CN108596781A - A kind of electric power system data excavates and prediction integration method - Google Patents
A kind of electric power system data excavates and prediction integration method Download PDFInfo
- Publication number
- CN108596781A CN108596781A CN201810333780.XA CN201810333780A CN108596781A CN 108596781 A CN108596781 A CN 108596781A CN 201810333780 A CN201810333780 A CN 201810333780A CN 108596781 A CN108596781 A CN 108596781A
- Authority
- CN
- China
- Prior art keywords
- prediction
- electric power
- power system
- excavates
- integration method
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000010354 integration Effects 0.000 title claims abstract description 18
- 238000012549 training Methods 0.000 claims abstract description 47
- 239000002245 particle Substances 0.000 claims abstract description 33
- 238000012880 independent component analysis Methods 0.000 claims abstract description 26
- 238000000605 extraction Methods 0.000 claims abstract description 10
- 238000011156 evaluation Methods 0.000 claims description 10
- 238000012937 correction Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 3
- 239000012535 impurity Substances 0.000 claims description 3
- 239000004615 ingredient Substances 0.000 claims description 2
- 238000013075 data extraction Methods 0.000 claims 1
- 238000005316 response function Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 4
- 238000004422 calculation algorithm Methods 0.000 description 12
- 230000006870 function Effects 0.000 description 10
- 238000005457 optimization Methods 0.000 description 10
- 238000007418 data mining Methods 0.000 description 6
- 230000005611 electricity Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000007812 deficiency Effects 0.000 description 4
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 201000004569 Blindness Diseases 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000007726 management method Methods 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
- 238000012806 monitoring device Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000000714 time series forecasting Methods 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"
-
- 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
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of electric power system data and excavates with prediction integration method, the combination based on independent component analysis and Support vector regression.This method includes input training sample data and carries out feature extraction using ICA；PSO is initialized, and sets accelerated factor c1 and c2, and current evolutionary generation is set to t by Inertia Weight w, maximum evolutionary generation Tmax, and S particle composition initial population X (t) is randomly generated in definition space Rn；Evaluate initial population X (t)；Termination condition, which meets then optimizing, to be terminated, and next step is gone to；The particle optimal location searched out i.e. optimized parameter vector C, σ, ε is assigned to SVR；It is trained with the training sample set pair SVR after ICA feature extractions, realizes the structure of prediction model；It is predicted using the prediction model and forecast sample established.The electric power system data that the present invention establishes excavates and predicts integration method, can reduce influence when interference signal builds prediction mode in predictive variable, improve the prediction effect of support vector regression.
Description
Technical field
The invention belongs to power system security stability control techniques fields, and in particular to a kind of electric power system data excavate with
Predict integration method, the combination based on independent component analysis and Support vector regression.
Background technology
Electric system has been achieved with flexibly depositing for a large amount of historical informations with the development of digitlization, automation and informationization
It takes, how therefrom to extract characteristic information and prediction variation preferably serves production run and becomes the pass that information resources efficiently use
Key.Being related to data mining and the field of prediction simultaneously in electric system has：Short-term load forecasting, long term load forecasting, short-term electricity
The forecasting of Gas Concentration etc. dissolved in valence prediction, electricity demand forecasting and transformer oil status predication, transformer.Wherein electric load
Prediction is an important routine work of electric power system dispatching and electricity consumption plan, is to formulate Transaction algorithm and the arrangement method of operation
Main Basiss.Currently, can all be recorded by various monitoring devices and management system with the values of the relevant various variables of load,
Storage forms large data record/database.How to be rejected from these complicated huge data wrong and useless
Data excavate the factor for lying in these data and really determining load, obtain the rule of this area's load variations, and then shape
The real accurate prediction of pairs of load is the core of load prediction work.
Data mining technology main purpose be picked out from data it is correct, novel, have potential utility value with
And it can finally become useful knowledge.The major technique of data mining includes with method：Statistical analysis technique, decision tree, people
Artificial neural networks, genetic algorithm, rough set method and support vector machines etc..Wherein there are many different data mining algorithms at present
For load prediction.
