CN107491828A - A kind of distributed new based on the optimal multi-kernel function of multi-source time-variable data goes out force prediction method - Google Patents
A kind of distributed new based on the optimal multi-kernel function of multi-source time-variable data goes out force prediction method Download PDFInfo
- Publication number
- CN107491828A CN107491828A CN201610420301.9A CN201610420301A CN107491828A CN 107491828 A CN107491828 A CN 107491828A CN 201610420301 A CN201610420301 A CN 201610420301A CN 107491828 A CN107491828 A CN 107491828A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- mtd
- kernel function
- munderover
- 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
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000009826 distribution Methods 0.000 claims abstract description 32
- 102000008297 Nuclear Matrix-Associated Proteins Human genes 0.000 claims abstract description 10
- 108010035916 Nuclear Matrix-Associated Proteins Proteins 0.000 claims abstract description 10
- 210000000299 nuclear matrix Anatomy 0.000 claims abstract description 10
- 230000006870 function Effects 0.000 claims description 145
- 241001269238 Data Species 0.000 claims description 16
- 239000000203 mixture Substances 0.000 claims description 8
- 238000012706 support-vector machine Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000008689 nuclear function Effects 0.000 claims description 3
- 230000004927 fusion Effects 0.000 description 4
- 201000004569 Blindness Diseases 0.000 description 3
- 238000007418 data mining Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000002360 explosive Substances 0.000 description 2
- 206010068052 Mosaicism Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 210000003765 sex chromosome Anatomy 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
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"
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
Abstract
The invention discloses a kind of distributed new based on the optimal multi-kernel function of multi-source time-variable data to go out force prediction method, by using multi-source time-variable data distribution characteristics method, single source data SVM prediction or classification, the nuclear matrix rank space method of difference, structure reflection influences the optimal multi-kernel function for the multi-source time-variable data characteristic distributions that distributed new is contributed, this optimal multi-kernel function is used when distributed new is contributed and predicted, multi-source time-variable data can be merged, it is energy output precision of prediction to improve distribution.
Description
Technical field
The present invention relates to the distributed new for sending out a kind of based on the optimal multi-kernel function of multi-source time-variable data output prediction side
Method, belong to Data Mining.
Background technology
In recent years, with the fast development and popularization and application of computer and information technology, the scale of sector application system
Rapid to expand, data caused by sector application are in explosive increase.The data volume of this explosive increase to multi-source, isomery,
Higher-dimension, distribution, the data of uncertainty propose challenge, and big data thinking is exactly the product in this environment, big data it is notable
One of feature is exactly multi-source heterogeneous characteristic.
The data for influenceing distributed new prediction have a lot, such as distributed new access and operation information, power network
Production management information, geographical weather information etc., these data are exactly multi-source data, if effectively utilizing point of multi-source time-variable data
Boot point carries out distributed new output prediction, then can improve precision of prediction.
Multi-kernel function is to realize the effective ways of multi-source heterogeneous data fusion, and its base nuclear energy in forming is enough as a series of biographies
Sensor equally perceives to respective information, using the teaching of the invention it is possible to provide more flexible and effective information tissue and data mining duty.At present
The method that selection is formed on multi-kernel function, mainly there is two methods:
1st, a global kernel function and a local kernel function are directly chosen, mixed kernel function is formed after simple weighted, this
The shortcomings that kind of method is:1. the composition of limitation multi-kernel function only has two monokaryon functions;2. do not account for data with existing to concentrate respectively
The feature of individual multi-source heterogeneous data;3. conventional global kernel function has Polynomial kernel function Polynomial and Sigmoid core letter
Number, this method select global kernel function when be typically optional one, there is blindness.
2nd, it is predicted or classifies first with monokaryon SVM, then chooses precision two kinds of kernel functions of highest and be combined
Form multinuclear letter.This method have ignored the variety classes feature from different data sources, not make full use of multi-kernel function pair
Multi-source heterogeneous data are effectively treated.
The content of the invention
The purpose of the present invention is to overcome above-mentioned the deficiencies in the prior art, there is provided one kind is based on the optimal multinuclear of multi-source time-variable data
The distributed new of function goes out force prediction method, and this method is built by multi-kernel function, function composition compares, function performance ratio
Parameter when can effectively handle distributed new related multi-source compared with three steps come optimal multi-kernel function when determining SVM predictions
According to there is blindness sex chromosome mosaicism, raising precision of prediction to solve the selection of current support vector machine method Kernel Function.
