CN113449934B - Wind power generation power prediction method and device based on data migration - Google Patents
Wind power generation power prediction method and device based on data migration Download PDFInfo
- Publication number
- CN113449934B CN113449934B CN202111008007.4A CN202111008007A CN113449934B CN 113449934 B CN113449934 B CN 113449934B CN 202111008007 A CN202111008007 A CN 202111008007A CN 113449934 B CN113449934 B CN 113449934B
- Authority
- CN
- China
- Prior art keywords
- data
- training set
- model
- sample
- power prediction
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000013508 migration Methods 0.000 title claims abstract description 21
- 230000005012 migration Effects 0.000 title claims abstract description 21
- 238000010248 power generation Methods 0.000 title claims abstract description 14
- 238000012549 training Methods 0.000 claims abstract description 59
- 238000012360 testing method Methods 0.000 claims abstract description 25
- 238000012805 post-processing Methods 0.000 claims abstract description 21
- 238000007781 pre-processing Methods 0.000 claims abstract description 19
- 230000008569 process Effects 0.000 claims abstract description 11
- 238000012163 sequencing technique Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 8
- 238000004140 cleaning Methods 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 7
- 238000009826 distribution Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012888 cubic function Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Images
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Wind Motors (AREA)
Abstract
The invention provides a wind power generation power prediction method and device based on data migration, which are characterized in that data migration is performed through a tree model in a data set preprocessing process, and the problem that a training set and a test set are not distributed uniformly is solved; the tree model has the highest efficiency, can be processed in parallel, is suitable for a large amount of data, and is faster than the existing method and has low requirement on data scale; the invention provides a post-processing module, which makes up the expression capability of the model and the algorithm, so that the prediction precision is higher and the accuracy is greatly improved.
Description
Technical Field
The invention belongs to the field of new energy wind power generation, and particularly relates to a wind power generation power prediction method and device based on data migration.
Background
At present, traditional machine learning algorithms such as linear models (lasso), tree models (GBM) and the like and advanced learning such as CNN, LSTM and the like are mostly adopted for power prediction related to wind power generation. Although the expression capability and the generalization capability of the algorithms are gradually strengthened, the algorithms have a large assumption premise that the training set and the test set meet independent and same distribution, and the assumption cannot be met actually, so that the accuracy of wind power prediction is influenced.
In order to solve the problem that the training set and the test set are distributed differently, other types of algorithms such as TCA, JDA, and deep learning-based correlation migration algorithms are also used. However, these algorithms also have advantages and disadvantages, and TCA and JDA are suitable for situations with small data size, such as hundreds to thousands of data, because they require large computational resources; deep learning generally requires a large amount of data to enable the model to have strong generalization.
Disclosure of Invention
The invention provides a power prediction method and device based on data migration, which solve the problem of inconsistent distribution of a training set and a test set by a tree model method, thereby greatly improving the accuracy of wind power prediction.
In order to achieve the purpose, the technical scheme of the invention is realized as follows:
a wind power generation power prediction method based on data migration is characterized in that a power prediction model is built by using a data set, and a tree model is used for preprocessing the data set before the power prediction model is modeled, wherein the preprocessing process specifically comprises the following steps:
s1, dividing the data set into a training set, a verification set and a test set according to the time stamp; carrying out data cleaning;
s2, modeling the training set by using a tree model;
s3, predicting the sample in the test set by using the model, counting the path length of each sample, and sequencing the path lengths;
and S4, selecting a sample with a lower p quantile according to the sorting result of S3, searching data in the training set by adopting a nearest neighbor method, and generating a new sample to be added into the training set.
Further, the method also comprises the following steps:
and S5, performing feature engineering processing on the training set data.
Furthermore, after a power prediction model is built, post-processing is carried out on the prediction data according to the power curve parameters of the fan.
Further, the post-processing comprises: the curve is divided into a plurality of sections according to the inflection point of the power curve of the fan, and each section is optimized according to the loss function.