Support vector machines is to establish model with the principle of risk minimization based on Statistical Learning Theory.It is real
The minimum of existing empiric risk and fiducial range can also obtain good system to reach in the case where statistical sample amount is less
Count the purpose of rule.By support vector machines with return combine form support vector regression because it can catch it is mutual in variable
Dynamic relationship, need not have data too many it is assumed that being widely used in time series forecasting in recent years.Although supporting vector
Machine regression model is not required to too many it is assumed that predictive variable can be used arbitrarily, but be easy to cause prediction model over training or
The problem of undertrained.
Invention content
The present invention is directed to solve the technical problems existing in the prior art at least to some extent, a kind of electric system is proposed
Data mining is with prediction integration method, the combination based on independent component analysis and Support vector regression, can reduce prediction and become
Influence when interference signal builds prediction mode in amount improves the prediction effect of support vector regression.
In order to realize the object of the invention, according to a first aspect of the present invention, embodiment of the present invention provides a kind of electric system
Data mining and prediction integration method, specifically use following technical scheme：
Specifically, this method comprises the following steps：
Step S101 inputs training sample data, is carried out using Independent Component Analysis (ICA) to training sample data
Feature extraction；
Step S102, particle group optimizing (PSO) initialization, setting accelerated factor c1 and c2, Inertia Weight w, maximum are evolved
Current evolutionary generation is set to t by algebraically Tmax, S particle x1, x2 ..., xs is randomly generated in definition space Rn, composition is just
Beginning population X (t)；
Step S103, evaluation initial population X (t)；
Step S104, checks whether termination condition meets, if satisfied, then optimizing terminates, goes to step S105；
Step S105, the particle optimal location that will be searched out, i.e. optimized parameter vector C, σ, ε are assigned to support vector regression
(SVR), wherein C is correction factor, and σ is nuclear parameter, and ε is prediction error；
Step S106, with the training sample set pair supporting vector after Independent Component Analysis (ICA) feature extraction
It returns (SVR) to be trained, realizes the structure of prediction model；
Step S107 is predicted using the prediction model and forecast sample established.
Further, the step S102 further includes：The initial velocity v1, v2 ..., vs of each particle are randomly generated, is formed
Rate matrices V (t).
Further, the step S103 is specifically included：It defines fitness function and carries out population evaluation, the fitness of particle
Functional value is bigger, then particle position is better.
Further, the fitness function is defined as： yiRespectively support vector regression
(SVR) training output valve and desired output.
Further, wherein the termination condition of step S104 is that optimizing reaches maximum evolutionary generation Tmax or evaluation of estimate
Less than given accuracy.
Further, the step S104 further includes：If the termination condition is unsatisfactory for, t=t+1, more new particle
Speed and position generate new population X (t+1), and go to step S103.
Further, the feature extraction in the step S101 specifically includes：Independent element is used to training sample data
Analysis method (ICA) is then excavated from the independent element isolated to isolate independent element and represents the only of impurity signal
Vertical ingredient, and remaining independent element is rebuild to obtain filtering out the training sample data after interference.
Further, the training sample set in the step S106 is the training sample data filtered out after interference.
Compared with prior art, the method for the above embodiment is returned due to introducing independent component analysis and supporting vector
Return, can reduce influence when interference signal builds prediction mode in predictive variable, when reducing Support vector regression modeling because
Being interfered in by sample is influenced and leads to the problem of over-fitting or be fitted deficiency, and then improves Support vector regression prediction result
Accuracy；Since support vector regression is longer to the training time of higher-dimension training sample data, using independent component analysis side
Method pre-processes sample, can also reduce the dimension of training sample and reduce the training time；The prediction of support vector regression
The problem of quality and Model Parameter are set with sizable relationship, therefore when using support vector regression, most critical be as
The parameter value of what preference pattern is to guarantee to obtain good prediction result；Particle swarm optimization algorithm is a kind of emerging based on group
The Stochastic Optimization Algorithms of intelligence, are simply easily achieved, and have stronger global optimization ability, using particle cluster algorithm to supporting
The parameter of vector regression model optimizes, and improves the prediction effect of support vector regression, overcomes prediction model parameters choosing
The blindness selected and the time for reducing parameter training.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Obtain other attached drawings according to these attached drawings.
Fig. 1 is the flow chart that a kind of electric power system data that embodiment of the present invention provides excavates and predicts integration method.