The present invention is as follows to solve the technical solution of above-mentioned technical problem:
A kind of distributed new of pin based on the optimal multi-kernel function of multi-source time-variable data goes out force prediction method, specific steps
It is as follows:
Step 1, determine to influence the time-variable data source of distributed new output prediction:Source of meteorological data, geodata source,
New energy operation information data source;
Step 2, the A kind meteorological datas from the weather station of distributed new plant stand collection source of meteorological data, from generalized information system
The B kind geodatas of geographical data source are gathered, from the C kinds new energy fortune of SCADA system collection new energy operation information data source
Row data, Q=A+B+C kinds time-variable data, composition collection are combined into altogetherD represents any one time-variable data, daily collection
For Xd=[x1,x2,…,xT]∈R1×T, T represents the data for gathering T moment altogether daily, in new energy service data, is distributed
It is Y that formula new energy, which is contributed,i=[y1,y2,…,yT]∈R1×T, gather N days altogether;
Step 3, utilize multi-source time-variable data distribution characteristics method, single source data SVM prediction or classification, nuclear moment
The battle array rank space method of difference, optimal multi-kernel function K of the structure for distributed new time-variable data source distribution featureopt(xi,xj),
I, j represent every kind of time-variable data any two moment;
Step 4, step 4, output prediction is carried out to distributed new using support vector machines, kernel function is using step
The optimal multi-kernel function K obtained in rapid 3opt(xi,xj)。
Structure is for the optimal multi-kernel function of distributed new time-variable data source distribution feature, specific steps in step 3
It is as follows:
(1) multi-kernel function is built:
1A. is for d kind time-variable datas, the time-variable data X of common N daysdCharacteristic distributions in corresponding coordinate system following table
Show, according to the distribution of data in figure, judge the distribution character of d-th of data source time-variable data, according to distribution character, select its right
Answer optimal monokaryon function Kd1(xi,xj), all Q time-variable datas are carried out with the operation, structure distribution characteristics multinuclear kernel function K1
(xi,xj):
Wherein:Kd1(xi,xj) it is in linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid kernel functions
One kind;
1B. is for d kind time-variable datas, using it as linear kernel function, Polynomial kernel function, Radial basis kernel function
With the input feature vector amount of Sigmoid kernel function support vector machines, comparison prediction or nicety of grading, precision of prediction highest core is selected
Kernel function K of the function as d-th of data source time-variable datad2(xi,xj), the operation is carried out to all Q time-variable datas, built
SVM multinuclear kernel functions K2(xi,xj):
Wherein:Kd2(xi,xj) it is in linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid kernel functions
One kind;
(2) multi-kernel function is formed and compared:If K1(xi,xj) and K2(xi,xj) kernel function form identical, then this multi-kernel function
For optimal multi-kernel function corresponding to multi-source time-variable data:Kopt(xi,xj)=K1(xi,xj)=K2(xi,xj), if K1(xi,xj) and K2
(xi,xj) composition it is different, then turn (3);
(3) multi-kernel function performance comparision:
3A. calculates optimal weighting coefficientses λ:
3B. is to multi-kernel function K1(xi,xj) and K2(xi,xj) be weighted, form weighting multinuclear kernel function K3(xi,xj):
Wherein:Wherein:Kd3(xi,xj) it is linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid cores
One kind in function, λ are the optimal weighting coefficientses that calculate in 3A, 0<λ<1;
3C. is calculated in K for the time-variable data of d kind data sourcesd1(xi,xj)、Kd2(xi,xj)、Kd3(xi,xj) kernel function
Nuclear matrix under mapping:
Wherein:D=1,2 ..., Q, s=1,2,3, s represent K1(xi,xj)、K2(xi,xj)、K3(xi,xj) three seed nucleus function pairs
The subscript answered;
3D. utilizes nuclear matrix rank space method of difference method, calculates the A obtained by 3B1,s,A2,s,…,AQ,s, s=1,2,3 order
Spatial Difference is measured:
Wherein:D=1,2 ..., Q, s=1,2,3, rank (A) are the order for seeking matrix A;
3E. compares RSD1、RSD2With RSD3The size of value, function corresponding to selection maximum melt as multi-source time-variable data
The optimal multi-kernel function K closedopt(xi,xj)。
Compared with prior art, the beneficial effects of the invention are as follows:
The Selection of kernel function of traditional support vector machine is always an open question, for that can realize multi-source heterogeneous data
The multi-kernel function selection of fusion more has blindness so that distributed new output precision of prediction is relatively low, and of the invention adopts
With multi-source time-variable data distribution characteristics method, single source data SVM prediction or classification, the nuclear matrix rank space method of difference, energy
Enough judge the characteristic distributions of all distributed new correlation multi-source time-variable datas, local kernel function is selected still for every kind of data
Global kernel function provides foundation, constructs optimal multi-kernel function, realizes the fusion of multi-source time-variable data, and it is energy to improve distribution
Source output precision of prediction.