Further, the method for generating a new sample in step S4 includes: and modeling the training set by using a nearest neighbor algorithm, and selecting data serving as a new sample to be added into the training set by using the built model for the data meeting the lower p quantile.
The invention also provides a wind power generation power prediction device based on data migration, which comprises a power prediction model building module and a data preprocessing module;
a power prediction model construction module for constructing a power prediction model using the data set;
a data pre-processing module for pre-processing a data set using a tree model, comprising:
the data dividing unit is used for dividing the data set into a training set, a verification set and a test set according to the time stamps; carrying out data cleaning;
a tree model modeling unit for modeling the training set using a tree model;
the test set ordering unit is used for predicting the test set samples by using the model constructed by the tree model modeling unit, counting the path length of each sample and then ordering the path lengths;
and the new sample generation unit is used for selecting a sample with a lower p quantile according to the sequencing result of the sequencing unit of the test set, searching data in the training set by adopting a nearest neighbor method, and generating a new sample to be added into the training set.
Further, the data preprocessing module further includes:
and the characteristic engineering unit is used for carrying out characteristic engineering processing on the training set data.
Furthermore, the device also comprises a post-processing module which is used for post-processing the prediction data according to the power curve parameters of the fan after a power prediction model is built.
Furthermore, the post-processing module divides the curve into a plurality of sections according to the inflection point of the power curve of the fan, and each section is optimized according to the loss function.
Furthermore, the new sample unit models the training set by using a nearest neighbor algorithm, and selects data which meet the lower p quantile and are used as new samples to be added into the training set by using the built model.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention innovatively solves the problem of inconsistent distribution of the training set and the test set by using the tree model, and has the advantages of higher speed and low requirement on data scale compared with the existing method;
2. the invention provides a post-processing module, which makes up the expression capability of a model and an algorithm, so that the prediction precision is higher and the accuracy is greatly improved.
Drawings
FIG. 1 is a schematic flow diagram of an embodiment of the present invention;
FIG. 2 is a first algorithm diagram for data preprocessing according to an embodiment of the present invention;
FIG. 3 is a second algorithm for data preprocessing according to an embodiment of the present invention;
FIG. 4 is a third algorithm diagram for data preprocessing according to an embodiment of the present invention;
FIG. 5 is a feature engineering algorithm diagram of an embodiment of the present invention;
FIG. 6 is a loss function equation for post-processing of an embodiment of the present invention.
Detailed Description
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.
In order to make the objects and features of the present invention more comprehensible, embodiments accompanying the present invention are further described below.
The design idea of the invention is that (1) the condition of inconsistent distribution of a training set and a test set is solved through a tree model; (2) the expression capability of the power prediction model is compensated through post-processing. On the premise of improving the accuracy, the method has no requirement on the data scale, and has lower space complexity and time complexity.
The first embodiment is as follows:
fig. 1 shows a schematic flow chart of the present invention, which is specifically described as follows:
firstly, obtaining a data set for constructing a power prediction model, wherein the data set can comprise wind speed, temperature, wind direction, humidity, pressure, single-site data and space data, and the space data refers to data interpolated according to longitude and latitude; dividing the data set into a training set, a verification set and a test set according to the time stamp; training set: and (4) verification set: the test set general ratio was 6: 2: 2, or the division ratio is 6: 3: 1.
and (4) carrying out data cleaning on the data, wherein the data cleaning mainly comprises abnormal value elimination, missing value processing and the like. Wherein, (1) for the abnormal value, many rejecting methods can be used, such as Isolation Forest algorithm, which identifies and determines the abnormal value by the "separating" degree of the data, and then rejects. (2) For the missing value processing, methods such as mean filling, neighbor filling, median filling and direct deletion can be adopted.
