CN108062720A - A kind of load forecasting method based on similar day selection and random forests algorithm - Google Patents
A kind of load forecasting method based on similar day selection and random forests algorithm Download PDFInfo
- Publication number
- CN108062720A CN108062720A CN201711263874.6A CN201711263874A CN108062720A CN 108062720 A CN108062720 A CN 108062720A CN 201711263874 A CN201711263874 A CN 201711263874A CN 108062720 A CN108062720 A CN 108062720A
- Authority
- CN
- China
- Prior art keywords
- day
- similarity
- factor
- load
- sample
- 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
- 238000007637 random forest analysis Methods 0.000 title claims abstract description 30
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 21
- 238000013277 forecasting method Methods 0.000 title claims abstract description 9
- 238000012549 training Methods 0.000 claims abstract description 10
- 238000000034 method Methods 0.000 claims description 25
- 238000001556 precipitation Methods 0.000 claims description 11
- 238000005070 sampling Methods 0.000 claims description 9
- 238000003066 decision tree Methods 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 8
- 239000012535 impurity Substances 0.000 claims description 5
- 238000011835 investigation Methods 0.000 claims description 3
- 238000013138 pruning Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 239000013598 vector Substances 0.000 abstract description 4
- 238000004378 air conditioning Methods 0.000 description 5
- 230000011218 segmentation Effects 0.000 description 5
- 230000007423 decrease Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 238000010438 heat treatment Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 235000006508 Nelumbo nucifera Nutrition 0.000 description 2
- 240000002853 Nelumbo nucifera Species 0.000 description 2
- 235000006510 Nelumbo pentapetala Nutrition 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 206010009866 Cold sweat Diseases 0.000 description 1
- 108010074506 Transfer Factor Proteins 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- YEHCICAEULNIGD-MZMPZRCHSA-N pergolide Chemical compound C1=CC([C@H]2C[C@@H](CSC)CN([C@@H]2C2)CCC)=C3C2=CNC3=C1 YEHCICAEULNIGD-MZMPZRCHSA-N 0.000 description 1
- 229940088507 permax Drugs 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000003860 storage 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
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Business, Economics & Management (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- Marketing (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Probability & Statistics with Applications (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of load forecasting methods based on similar day selection and random forests algorithm, are by using similar day Algorithms of Selecting and random forests algorithm, load forecasting model is obtained using training sample set, so as to fulfill the load prediction to predicting day.It comprises the concrete steps that, calculate history day first and predicts the similarity between day, realize pretreatment to receiving money sample, the higher historical sample of similarity is chosen, so as to obtain similar day sample set；Then random forest regression model is trained, the feature vector for predicting day is inputted in trained model, the average value of all regression tree output results is taken as final load prediction results, so as to fulfill the load prediction to predicting day.
Description
Technical field
The present invention relates to Load characteristics index prediction, more particularly to it is a kind of chosen based on similar day and random forest it is negative
Lotus characteristic index Forecasting Methodology.
Background technology
Electric system is a large-scale electrical energy production, conveying, distribution, the network of consumption, and electric energy is unable to mass storage,
There is instantaneity.Holding electrical energy production and the balance of consumption are the primary conditions of power system stability operation, and load
Prediction is to ensure the important step of this balance.
In recent years, with the further opening of electricity market, power generation and consumption are more market-oriented, and load prediction is existing
It is more shown for the effect in electric power system dispatching and management, therefore higher is proposed to the accuracy of Load Prediction In Power Systems
Requirement.China experienced the attention degree of load prediction one longer process, to promote holding for entire electricity market
Supervention exhibition, load prediction are also the necessary requirement that China realizes electricity market, have important theory significance and practical value.For
Comply with the trend in epoch, it is necessary to deeper into the principle and method of studying intensively load prediction, improve load forecasting method, improve load
The precision of prediction.
The content of the invention
For more than technical problem, the present invention proposes a kind of Load characteristics index based on similar day selection and random forest
Forecasting Methodology obtains similar day sample, so as to be realized to historical data by similar day Algorithms of Selecting from historical sample concentration
Pretreatment, and then be trained using random forest, build load forecasting model.