Brief description of the drawings
Fig. 1 is that the distributed new based on the optimal multi-kernel function of multi-source time-variable data goes out force prediction method
Fig. 2 is using the flow chart of multi-source time-variable data distribution characteristics method construction multi-kernel function in step 3
Fig. 3 is the flow chart that optimal multi-kernel function is selected after two multi-kernel function linear weighted functions of step 3
Embodiment
To become apparent the present invention, hereby to be preferable to carry out example, and accompanying drawing is coordinated to be described in detail below.Fig. 1 is more
The distributed new of the optimal multi-kernel function of source time-variable data goes out force prediction method overall flow figure, i.e. following step 1 to step
4。
Step 1, determine to influence the time-variable data source of distributed new output prediction:Source of meteorological data, geodata source,
New energy operation information data source;
Step 2, the A kind meteorological datas from the weather station of distributed new plant stand collection source of meteorological data, from generalized information system
The B kind geodatas of geographical data source are gathered, from the C kinds new energy fortune of SCADA system collection new energy operation information data source
Row data, Q=A+B+C kinds time-variable data, composition collection are combined into altogetherD represents any one time-variable data, daily collection
For Xd=[x1,x2,…,xT]∈R1×T, T represents the data for gathering T moment altogether daily, gathered N days altogether, to be predicted distributed new
It is Y that the energy, which is contributed,i=[y1,y2,…,yT]∈R1×T;
Step 3, utilize multi-source time-variable data distribution characteristics method, single source data SVM prediction or classification, nuclear moment
The battle array rank space method of difference, optimal multi-kernel function K of the structure for distributed new time-variable data source distribution featureopt(xi,xj),
I, j represent every kind of time-variable data any two moment, comprise the following steps that:
(1) multi-kernel function is built:
1A. is for d kind time-variable datas, the time-variable data X of common N daysdCharacteristic distributions in corresponding coordinate system following table
Show, according to the distribution of data in figure, judge the distribution character of d-th of data source time-variable data, according to distribution character, select its right
Answer optimal monokaryon function Kd1(xi,xj), all Q time-variable datas are carried out with the operation, structure distribution characteristics multinuclear kernel function K1
(xi,xj):
Wherein:Kd1(xi,xj) it is in linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid kernel functions
One kind;
1B. is for d kind time-variable datas, using it as linear kernel function, Polynomial kernel function, Radial basis kernel function
With the input feature vector amount of Sigmoid kernel function support vector machines, comparison prediction or nicety of grading, precision of prediction highest core is selected
Kernel function K of the function as d-th of data source time-variable datad2(xi,xj), the operation is carried out to all Q time-variable datas, built
SVM multinuclear kernel functions K2(xi,xj):
Wherein:Kd2(xi,xj) it is in linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid kernel functions
One kind;
(2) multi-kernel function is formed and compared:If K1(xi,xj) and K2(xi,xj) kernel function form identical, then this multi-kernel function
For optimal multi-kernel function corresponding to multi-source time-variable data:Kopt(xi,xj)=K1(xi,xj)=K2(xi,xj), if K1(xi,xj) and K2
(xi,xj) composition it is different, then turn (3);
(3) multi-kernel function performance comparision:
3A. calculates optimal weighting coefficientses λ:
1. in multi-kernel function, the function after mapping is:
In formula:φ3、φ1、φ2Respectively weight multi-kernel function K3(xi,xj), multi-source time-variable data distribution characteristics method multinuclear
Function K1(xi,xj), single source data SVM prediction or classification multi-kernel function K2(xi,xj) corresponding to Function Mapping;
2. distance is:
3. seek the One- place 2-th Order multinomial on λ:
4. maximizing L (λ), that is, the polynomial max problem of One- place 2-th Order is sought, then λ is tried to achieve by following formula:
3B. is to multi-kernel function K1(xi,xj) and K2(xi,xj) be weighted, form weighting multinuclear kernel function K3(xi,xj):
Wherein:Wherein:Kd3(xi,xj) it is linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid cores
One kind in function, λ are the optimal weighting coefficientses that calculate in 3A, 0<λ<1;
