity degree；5th data acquisition subelement, for based on the multiple first dynamic similarity degree and the multiple second dynamic similarity degree Obtain the dynamic similarity sequence of the parameter to be measured.
- 8. device according to claim 7, it is characterised in that the 5th data acquisition subelement includes：First data acquisition module, for obtaining the target component dynamic similarity degree weight corresponding to the first dynamic similarity degree With the influence factor dynamic similarity degree weight corresponding to the second dynamic similarity degree, wherein, the target component dynamic similarity It is 1 to spend weight with the influence factor dynamic similarity degree weight sum；Second data acquisition module, weighed for obtaining each first dynamic similarity degree with the target component dynamic similarity degree First product of weight；3rd data acquisition module, weighed for obtaining each second dynamic similarity degree with the influence factor dynamic similarity degree Second product of weight；Data processing module, for obtaining the dynamic similarity of the parameter to be measured based on first product and second product Sequence.
- 9. device according to claim 8, it is characterised in that the data processing module includes：First data acquisition submodule, for obtaining first product and second sum of products；Data processing submodule, for using first product and second sum of products as comprehensive dynamic similarity degree；Second data acquisition submodule, corresponding subsequence during for numerical value maximum based on the comprehensive dynamic similarity degree Sequence number；3rd data acquisition submodule, for using target component subsequence corresponding to the sequence number as the parameter to be measured Dynamic similarity sequence.
- 10. device according to claim 9, it is characterised in that the comprehensive dynamic similarity degree meets：Wherein, i=1,2 ..., r-1, β0For the target component dynamic similarity degree weight, βc(c=1,2 ..., C) is described Influence factor dynamic similarity degree weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710753494.4A CN107516114B (en) | 2017-08-28 | 2017-08-28 | Time sequence processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710753494.4A CN107516114B (en) | 2017-08-28 | 2017-08-28 | Time sequence processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107516114A true CN107516114A (en) | 2017-12-26 |
| CN107516114B CN107516114B (en) | 2020-06-19 |
Family
ID=60724328
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710753494.4A Active CN107516114B (en) | 2017-08-28 | 2017-08-28 | Time sequence processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107516114B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537466A (en) * | 2018-04-25 | 2018-09-14 | 北京腾云天下科技有限公司 | A kind of method and computing device of statistics application operation indicator |
| CN110674375A (en) * | 2019-09-25 | 2020-01-10 | 联想(北京)有限公司 | Data processing method and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101082972A (en) * | 2007-05-30 | 2007-12-05 | 华为技术有限公司 | Method and device for forecasting user's interest to commercial product and method for publishing advertisement thereof |
| CN101739814A (en) * | 2009-11-06 | 2010-06-16 | 吉林大学 | SCATS coil data-based traffic state online quantitative evaluation and prediction method |
| CN101877077A (en) * | 2009-11-25 | 2010-11-03 | 天津工业大学 | Time series predicting model |
| CN103577562A (en) * | 2013-10-24 | 2014-02-12 | 河海大学 | Multi-measurement time series similarity analysis method |
| CN105608096A (en) * | 2014-11-21 | 2016-05-25 | 财团法人资讯工业策进会 | Time series data processing method and system |
-
2017
- 2017-08-28 CN CN201710753494.4A patent/CN107516114B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101082972A (en) * | 2007-05-30 | 2007-12-05 | 华为技术有限公司 | Method and device for forecasting user's interest to commercial product and method for publishing advertisement thereof |
| CN101739814A (en) * | 2009-11-06 | 2010-06-16 | 吉林大学 | SCATS coil data-based traffic state online quantitative evaluation and prediction method |
| CN101877077A (en) * | 2009-11-25 | 2010-11-03 | 天津工业大学 | Time series predicting model |
| CN103577562A (en) * | 2013-10-24 | 2014-02-12 | 河海大学 | Multi-measurement time series similarity analysis method |
| CN105608096A (en) * | 2014-11-21 | 2016-05-25 | 财团法人资讯工业策进会 | Time series data processing method and system |
Non-Patent Citations (2)
| Title |
|---|
| MARTINEZ V，ET AL.: "CONVEX:similarity-based algorithms for forecasting group behavior", 《IEEE INTELLIGENT SYSTEMS》 * |
| 莫维仁等: "短期负荷预测中选择相似日的探讨", 《清华大学学报（自然科学版）》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537466A (en) * | 2018-04-25 | 2018-09-14 | 北京腾云天下科技有限公司 | A kind of method and computing device of statistics application operation indicator |
| CN108537466B (en) * | 2018-04-25 | 2020-09-15 | 北京腾云天下科技有限公司 | Method for counting application operation indexes and computing equipment |
| CN110674375A (en) * | 2019-09-25 | 2020-01-10 | 联想(北京)有限公司 | Data processing method and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107516114B (en) | 2020-06-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Azar et al. | Dynamic policy programming | |
| CN107526667A (en) | A kind of Indexes Abnormality detection method, device and electronic equipment | |
| CN105224543A (en) | For the treatment of seasonal effect in time series method and apparatus | |
| CN107516114A (en) | A kind of time Series Processing method and device | |
| CN103036231A (en) | Forecasting method, device, and upper computer of power load | |
| CN111831704A (en) | Abnormal data determination method and device, storage medium and electronic equipment | |
| JP2020501232A (en) | Risk control event automatic processing method and apparatus | |
| CN107370766A (en) | A kind of network flow abnormal detecting method and system | |
| CN104517020A (en) | Characteristic extraction method and device used for cause and effect analysis | |
| Vogel et al. | A novel method to optimize electricity generation from wind energy | |
| Wang | A big data framework for stock price forecasting using fuzzy time series | |
| Cormier et al. | Simple models and insights for warehouse sizing | |
| CN102023993B (en) | Cluster page ranking equipment and method based on clustering/classification and time | |
| CN116384574A (en) | W-LSTM-based non-parameter load interval prediction method | |
| Haug | Closed form valuation of American barrier options | |
| CN111175810B (en) | Microseismic signal arrival time picking method, device, equipment and storage medium | |
| CN117092582A (en) | Electric energy meter abnormality detection method and device based on contrast self-encoder | |
| Xylogiannopoulos et al. | Periodicity data mining in time series using suffix arrays | |
| CN115238779B (en) | Cloud disk abnormality detection method, device, equipment and medium | |
| Prodinger | Approximate counting with m counters: a detailed analysis | |
| US20190205804A1 (en) | Information processing device, information processing method and computer readable medium | |
| CN113850427A (en) | Time series prediction method, electronic device, and computer storage medium | |
| CN114358382A (en) | Power system equivalent inertia probability prediction method, device, product and storage medium | |
| CN107392220A (en) | The clustering method and device of data flow | |
| CN107605472B (en) | Method and device for determining oil reservoir recovery ratio |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 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 |