Then the important step in the present invention-the pre-processing process of performing data migration; the execution of data migration mainly deals with the problem of different distributions, and the specific process is as follows:
1. modeling the training set by using a tree model, wherein the tree model also selects an isolated forest, and the modeling principle is as follows: assuming the number psi of the training set samples, randomly selecting a feature as an initial node, randomly selecting a value in the value range of the feature, performing binary division on the psi samples, dividing the samples smaller than the value into left branches, and dividing the samples larger than the value into right branches. Such binary division operation is then repeated at both the left and right branches. Until the conditions that 1) the data itself is not re-partitionable and 2) the binary tree reaches a defined maximum depth are met, as shown in fig. 2 and 3, which are references to modeling algorithms that may be used when this step is implemented.
2. After modeling of the training set, predicting samples in the test set by using the model, counting the path length of each sample, and then sequencing the path lengths; fig. 4 shows the main contents of the algorithm that may be used when this step is implemented.
3. Since the density of the area where the sample is located is larger as the path length is larger, the model is easier to predict, so that data migration is performed only for the sample with the short path length,
in the first step of data migration, a next P quantile and samples belonging to the next P quantile are determined, and the calculation method is as follows: 1) arranging the path length data from small to large; 2) determining the position of a next P quantile, wherein the next P quantile is equivalent to the P quantile from the beginning, and the parameter P is determined according to the data set; 3) and determining a specific numerical value of the lower P quantile, namely the sample belonging to the lower P quantile.
And the second step of data migration is to find the data in the training set by adopting a nearest neighbor method aiming at the samples belonging to the lower quantile p, and then generate new samples to be added into the training set. The specific process comprises the following steps: with the knn algorithm, firstly, knn modeling is performed on a training set, then k data are selected by using a built knn model aiming at a sample m belonging to the lower P quantile, and two strategies are adopted for generating a new sample: 1) directly adding the mk samples into a training set; 2) and generating new mk samples by a smote algorithm and adding the new mk samples into a training set.
After data migration is completed and before a wind power prediction model is constructed, feature engineering processing needs to be performed on training set data after a new sample is generated, and new features are constructed, wherein the main concept of the feature engineering in the invention is stacking: firstly, training a plurality of different models, wherein different feature combinations can be used for training, such as wind direction, pressure intensity and the like, and then, the output of each model is used as the input to train one model so as to obtain a final output as a new feature, such as a new feature constructed by wind speeds at different heights; or to construct new features with all features of the same height, etc.
As shown in fig. 5, the step of performing feature engineering by the stacking method is as follows: in the step, the first-level algorithm selects different algorithms as much as possible, for example, gbdt, randomfortest, lasso and svm.
After the new feature is constructed after the feature engineering, the model for wind power prediction is built by adding the original feature and the newly constructed feature, the model selected in the embodiment is xgboost,
after power modeling, another important step of the method is a post-processing process which is mainly used for solving the problem of insufficient expression capacity of the model. The power prediction curve of wind is similar to an S-shaped curve (a physical curve is a cubic function), the prediction accuracy of the existing model for large wind speed and small wind speed is not high, and the problem is solved in the post-processing process. The predicted data can be post-processed according to the power curve parameters of the fan. The power curve parameters of the wind turbine are a set of data provided by a manufacturer, namely theoretical corresponding data of wind speed and power. Because the power curve is similar to a sigmoid curve and has two inflection points, the curve is divided into 3 sections according to the inflection points of the power curve of the fan, and each section is optimized according to a loss function.
The modeled loss function is rmse or mae, and the equations for these two loss functions are shown in FIG. 6.
Example two:
in a second embodiment, the invention provides a device for implementing the wind power generation power prediction method based on data migration in the first embodiment, and the device includes a data preprocessing module, a power prediction model building module, and a post-processing module.
1. The data preprocessing module is used for preprocessing the data set by using the tree model, and specifically comprises the following units:
the data dividing unit is used for dividing the data set into a training set, a verification set and a test set according to the time stamps; carrying out data cleaning;
a tree model modeling unit for modeling the training set using a tree model;
the test set ordering unit is used for predicting the test set samples by using the model constructed by the tree model modeling unit, counting the path length of each sample and then ordering the path lengths;
the new sample generation unit is used for selecting a sample with a lower p quantile according to the sequencing result of the sequencing unit of the test set, searching data in the training set by adopting a nearest neighbor method, and generating a new sample to be added into the training set;
the characteristic engineering unit is used for carrying out characteristic engineering processing on the training set data;
2. a power prediction model building module for building a power prediction model using the data set (including the training set new samples generated by the new sample generating unit and the newly constructed features by the feature engineering unit);
3. and the post-processing module is used for post-processing the predicted data according to the power curve parameters of the fan after a power prediction model is constructed. The post-processing module divides the curve into 3 sections according to the inflection point of the power curve of the fan, and each section is optimized according to the loss function.