A kind of load forecasting model based on similar day selection and random forests algorithm, comprises the following steps：
S110, data are collected according to investigation on the spot, from economic factor, climatic factor, time factor, geographic factor etc.
An important factor for setting out, choosing temperature, humidity, precipitation, day type etc. multiple influence load variations；
S120, according to selected load variations influence factor, determine the similarity calculating method of each influence factor；
S130, the similarity based method obtained according to different affecting factors calculate each history day to predicting the similar of day respectively
Degree builds similarity matrix；
S140, according to similarity matrix, the similarity of history day each factor is multiplied, calculates total similarity.History day
Total similarity is represented with Fd；
S150, the threshold value for determining total similarity choose the sample composition similar day sample set that similarity is more than threshold value；
The quantity K of S160, the initial regression tree of setting extract K son using Bootstrap double sampling methods from S sample
Sample set, generates K regression tree to greatest extent, and beta pruning is not required in the process of generation；
S170, the sample set obtained according to double sampling, while K decision tree of training, choose r characteristic factor as current
The disruptive features collection of node, r are less than or equal to log2 (H+1)；
S180, the division that node is carried out according to Geordie impurity level minimum criteria；
S190, using OOB data as test sample, estimation error is carried out to random forest regression model, is missed according to prediction
The quantity K of regression tree in difference adjustment model；
S200, according to input prediction day sample, the average value of each regression tree output is taken to can obtain final prediction knot
Fruit.
Further, influencing the factor of load variations includes following 6 aspects：(1) maximum temperature；(2) minimum temperature；
(3) humidity；(4) precipitation；(5) day type；(6) distance on date；
Further, total similarity calculating method is by the way of accumulation.
Further, random forest node split algorithm carries out the division of node using Geordie impurity level minimum criteria.
Further, the model evaluation method uses OOB false segmentation rates, and the false segmentation rate of all trees is averaged to obtain OOB
False segmentation rate, it is possible to obtain an OOB estimation error.
A kind of load forecasting method based on similar day selection and random forests algorithm of the present invention is by using phase
Like day Algorithms of Selecting and random forests algorithm, load forecasting model is obtained using training sample set, so as to fulfill to prediction day
Load prediction.First calculate history day and predict day between similarity, to receive money sample realize pretreatment, choose similarity compared with
High historical sample, so as to obtain similar day sample set.Then random forest regression model is trained, will predicts the spy of day
Sign vector is inputted in trained model, takes the average value of all regression trees output results as final load prediction results,
So as to fulfill the load prediction to predicting day.
Compared with the prior art, the invention has the beneficial effects that：Similar day can be carried out from multiple influence factors
Selection；The parameter that random forests algorithm needs are adjusted is less, gives training set sample, characteristic quantity number of random forest etc. really
After fixed, it is only necessary to determine the number of decision tree；It is not easy over-fitting occur and the fitting precision of algorithm is also higher,
And the mathematical definition of random forest and the structural property of decision tree ensure that convergence.
Description of the drawings
Fig. 1 is the method flow diagram of the load forecasting method based on similar day selection and random forests algorithm of the present invention；
Fig. 2 is random forest structure chart；
Fig. 3 is influence coefficient correspondence figure of the temperature to load.
Specific embodiment
As shown in Figure 1, Figure 2 and Fig. 3, the shown present invention be related to it is a kind of chosen based on similar day and random forests algorithm it is negative
Lotus Forecasting Methodology is by using similar day Algorithms of Selecting and random forests algorithm, load prediction is obtained using training sample set
Model, so as to fulfill the load prediction to predicting day.The embodiment calculates history day and predicts the similarity between day, to receiving
It provides sample and realizes pretreatment, the higher historical sample of similarity is chosen, so as to obtain similar day sample set.Then to random forest
Regression model is trained, and the feature vector for predicting day is inputted in trained model, takes all regression tree output results
Average value is as final load prediction results, so as to fulfill the load prediction to predicting day.