Specific implementation mode
In being described below, for illustration and not for limitation, it is proposed that such as tool of particular system structure, technology etc
Body details understands the embodiment of the present invention to cut thoroughly.However, it will be clear to one skilled in the art that there is no these specific
The present invention can also be realized in the other embodiments of details.In other situations, it omits to well-known system, device, electricity
The detailed description of road and method, in case unnecessary details interferes description of the invention.
In order to illustrate technical solutions according to the invention, illustrated below by specific implementation mode combination attached drawing.
A kind of electric power system data of an embodiment of the invention offer excavates and predicts integration method, based on independently
The combination of constituent analysis and Support vector regression, wherein the electric power system data is excavated with the flow of prediction integration method such as
Shown in Fig. 1.
Specifically, this approach includes the following steps S101~S107：
The step S101 of present embodiment includes：Training sample data are inputted, independent element is used to training sample data
Analysis method (ICA) carries out feature extraction；
In embodiments of the present invention, it should be noted that the electric power system data includes but not limited to electric system
Short term, long-term load, short-term electricity price, the gas concentration etc. dissolved in electricity consumption and transformer oil, load forecast
Be in power planning and system operation it is very important one work, heavy workload, and need be repeated, it usually according to
It,, can be according to reality in specific implementation process if weather and load data early period establish prediction model according to certain information
Electric power system data needed for the selection of border technical need, other unmentioned related datas in present embodiment, herein not one by one
It repeats.
Feature extraction in step S101 specifically includes：Independent Component Analysis (ICA) is used to training sample data
To isolate independent element, the independent element for representing impurity signal is then excavated from the independent element isolated, and will remain
Remaining independent element is rebuild to obtain filtering out the training sample data after interference.
In embodiments of the present invention, introduce Independent Component Analysis, can reduce Support vector regression modeling when because
Being interfered in by sample is influenced and leads to the problem of over-fitting or be fitted deficiency, and dimension and the reduction of training sample can also be reduced
Training time, and then improve the accuracy of Support vector regression prediction result.
The step S102 of present embodiment method includes：Particle group optimizing (PSO) initialize, setting accelerated factor c1 and
Current evolutionary generation is set to t, S grain is randomly generated in definition space Rn by c2, Inertia Weight w, maximum evolutionary generation Tmax
Sub- x1, x2 ..., xs, composition initial population X (t).
Further include the initial velocity v1, v2 ..., vs for randomly generating each particle in wherein step S102, forms rate matrices
V(t)。
In embodiments of the present invention, introduce particle swarm optimization algorithm, particle swarm optimization algorithm be it is a kind of it is emerging based on
The Stochastic Optimization Algorithms of colony intelligence, are simply easily achieved, and have stronger global optimization ability.
The step S103 of present embodiment is evaluation initial population X (t).It obtains terminating item by evaluating population
Part finds particle optimal location.
The step S103 of embodiment of the present invention is specifically included：It defines fitness function and carries out population evaluation, particle is fitted
Response functional value is bigger, then particle position is better.The purpose for defining fitness function evaluation population is sought by the result of quantization
Look for accurate optimum particle position.
Wherein, the fitness function is defined as： yiRespectively support vector regression (SVR)
Training output valve and desired output.By the function it is found that working as the training output valve and desired output of support vector regression (SVR)
When difference minimum between value, the functional value is maximum, also this means that the position of particle is better.
The step S104 of present embodiment includes：It checks whether termination condition meets, if satisfied, then optimizing terminates, goes to
Step S105.It needs to carry out end judgement after evaluating population, needs to stop optimizing if meeting termination condition.Its
Described in termination condition be that optimizing reaches maximum evolutionary generation Tmax or evaluation of estimate and is less than given accuracy.
Wherein step S104 further includes：If the termination condition is unsatisfactory for, t=t+1, the speed of more new particle and position
It sets, generates new population X (t+1), and go to step S103.When termination condition is unsatisfactory for, speed and the position of more new particle are needed
It sets to establish new population, and re-starts optimizing, until finding optimal value, meet termination condition.
The step S105 of present embodiment includes：The particle optimal location that will be searched out, i.e. optimized parameter vector C, σ, ε are assigned
It is correction factor to give support vector regression (SVR), wherein C, and σ is nuclear parameter, and ε is prediction error.