3C. is calculated in K for the time-variable data of d kind data sourcesd1(xi,xj)、Kd2(xi,xj)、Kd3(xi,xj) kernel function
Nuclear matrix under mapping:
Wherein:D=1,2 ..., Q, s=1,2,3, s represent K1(xi,xj)、K2(xi,xj)、K3(xi,xj) three seed nucleus function pairs
The subscript answered;
3D. utilizes nuclear matrix rank space method of difference method, calculates the A obtained by 3B1,s,A2,s,…,AQ,s, s=1,2,3 order
Spatial Difference is measured:
Wherein:D=1,2 ..., Q, s=1,2,3, rank (A) are the order for seeking matrix A;
3E. compares RSD1、RSD2With RSD3The size of value, function corresponding to selection maximum melt as multi-source time-variable data
The optimal multi-kernel function K closedopt(xi,xj)。
Step 4, output prediction is carried out to distributed new using support vector machines, kernel function is used in step 3 and obtained
The optimal multi-kernel function K arrivedopt(xi,xj)。
The present invention is by using using multi-source time-variable data distribution characteristics method, single source data SVM prediction or classification
Method, the nuclear matrix rank space method of difference, construct optimal multi-kernel function, realize the fusion of multi-source time-variable data, improve distributed
New energy output precision of prediction, is suitable for the analysis of multi-source data.
Claims (2)
1. a kind of distributed new based on the optimal multi-kernel function of multi-source time-variable data goes out force prediction method, it is characterised in that
This method comprises the following steps:
Step 1, determine to influence the time-variable data source of distributed new output prediction, including source of meteorological data, geodata source
With new energy service data source;
Step 2, the A kind meteorological datas from the weather station of distributed new plant stand collection source of meteorological data, are gathered from generalized information system
The B kind geodatas in geodata source, number is run from the C kinds new energy of SCADA system collection new energy operation information data source
According to Q=A+B+C kinds time-variable data, composition collection are combined into altogetherD represents any one time-variable data, and collection is X dailyd
=[x1,x2,…,xT]∈R1×T, the daily data for gathering T moment altogether of T representatives, gather N days altogether, distributed new to be predicted
Contribute as Yi=[y1,y2,…,yT]∈R1×T;
Step 3, utilize multi-source time-variable data distribution characteristics method, single source data SVM prediction or classification, nuclear matrix order
Spatial diversity method, optimal multi-kernel function K of the structure for distributed new time-variable data source distribution featureopt(xi,xj), i, j
Represent every kind of time-variable data any two moment;
Step 4, output prediction is carried out to distributed new using support vector machines, kernel function uses what is obtained in step 3
Optimal multi-kernel function Kopt(xi,xj)。
2. optimal multinuclear letter of step 3 structure for distributed new time-variable data source distribution feature in the claim 1
Number, it is characterised in that the step 3 is specific as follows:
(1) multi-kernel function is built:
1A. is for d kind time-variable datas, the time-variable data X of common N daysdCharacteristic distributions represented under corresponding coordinate system, root
According to the distribution of data in coordinate system, judge the distribution character of d-th of data source time-variable data, select corresponding optimal monokaryon function Kd1
(xi,xj), all Q time-variable datas are carried out with the operation, structure distribution characteristics multinuclear kernel function K1(xi,xj):
<mrow>
<msub>
<mi>K</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>d</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>Q</mi>
</munderover>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mn>1</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
Wherein:Kd1(xi,xj) be linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid kernel functions in one
Kind;
1B. for d kind time-variable datas, using its as linear kernel function, Polynomial kernel function, Radial basis kernel function with
The input feature vector amount of Sigmoid kernel function support vector machines, comparison prediction or nicety of grading, select precision of prediction highest core letter
Kernel function K of the number as d-th of data source time-variable datad2(xi,xj), the operation is carried out to all Q time-variable datas, built
SVM multinuclear kernel functions K2(xi,xj):
<mrow>
<msub>
<mi>K</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>d</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>Q</mi>