In wherein a kind of embodiment, a kind of load based on similar day selection and random forests algorithm that the present invention relates to is pre-
Survey method comprises the following steps：
Step S110 collects data, from economic factor, climatic factor, time factor, geographic factor etc. according to investigation on the spot
Aspect is set out, an important factor for choosing temperature, humidity, precipitation, day type etc. multiple influence load variations.Above-mentioned influence factor
In, temperature needs to consider per max. daily temperature and minimum temperature.Day type refers to working day or weekend, if for special day etc.
Deng.
Step S120 according to selected load variations influence factor, determines the similarity calculation side of each influence factor
Method, above-mentioned influence factor can be divided into：(1) maximum temperature；(2) minimum temperature；(3) humidity；(4) precipitation；(5) day type；(6)
The distance on date；The similarity calculating method of different affecting factors, is described in detail below：
(1) maximum temperature
Influence of the temperature to load is mainly reflected in the use aspect of air-conditioning equipment, passes through the adjusting of air-conditioning equipment so that
The comfort level of human body reaches optimal.When temperature is relatively low, air-conditioning is presented as heating load, and temperature is lower, and load is higher；Work as temperature
When higher, air-conditioning is presented as temperature-lowering load, and temperature is higher, and load is higher.Therefore influence of the temperature to load be rendered as it is non-thread
Property, temperature influences coefficient as shown in figure 3, using the mapping table similar to Fig. 3 to load, and the highest temperature is mapped as one and is reflected
Value is penetrated, similarity is calculated according to mapping value.
(2) minimum temperature
Mainly in winter, the lowest temperature is lower for influence of the lowest temperature to load, and the proportion of air conditioner load rises.But winter
Air conditioner load is significantly lower than summer, and the mode of winter heating is not limited to air-conditioning, more effective using coal fired central heating, because
Influence of this minimum temperature to load again may be by obtaining similar to the mapping table of Fig. 1, but need reasonably to turn down minimum
Temperature is to the size of the Intrusion Index of load.
(3) humidity
When weather is in clammy state, " coldness index " will be caused to rise, human comfort can decline, so as to cause sky
Load is adjusted to rise；When temperature is placed in the middle, the influence of humidity is smaller, unobvious.When temperature is in damp and hot state, similary human body relaxes
Appropriateness can decline, so as to cause the apparent rising of air conditioner load.Consideration more than phenomenon, when assessing the similarity of humidity factor,
It needs to consider the influence of temperature, determines influence coefficient of the different humidity to load, similarity is calculated further according to coefficient is influenced.
(4) precipitation
Influence of the precipitation to load is mainly reflected in the indirect influence for temperature and humidity.When precipitation is larger and holds
When continuous time is longer, humidity increases, and temperature decreases, so as to drive the rise of air conditioner load indirectly.But when precipitation is larger,
Ye Huishi small power stations, which contribute, to be increased, and network for the load reduces.Therefore, it is necessary first to determine direction of the precipitation to loading effects, according to
Rainfall size sorts, and determines degree of the various rainfall forms to loading effects successively.
(5) day type
Day type can be divided into working day or weekend or special day.The shape of the load curve at general work day and weekend is
There is significant difference, similarly for some special days, apparent movement can also occur for load curve.If day, type was identical, it is somebody's turn to do
Factor maximum similarity 1.For special day, it is necessary to consider the case when：It it is all special day, and the date is identical；It is all special day,
But the date is different；One is special day, and one is no special day；It is no special day.
(6) distance on date
The far and near size for also influencing similarity away from prediction day.Generally, range prediction day is nearer, and similarity is higher.Its phase
It is as follows like degree calculation formula：
Wherein pidFor the similarity of d days dates distance factor；γ is attenuation coefficient；A is the minimum similarity of the factor.
Step S130, the similarity based method obtained according to different affecting factors calculate each history day with predicting day respectively
Similarity builds similarity matrix.Assuming that considering M kind factors, D history day, the similarity matrix of acquisition is X, specific to represent
Method is as follows：
X=[Pid] i=1,2,3..., M；D=1,2,3...D
Wherein X is similarity matrix；M is the number of influence factor；D is the number of history day.Pid each history days are each
The similarity of influence factor.
Step S140 according to similarity matrix, the similarity of history day each factor is multiplied, calculates total similarity.History
Day total similarity represented with Fd, circular such as following formula：
Step S150 determines the threshold value of total similarity, chooses the sample composition similar day sample set that similarity is more than threshold value.