Parameter is set with sizable relationship, therefore use in the prediction quality and pattern of support vector regression (SVR)
It is how to go the value of selection parameter the problem of most critical when SVR to guarantee to obtain good prediction result.Correction factor C, core
The selection of parameter σ and prediction error ε influences the SVR precision estimated very big.According to the characteristic of sample data, correction factor C
Determine the complexity of model and the punishment degree to the fitness bias more than ε.C values excessive (be more than 100) too small (are less than
10) all it can learn because crossing or owe to learn that the Generalization Capability of system is made to be deteriorated.The σ explications knot of high-dimensional feature space φ (x)
Structure, thus the complexity of last solution is controlled, same σ values excessive (being more than 10) or too small (being less than 1) can all make the extensive of system
Degradation.ε shows expectation of the system to estimation function error in sample data, and ε values are bigger, and supporting vector number is fewer,
The expression of solution is more sparse, but big ε can also reduce the precision of regression estimates.The present invention is using particle group optimizing (PSO) to ginseng
Number optimizes, and overcomes method-commonly used at present and determines that the needs present in SVR parameters are anti-by cross validation tentative calculation
Retrial tests, takes longer defect.
The step S106 of present embodiment includes：With the training after Independent Component Analysis (ICA) feature extraction
Sample set is trained support vector regression (SVR), realizes the structure of prediction model.The training sample in step S106
Collection is the training sample data filtered out after interference.
In embodiments of the present invention, introduce Independent Component Analysis, can reduce Support vector regression modeling when because
Being interfered in by sample is influenced and leads to the problem of over-fitting or be fitted deficiency, and dimension and the reduction of training sample can also be reduced
Training time, and then improve the accuracy of Support vector regression prediction result.
The step S107 of present embodiment includes：It is predicted using the prediction model and forecast sample established.By
Step S101 to step S106 has obtained the training sample data after rejecting interference, and has established prediction model, according to training sample
Notebook data and prediction model can carry out final data prediction.
The present invention carries out daily load prediction by sample of EUNITE historical loads.The input quantity of Selection Model is 18, point
It Wei not day to be measured and 1 day highest, minimum temperature a few days ago to be measured；1 day maximum, minimum and average load a few days ago to be measured, it is to be measured
Take within 1 day a few days ago and first 2 days 5 load values centered on prediction time respectively, upper one year load value in the same time.Output quantity 1 is
The load value at moment day to be measured.Using totally two years 35040 data samples on December 31,1 day to 1998 January in 1997,
Wherein training sample set is divided into totally 17520 samples on December 31,1 day to 1997 January in 1997, separately retains January 1 in 1998
Day on December 31st, 1998 totally 17520 samples as test sample collection.
In simulations, population scale is set as 200；Solution space is 3 dimensions (corresponding to C, σ and ε respectively), and Inertia Weight w's is first
Initial value is 0.9, maximum evolutionary generation TmaxIt is 150, accelerated factor c1=c2=1；C value ranges are [0.1,100], σ value models
It encloses for [0.1,10], ε value ranges are [0,1], and the speed maximum value vector of corresponding (C, σ, ε) is (0.5,0.1,0.1).
The prediction result of the different independent element numbers of table 1
Independent element number it is excessive or it is very few can all precision of prediction be made to reduce because independent element contains when number is excessive
Much noise signal；Independent element is lost the characteristic information of too many input variable when very few.In ginsengs such as C, σ and ε of SVR models
It under conditions of number immobilizes, investigates under same training sample, the influence of the numbers of different independent elements to ICA-SVR models,
Its prediction result is shown in Table 1.Simulation result shows：When number for the training sample independent element is 9, prediction essence
Degree is best.
In order to verify the estimated performance for illustrating ICA-SVR methods, by itself and the prediction side for directly using SVR (standard SVR)
Method is compared.ICA-SVR methods are better than standard SVR methods in terms of accuracy and speed.