</munderover>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mn>2</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
Wherein:Kd2(xi,xj) be linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid kernel functions in one
Kind;
(2) multi-kernel function is formed and compared:If K1(xi,xj) and K2(xi,xj) kernel function form it is identical, then this multi-kernel function is more
Optimal multi-kernel function corresponding to the time-variable data of source:Kopt(xi,xj)=K1(xi,xj)=K2(xi,xj), if K1(xi,xj) and K2(xi,
xj) composition it is different, then turn (3);
(3) multi-kernel function performance comparision:
3A. calculates optimal weighting coefficientses λ:
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>D</mi>
<mo>=</mo>
<msub>
<mi>K</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>K</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>K</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>E</mi>
<mo>=</mo>
<msub>
<mi>K</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mn>2</mn>
<msub>
<mi>K</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>+</mo>
<msub>
<mi>K</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>A</mi>
<mo>=</mo>
<mi>D</mi>
<mo>+</mo>
<mi>E</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>B</mi>
<mo>=</mo>
<mo>-</mo>
<mn>2</mn>
<mi>E</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mrow>
<mi>&lambda;</mi>
<mo>=</mo>
<mo>-</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>T</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>By</mi>
<mi>i</mi>
</msub>
<msub>
<mi>y</mi>
<mi>j</mi>
</msub>
</mrow>
<mrow>
<mn>2</mn>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>T</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>Ay</mi>
<mi>i</mi>
</msub>
<msub>
<mi>y</mi>
<mi>j</mi>
</msub>
</mrow>
</mfrac>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>T</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>Ey</mi>
<mi>i</mi>
</msub>
<msub>
<mi>y</mi>
<mi>j</mi>
</msub>
</mrow>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>T</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mrow>
<mi>i</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<mrow>
<mo>(</mo>
<mrow>
<mi>D</mi>
<mo>+</mo>
<mi>E</mi>
</mrow>
<mo>)</mo>
</mrow>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<msub>
<mi>y</mi>
<mi>j</mi>
</msub>
</mrow>
</mfrac>
</mrow>
3B. is to multi-kernel function K1(xi,xj) and K2(xi,xj) be weighted, form weighting multinuclear kernel function K3(xi,xj):
<mfenced open = "" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>K</mi>
<mn>3</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>&lambda;K</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mrow>
<mo>(</mo>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>&lambda;</mi>
</mrow>
<mo>)</mo>
</mrow>
<msub>
<mi>K</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>=</mo>
<mi>&lambda;</mi>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>d</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>Q</mi>
</munderover>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mn>1</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mrow>
<mo>(</mo>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>&lambda;</mi>
</mrow>
<mo>)</mo>
</mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>d</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>Q</mi>
</munderover>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mn>2</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>d</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>Q</mi>
</munderover>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<msub>
<mi>&lambda;K</mi>
<mrow>
<mi>d</mi>
<mn>1</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mrow>
<mo>(</mo>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>&lambda;</mi>
</mrow>
<mo>)</mo>
</mrow>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mn>2</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mo>&rsqb;</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>d</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>Q</mi>
</munderover>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mn>3</mn>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Wherein:Wherein:Kd3(xi,xj) it is in linear kernel function, Polynomial kernel function, Radial basis kernel function and Sigmoid kernel functions
One kind, a is the optimal weighting coefficientses that calculate in 3A, 0<λ<1;
3C. is calculated in K for the time-variable data of d kind data sourcesd1(xi,xj)、Kd2(xi,xj)、Kd3(xi,xj) kernel function mapping
Under nuclear matrix:
<mrow>
<msub>
<mi>A</mi>
<mrow>
<mi>d</mi>
<mo>,</mo>
<mi>s</mi>
</mrow>