λ is the threshold value of total similarity, chooses Fd >=λ and forms similar day sample set, the number of similar day is S.
Step S160 sets the quantity K of initial regression tree, and K is extracted from S sample using Bootstrap double sampling methods
A sub- sample set, generates K regression tree to greatest extent, and beta pruning is not required in the process of generation.In order to which decision tree is made not generate office
Portion's optimal solution, for random forest using there is the Bootstrap double sampling methods put back to, random sampling technology generates K training sample
Set.
Step S170 according to the sample set that double sampling obtains, while trains K decision tree, chooses r characteristic factor conduct
The disruptive features collection of present node, r are less than or equal to log2 (H+1)；Decision tree can be random in all features in node split
A selection part is used to calculate optimal Split Attribute, it is impossible to participate in whole characteristic attributes in calculating；R is usually chosen to be less than or equal to
Log2 (H+1) or n=√ H, wherein H are the dimension of sampling feature vectors.Here it is 6 to choose H, therefore r=3.
Step S180 carries out the division of node according to Geordie impurity level minimum criteria, is described in detail below：
The calculating of Gini impurity level minimum criterias is as follows：
P (j | t) represent the probability that classification is j on node t, when all samples of node t belong to same class, gini index
Minimum value 0 is got, at this time sample class purity highest；When Gini indexes are maximized 1, the sample class of present node is in
Purity is minimum.
The Gini indexes of present node are
Wherein z is the number of child node, and ni is the sample size at child node i, and n is the number of samples at upper layer node.
The Gini indexes of each attribute are calculated in training process, and select one there is the variable of minimum Gini indexs to present node
Into line splitting, until reaching stop condition.
Step S190 using OOB data as test sample, carries out estimation error, according to pre- to random forest regression model
The quantity K of regression tree in error transfer factor model is surveyed, random forest is the generation that collection is trained using Bootstrap methods,
When generating these data sets, initial training, which is concentrated with some samples, to be extracted, and be referred to as the outer data of bag, i.e.,
OOB data in units of each decision tree, utilize the set for all training sample points do not chosen by the forest, statistics
The false segmentation rate of all trees is averaged to obtain OOB points of rate, it is possible to obtain an OOB estimation error by the OOB false segmentation rates of the tree.
Step S200, according to input prediction day sample, take each regression tree output average value can obtain it is final pre-
Survey result.
Embodiment described above only expresses the several embodiments of the present invention, and description is more specific and detailed, but simultaneously
Cannot limiting the scope of the invention therefore be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to the guarantor of the present invention
Protect scope.Therefore, protection scope of the present invention should be determined by the appended claims.
Claims (3)
1. a kind of load forecasting method based on similar day selection and random forests algorithm, which is characterized in that comprise the following steps：
S110, data are collected according to investigation on the spot, from economic factor, climatic factor, time factor, geographic factor aspect,
Choose temperature, humidity, precipitation, the day multiple factors for influencing load variations of type；
S120, according to selected load variations influence factor, determine the similarity calculating method of each influence factor；
S130, the similarity based method obtained according to different affecting factors calculate each history day with predicting the similarity of day, structure respectively
Build similarity matrix；
S140, according to similarity matrix, the similarity of history day each factor is multiplied, calculates total similarity.Total phase of history day
It is represented like degree with Fd；
S150, the threshold value for determining total similarity choose the sample composition similar day sample set that similarity is more than threshold value；
The quantity K of S160, the initial regression tree of setting, K subsample is extracted using Bootstrap double sampling methods from S sample
Collection, generates K regression tree to greatest extent, and beta pruning is not required in the process of generation；
S170, the sample set obtained according to double sampling, while K decision tree of training, choose r characteristic factor as present node
Disruptive features collection, r be less than or equal to log2 (H+1)；
S180, the division that node is carried out according to Geordie impurity level minimum criteria；
S190, using OOB data as test sample, estimation error is carried out to random forest regression model, according to prediction error tune
The quantity K of regression tree in integral mould；
S200, according to input prediction day sample, the average value of each regression tree output is taken to can obtain final prediction result.