From the above, it can be seen that the method for embodiment of the present invention due to introduce independent component analysis and support to
Amount returns, and can reduce influence when interference signal builds prediction mode in predictive variable, reduces Support vector regression modeling
Shi Yin is interfered in by sample to be influenced and leads to the problem of over-fitting or be fitted deficiency, and then improves Support vector regression prediction
As a result accuracy；Since support vector regression is longer to the training time of higher-dimension training sample data, using independent component point
Analysis method pre-processes sample, can also reduce the dimension of training sample and reduce the training time；Support vector regression
The problem of prediction quality and Model Parameter are set with sizable relationship, therefore when using support vector regression, most critical
It is how the parameter value of preference pattern is to guarantee to obtain good prediction result；Particle swarm optimization algorithm is a kind of emerging base
It in the Stochastic Optimization Algorithms of colony intelligence, is simply easily achieved, and there is stronger global optimization ability, using particle cluster algorithm pair
The parameter of support vector regression model optimizes, and improves the prediction effect of support vector regression, overcomes prediction model ginseng
The blindness of number selection and the time for reducing parameter training.
Part not deployed in its method in embodiment of the present invention can refer to the correspondence portion of the method for embodiment of above
Point, it is no longer developed in details herein.
In the description of this specification, reference term " embodiment ", " some embodiments ", " schematically implementation
The description of mode ", " example ", " specific example " or " some examples " etc. means the tool described in conjunction with the embodiment or example
Body characteristics, structure, material or feature are contained at least one embodiment or example of the present invention.In the present specification,
Schematic expression of the above terms are not necessarily referring to identical embodiment or example.Moreover, the specific features of description, knot
Structure, material or feature can be combined in any suitable manner in any one or more embodiments or example.
Any process described otherwise above or method description are construed as in flow chart or herein, and expression includes
It is one or more for realizing specific logical function or process the step of executable instruction code module, segment or portion
Point, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discuss suitable
Sequence, include according to involved function by it is basic simultaneously in the way of or in the opposite order, to execute function, this should be of the invention
Embodiment person of ordinary skill in the field understood.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art within the scope of the invention can be to above-mentioned
Embodiment is changed, changes, replacing and modification.
Claims (8)
1. a kind of electric power system data excavates and prediction integration method, which is characterized in that this method comprises the following steps：
Step S101 inputs training sample data, and feature is carried out using Independent Component Analysis (ICA) to training sample data
Extraction；
Step S102, particle group optimizing (PSO) initialization, sets accelerated factor c1 and c2, Inertia Weight w, maximum evolutionary generation
Current evolutionary generation is set to t by Tmax, and S particle x1, x2 ..., xs are randomly generated in definition space Rn, forms initial kind
Group X (t)；
Step S103, evaluation initial population X (t)；
Step S104, checks whether termination condition meets, if satisfied, then optimizing terminates, goes to step S105；
Step S105, the particle optimal location that will be searched out, i.e. optimized parameter vector C, σ, ε are assigned to support vector regression (SVR),
Wherein C is correction factor, and σ is nuclear parameter, and ε is prediction error；
Step S106, with the training sample set pair support vector regression after Independent Component Analysis (ICA) feature extraction
(SVR) it is trained, realizes the structure of prediction model；
Step S107 is predicted using the prediction model and forecast sample established.
2. a kind of electric power system data according to claim 1 excavates and prediction integration method, which is characterized in that the step
Suddenly S102 further includes：
Randomly generate the initial velocity v1, v2 ..., vs of each particle, composition rate matrices V (t).
3. a kind of electric power system data according to claim 1 excavates and prediction integration method, which is characterized in that the step
Rapid S103 is specifically included：
It defines fitness function and carries out population evaluation, the fitness function value of particle is bigger, then particle position is better.
4. a kind of electric power system data according to claim 3 excavates and prediction integration method, which is characterized in that described suitable
Response function is defined as：
yiRespectively support vector regression (SVR) training output valve and desired output.
5. a kind of electric power system data according to claim 1 excavates and prediction integration method, which is characterized in that wherein walk
The termination condition of rapid S104 is that optimizing reaches maximum evolutionary generation Tmax or evaluation of estimate is less than given accuracy.
6. a kind of electric power system data according to claim 1 excavates and prediction integration method, which is characterized in that the step
Suddenly S104 further includes：
If the termination condition is unsatisfactory for, t=t+1, the speed of more new particle and position generate new population X (t+1), and turn
To step S103.
7. a kind of electric power system data according to claim 1 excavates and prediction integration method, which is characterized in that the step
Feature extraction in rapid S101 specifically includes：
It is then only from what is isolated to training sample data using Independent Component Analysis (ICA) to isolate independent element
The independent element for representing impurity signal is excavated in vertical ingredient, and remaining independent element is rebuild to obtain filtering out interference
Training sample data afterwards.