</msub>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mi>s</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mi>L</mi>
</mtd>
<mtd>
<mrow>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mi>s</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mn>1</mn>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>T</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>M</mi>
</mtd>
<mtd>
<mi>O</mi>
</mtd>
<mtd>
<mi>M</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mi>s</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>T</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
<mtd>
<mi>L</mi>
</mtd>
<mtd>
<mrow>
<msub>
<mi>K</mi>
<mrow>
<mi>d</mi>
<mi>s</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>T</mi>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mi>T</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Wherein:D=1,2 ..., Q, s=1,2,3, s represent K1(xi,xj)、K2(xi,xj)、K3(xi,xj) corresponding to three seed nucleus functions
Subscript;
3D. utilizes nuclear matrix rank space method of difference method, calculates the A obtained by 3B1,s,A2,s,…,AQ,s, s=1,2,3 rank space
Diversity measure:
<mrow>
<msub>
<mi>RSD</mi>
<mi>s</mi>
</msub>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>A</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
</msub>
<mo>,</mo>
<msub>
<mi>A</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
</msub>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<msub>
<mi>A</mi>
<mrow>
<mi>Q</mi>
<mo>,</mo>
<mi>s</mi>
</mrow>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>Q</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>d</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>Q</mi>
</munderover>
<mfenced open = "{" close = "}">
<mtable>
<mtr>
<mtd>
<mrow>
<mn>2</mn>
<mo>&times;</mo>
<mi>r</mi>
<mi>a</mi>
<mi>n</mi>
<mi>k</mi>
<mrow>
<mo>(</mo>
<msup>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msubsup>
<mi>A</mi>
<mrow>
<mi>d</mi>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
</mtd>
<mtd>
<mrow>
<msubsup>
<mi>A</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
<mo>...</mo>
<msubsup>
<mi>A</mi>
<mrow>
<mi>d</mi>
<mo>-</mo>
<mn>1</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
<msubsup>
<mi>A</mi>
<mrow>
<mi>d</mi>
<mo>+</mo>
<mn>1</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
<mo>...</mo>
<msubsup>
<mi>A</mi>
<mrow>
<mi>Q</mi>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mi>T</mi>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>-</mo>
<mi>r</mi>
<mi>a</mi>
<mi>n</mi>
<mi>k</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mrow>
<mi>d</mi>
<mo>,</mo>
<mi>s</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>r</mi>
<mi>a</mi>
<mi>n</mi>
<mi>k</mi>
<mrow>
<mo>(</mo>
<msup>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<msubsup>
<mi>A</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
<mo>...</mo>
<msubsup>
<mi>A</mi>
<mrow>
<mi>d</mi>
<mo>-</mo>
<mn>1</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
<msubsup>
<mi>A</mi>
<mrow>
<mi>d</mi>
<mo>+</mo>
<mn>1</mn>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
<mo>...</mo>
<msubsup>
<mi>A</mi>
<mrow>
<mi>Q</mi>
<mo>,</mo>
<mi>s</mi>
</mrow>
<mi>T</mi>
</msubsup>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mi>T</mi>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Wherein:D=1,2 ..., Q, s=1,2,3, rank (A) are the order of matrix A;
3E. compares RSD1、RSD2With RSD3The size of value, what function corresponding to selection maximum merged as multi-source time-variable data
Optimal multi-kernel function Kopt(xi,xj)。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610420301.9A CN107491828B (en) | 2016-06-13 | 2016-06-13 | Distributed new energy output prediction method based on multi-source time-varying data optimal multi-core function |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610420301.9A CN107491828B (en) | 2016-06-13 | 2016-06-13 | Distributed new energy output prediction method based on multi-source time-varying data optimal multi-core function |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107491828A true CN107491828A (en) | 2017-12-19 |
CN107491828B CN107491828B (en) | 2020-11-10 |
Family
ID=60642358
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610420301.9A Expired - Fee Related CN107491828B (en) | 2016-06-13 | 2016-06-13 | Distributed new energy output prediction method based on multi-source time-varying data optimal multi-core function |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107491828B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109471884A (en) * | 2018-09-12 | 2019-03-15 | 国网浙江省电力有限公司嘉兴供电公司 | The relevant multi-source heterogeneous data processing method of distributed new |