2. the load forecasting method according to claim 1 based on similar day selection and random forests algorithm, feature exist
In the day type refers to working day or weekend, if be special day.
3. the load forecasting method according to claim 1 based on similar day selection and random forests algorithm, feature exist
In above-mentioned influence factor can be divided into：Maximum temperature, minimum temperature, humidity, precipitation, day type and the distance on date.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711263874.6A CN108062720A (en) | 2017-12-05 | 2017-12-05 | A kind of load forecasting method based on similar day selection and random forests algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711263874.6A CN108062720A (en) | 2017-12-05 | 2017-12-05 | A kind of load forecasting method based on similar day selection and random forests algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108062720A true CN108062720A (en) | 2018-05-22 |
Family
ID=62135947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711263874.6A Pending CN108062720A (en) | 2017-12-05 | 2017-12-05 | A kind of load forecasting method based on similar day selection and random forests algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108062720A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537683A (en) * | 2018-04-13 | 2018-09-14 | 贵州电网有限责任公司 | A kind of load forecasting method based on similar day selection and random forests algorithm |
| CN109283378A (en) * | 2018-08-30 | 2019-01-29 | 番禺珠江钢管（珠海）有限公司 | A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium |
| CN109345011A (en) * | 2018-09-19 | 2019-02-15 | 中冶赛迪重庆信息技术有限公司 | A kind of Air-conditioning Load Prediction method and system returning forest based on depth |
| CN109409647A (en) * | 2018-09-10 | 2019-03-01 | 昆明理工大学 | A kind of analysis method of the salary level influence factor based on random forests algorithm |
| CN110135630A (en) * | 2019-04-25 | 2019-08-16 | 武汉数澎科技有限公司 | The short term needing forecasting method with multi-step optimization is returned based on random forest |
| CN111340281A (en) * | 2020-02-20 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Prediction model training method and device |
| CN112529262A (en) * | 2020-11-27 | 2021-03-19 | 北京京能高安屯燃气热电有限责任公司 | Short-term power prediction method, device, computer equipment and storage medium |
| CN113837459A (en) * | 2021-09-15 | 2021-12-24 | 浙江浙能技术研究院有限公司 | Intelligent power plant coal-fired power generator set short-term load prediction method based on RF-DTW |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105303262A (en) * | 2015-11-12 | 2016-02-03 | 河海大学 | Short period load prediction method based on kernel principle component analysis and random forest |
| CN105335800A (en) * | 2015-11-19 | 2016-02-17 | 国网天津市电力公司 | Method for forecasting electricity consumption of power consumers based on joint learning |
| CN105654196A (en) * | 2015-12-29 | 2016-06-08 | 中国电力科学研究院 | Adaptive load prediction selection method based on electric power big data |
| CN107016469A (en) * | 2017-04-13 | 2017-08-04 | 重庆大学 | Methods of electric load forecasting |
| CN107194600A (en) * | 2017-06-01 | 2017-09-22 | 国网山东省电力公司济南市历城区供电公司 | A kind of electric load Seasonal Characteristics sorting technique |
-
2017
- 2017-12-05 CN CN201711263874.6A patent/CN108062720A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105303262A (en) * | 2015-11-12 | 2016-02-03 | 河海大学 | Short period load prediction method based on kernel principle component analysis and random forest |
| CN105335800A (en) * | 2015-11-19 | 2016-02-17 | 国网天津市电力公司 | Method for forecasting electricity consumption of power consumers based on joint learning |
| CN105654196A (en) * | 2015-12-29 | 2016-06-08 | 中国电力科学研究院 | Adaptive load prediction selection method based on electric power big data |
| CN107016469A (en) * | 2017-04-13 | 2017-08-04 | 重庆大学 | Methods of electric load forecasting |
| CN107194600A (en) * | 2017-06-01 | 2017-09-22 | 国网山东省电力公司济南市历城区供电公司 | A kind of electric load Seasonal Characteristics sorting technique |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537683A (en) * | 2018-04-13 | 2018-09-14 | 贵州电网有限责任公司 | A kind of load forecasting method based on similar day selection and random forests algorithm |
| CN109283378A (en) * | 2018-08-30 | 2019-01-29 | 番禺珠江钢管（珠海）有限公司 | A kind of rotating arc welding is seamed into shape parameter detection method, system, device and medium |