8. a kind of electric power system data according to claim 1 excavates and prediction integration method, which is characterized in that the step
The training sample set in rapid S106 is the training sample data filtered out after interference.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810333780.XA CN108596781A (en) | 2018-04-13 | 2018-04-13 | A kind of electric power system data excavates and prediction integration method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810333780.XA CN108596781A (en) | 2018-04-13 | 2018-04-13 | A kind of electric power system data excavates and prediction integration method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108596781A true CN108596781A (en) | 2018-09-28 |
Family
ID=63622322
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810333780.XA Pending CN108596781A (en) | 2018-04-13 | 2018-04-13 | A kind of electric power system data excavates and prediction integration method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596781A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109408698A (en) * | 2018-10-31 | 2019-03-01 | 辽宁工程技术大学 | System and application method are supported in intelligent financial report based on data mining technology |
| CN109635994A (en) * | 2018-10-23 | 2019-04-16 | 广东精点数据科技股份有限公司 | A kind of crop yield prediction technique for realizing the multi-source heterogeneous fusion of influence factor |
-
2018
- 2018-04-13 CN CN201810333780.XA patent/CN108596781A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635994A (en) * | 2018-10-23 | 2019-04-16 | 广东精点数据科技股份有限公司 | A kind of crop yield prediction technique for realizing the multi-source heterogeneous fusion of influence factor |
| CN109408698A (en) * | 2018-10-31 | 2019-03-01 | 辽宁工程技术大学 | System and application method are supported in intelligent financial report based on data mining technology |
| CN109408698B (en) * | 2018-10-31 | 2022-01-14 | 辽宁工程技术大学 | Intelligent financial report support system based on data mining technology and use method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109508360B (en) | Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton | |
| CN105631483A (en) | Method and device for predicting short-term power load | |
| CN106022614A (en) | Data mining method of neural network based on nearest neighbor clustering | |
| Chen et al. | Groundwater level prediction using SOM-RBFN multisite model | |
| CN110135635A (en) | A kind of region electric power saturation load forecasting method and system | |
| CN113554466A (en) | Short-term power consumption prediction model construction method, prediction method and device | |
| CN111489259A (en) | Stock market risk prediction intelligent implementation method based on deep learning | |
| Solomatine | Applications of data-driven modelling and machine learning in control of water resources | |
| CN104732091A (en) | Cellular automaton river bed evolution prediction method based on natural selection ant colony algorithm | |
| CN112364560A (en) | Intelligent prediction method for working hours of mine rock drilling equipment | |
| CN113570132A (en) | Wind power prediction method for space-time meteorological feature extraction and deep learning | |
| CN113033081A (en) | Runoff simulation method and system based on SOM-BPNN model | |
| CN108596781A (en) | A kind of electric power system data excavates and prediction integration method | |
| CN113361774A (en) | Ecological space intensive utilization optimization method and system | |
| CN116960962A (en) | Mid-long term area load prediction method for cross-area data fusion | |
| Purwinarko et al. | Crude oil price prediction using Artificial Neural Network-Backpropagation (ANN-BP) and Particle Swarm Optimization (PSO) methods | |
| CN113537553A (en) | Regional power distribution network load demand hybrid prediction method | |
| CN115423146A (en) | Self-adaptive runoff forecasting method based on multi-factor nearest neighbor sampling regression and support vector machine | |
| CN115271154A (en) | Nonlinear regression flood element prediction method based on polynomial and partial least square coupling | |
| CN114372615A (en) | Short-term photovoltaic power prediction method and system considering spatial correlation | |
| CN114330818A (en) | Dynamic water demand prediction method based on main driving factor screening and deep learning | |
| CN114444763A (en) | Wind power prediction method based on AFSA-GNN | |
| CN113656707A (en) | Financing product recommendation method, system, storage medium and equipment | |
| CN113095466A (en) | Algorithm of satisfiability model theoretical solver based on meta-learning model | |
| CN112270123A (en) | Basin reservoir group runoff random generation method based on convolution generation countermeasure network |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180928 |
|
| RJ01 | Rejection of invention patent application after publication |