CN110019488A (en) * | 2018-09-12 | 2019-07-16 | 国网浙江省电力有限公司嘉兴供电公司 | Multi-source heterogeneous data fusion multi-core classification method |
CN111598156A (en) * | 2020-05-14 | 2020-08-28 | 北京工业大学 | PM based on multi-source heterogeneous data fusion2.5Prediction model |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040034612A1 (en) * | 2002-03-22 | 2004-02-19 | Nick Mathewson | Support vector machines for prediction and classification in supply chain management and other applications |
CN102542167B (en) * | 2011-12-31 | 2014-12-24 | 东北电力大学 | Wind-speed time series forecasting method for wind power station |
CN105184424B (en) * | 2015-10-19 | 2017-07-07 | 国网山东省电力公司菏泽供电公司 | Realize that the multi-kernel function of multi-source heterogeneous data fusion learns the Mapreduceization short-term load forecasting method of SVM |
-
2016
- 2016-06-13 CN CN201610420301.9A patent/CN107491828B/en not_active Expired - Fee Related
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109471884A (en) * | 2018-09-12 | 2019-03-15 | 国网浙江省电力有限公司嘉兴供电公司 | The relevant multi-source heterogeneous data processing method of distributed new |
CN110019488A (en) * | 2018-09-12 | 2019-07-16 | 国网浙江省电力有限公司嘉兴供电公司 | Multi-source heterogeneous data fusion multi-core classification method |
CN111598156A (en) * | 2020-05-14 | 2020-08-28 | 北京工业大学 | PM based on multi-source heterogeneous data fusion2.5Prediction model |
Also Published As
Publication number | Publication date |
---|---|
CN107491828B (en) | 2020-11-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103150609B (en) | Modeling method for short time traffic flow predicting model | |
CN108549929B (en) | A kind of photovoltaic power prediction technique based on deep layer convolutional neural networks | |
Gharehchopogh et al. | Slime mould algorithm: A comprehensive survey of its variants and applications | |
CN106971240A (en) | The short-term load forecasting method that a kind of variables choice is returned with Gaussian process | |
CN110532571A (en) | Text handling method and relevant apparatus | |
Sharma et al. | Identification of photovoltaic module parameters by implementing a novel teaching learning based optimization with unique exemplar generation scheme (TLBO-UEGS) | |
CN109659933A (en) | A kind of prediction technique of power quality containing distributed power distribution network based on deep learning model | |
Liu et al. | A logistic based cellular automata model for continuous urban growth simulation: A case study of the Gold Coast City, Australia | |
CN103679263A (en) | Thunder and lightning approach forecasting method based on particle swarm support vector machine | |
CN104951834A (en) | LSSVM (least squares support vector machine) wind speed forecasting method based on integration of GA (genetic algorithm) and PSO (particle swarm optimization) | |
CN104166731A (en) | Discovering system for social network overlapped community and method thereof | |
CN110473592A (en) | The multi-angle of view mankind for having supervision based on figure convolutional network cooperate with lethal gene prediction technique | |
CN110059875A (en) | Public bicycles Demand Forecast method based on distributed whale optimization algorithm | |
Yuan et al. | Regional land-use allocation using a coupled MAS and GA model: From local simulation to global optimization, a case study in Caidian District, Wuhan, China | |
CN107491828A (en) | A kind of distributed new based on the optimal multi-kernel function of multi-source time-variable data goes out force prediction method | |
CN110348665A (en) | A kind of low-voltage platform area electric power system data quality evaluating method and device | |
CN109492748A (en) | A kind of Mid-long term load forecasting method for establishing model of the electric system based on convolutional neural networks | |
CN105654136A (en) | Deep learning based automatic target identification method for large-scale remote sensing images | |
CN107423820A (en) | The knowledge mapping of binding entity stratigraphic classification represents learning method | |
CN105868906A (en) | Optimized method for analyzing maturity of regional development | |
Dai et al. | Agent-based model of land system: Theory, application and modelling framework | |
CN102857525A (en) | Community Discovery Method Based on Random Walk Strategy | |
Zhou et al. | Advanced orthogonal learning and Gaussian barebone hunger games for engineering design | |
CN104732091A (en) | Cellular automaton river bed evolution prediction method based on natural selection ant colony algorithm | |
CN102184328A (en) | Method for optimizing land use evolution CA model transformation rules |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201110 |
|
CF01 | Termination of patent right due to non-payment of annual fee |