| CN109409647A (en) * | 2018-09-10 | 2019-03-01 | 昆明理工大学 | A kind of analysis method of the salary level influence factor based on random forests algorithm |
| CN109345011A (en) * | 2018-09-19 | 2019-02-15 | 中冶赛迪重庆信息技术有限公司 | A kind of Air-conditioning Load Prediction method and system returning forest based on depth |
| CN110135630A (en) * | 2019-04-25 | 2019-08-16 | 武汉数澎科技有限公司 | The short term needing forecasting method with multi-step optimization is returned based on random forest |
| CN110135630B (en) * | 2019-04-25 | 2022-02-25 | 武汉数澎科技有限公司 | Short-term load demand prediction method based on random forest regression and multi-step optimization |
| CN111340281A (en) * | 2020-02-20 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Prediction model training method and device |
| CN111340281B (en) * | 2020-02-20 | 2022-06-03 | 支付宝(杭州)信息技术有限公司 | Prediction model training method and device |
| CN112529262A (en) * | 2020-11-27 | 2021-03-19 | 北京京能高安屯燃气热电有限责任公司 | Short-term power prediction method, device, computer equipment and storage medium |
| CN113837459A (en) * | 2021-09-15 | 2021-12-24 | 浙江浙能技术研究院有限公司 | Intelligent power plant coal-fired power generator set short-term load prediction method based on RF-DTW |
| CN113837459B (en) * | 2021-09-15 | 2024-05-03 | 浙江浙能数字科技有限公司 | RF-DTW-based short-term load prediction method for intelligent power plant coal-fired generator set |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108062720A (en) | A kind of load forecasting method based on similar day selection and random forests algorithm | |
| CN108537683A (en) | A kind of load forecasting method based on similar day selection and random forests algorithm | |
| CN108022001B (en) | Short-term load probability density prediction method based on PCA (principal component analysis) and quantile regression forest | |
| Sheng et al. | Short-term solar power forecasting based on weighted Gaussian process regression | |
| CN107516170B (en) | Difference self-healing control method based on equipment failure probability and power grid operation risk | |
| CN106485262B (en) | Bus load prediction method | |
| Yan et al. | Wind power grouping forecasts and its uncertainty analysis using optimized relevance vector machine | |
| CN106251001A (en) | A kind of based on the photovoltaic power Forecasting Methodology improving fuzzy clustering algorithm | |
| CN108921339B (en) | Quantile regression-based photovoltaic power interval prediction method for genetic support vector machine | |
| CN109376772B (en) | Power load combination prediction method based on neural network model | |
| CN110135630A (en) | The short term needing forecasting method with multi-step optimization is returned based on random forest | |
| CN106650784A (en) | Feature clustering comparison-based power prediction method and device for photovoltaic power station | |
| CN113837499A (en) | Ultra-short-term wind power prediction method and system | |
| CN107437135B (en) | Novel energy storage type selection method | |
| CN111461921B (en) | Load modeling typical user database updating method based on machine learning | |
| CN108376262A (en) | A kind of analysis model construction method of wind power output typical characteristics | |
| CN113762387B (en) | Multi-element load prediction method for data center station based on hybrid model prediction | |
| CN109617048B (en) | Power grid planning typical scene selection method based on multi-target linear programming | |
| CN109961173A (en) | A kind of intelligent Forecasting for distributed generation resource generated output | |
| CN106786499B (en) | Based on the short-term wind power forecast method for improving AFSA optimization ELM | |
| CN115912502A (en) | Intelligent power grid operation optimization method and device | |
| CN108694475B (en) | Short-time-scale photovoltaic cell power generation capacity prediction method based on hybrid model | |
| CN104850612B (en) | Distribution network user load characteristic classification method based on enhanced aggregation hierarchical clustering | |
| CN116316617B (en) | Multi-station intelligent fusion new energy generation power region prediction method and system | |
| CN116826710A (en) | Peak clipping strategy recommendation method and device based on load prediction and storage medium |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180522 |