The apparatus described in embodiment two may implement all the method steps described in embodiment one.
From the contents of the above embodiments, it can be seen that the practical innovation points of the present invention are as follows: data migration, there are many methods in sample density estimation, but the invention uses tree model with the highest efficiency and can process in parallel, suitable for a large amount of data; 2, because the wind power generation power prediction is more complex, the existing model expression capability is not enough to make the overall effect the best, the invention divides the curve into 3 sections by adding the post-processing module, and adopts the thought of learning segment by segment to make the overall effect the best.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (4)
1. A wind power generation power prediction method based on data migration is characterized in that a data set is preprocessed by using a tree model before modeling of a power prediction model, and the preprocessing process specifically comprises the following steps:
s1, dividing the data set into a training set, a verification set and a test set according to the time stamp; carrying out data cleaning;
s2, modeling the training set by using a tree model;
s3, predicting the sample in the test set by using the model, counting the path length of each sample, and sequencing the path lengths;
s4, aiming at the sequencing result of S3, selecting a sample with a lower p quantile, searching data in a training set by adopting a nearest neighbor method, and generating a new sample to be added into the training set;
further comprising: after a power prediction model is built, post-processing is carried out on prediction data according to the power curve parameters of the fan; the post-processing comprises: dividing a curve into a plurality of sections according to the inflection point of the power curve of the fan, and optimizing each section according to a loss function;
the method for generating a new sample in step S4 includes: modeling a training set by using a nearest neighbor algorithm, and selecting data by using the built model for data meeting a lower p quantile as a new sample to be added into the training set; the specific process comprises the following steps: with the knn algorithm, firstly, knn modeling is performed on a training set, then k data are selected by using a built knn model aiming at a sample m belonging to the lower P quantile, and two strategies are adopted for generating a new sample: 1) directly adding the mk samples into a training set; 2) and generating new mk samples by a smote algorithm and adding the new mk samples into a training set.
2. The method for wind power generation power prediction based on data migration according to claim 1, further comprising:
and S5, performing feature engineering processing on the training set data.
3. A wind power generation power prediction device based on data migration is characterized by comprising a power prediction model building module and a data preprocessing module;
a power prediction model construction module for constructing a power prediction model using the data set;
a data pre-processing module for pre-processing a data set using a tree model, comprising:
the data dividing unit is used for dividing the data set into a training set, a verification set and a test set according to the time stamps; carrying out data cleaning;
a tree model modeling unit for modeling the training set using a tree model;
the test set ordering unit is used for predicting the test set samples by using the model constructed by the tree model modeling unit, counting the path length of each sample and then ordering the path lengths;
the new sample generation unit is used for selecting a sample with a lower p quantile according to the sequencing result of the sequencing unit of the test set, searching data in the training set by adopting a nearest neighbor method, and generating a new sample to be added into the training set;
the post-processing module is used for post-processing the predicted data according to the power curve parameters of the fan after a power prediction model is built; the post-processing module divides the curve into a plurality of sections according to the inflection point of the power curve of the fan, and each section is optimized according to a loss function;
the new sample unit models the training set by using a nearest neighbor algorithm, selects data by using the built model for the data meeting the lower p quantile and adds the data as a new sample into the training set; the specific process comprises the following steps: with the knn algorithm, firstly, knn modeling is performed on a training set, then k data are selected by using a built knn model aiming at a sample m belonging to the lower P quantile, and two strategies are adopted for generating a new sample: 1) directly adding the mk samples into a training set; 2) and generating new mk samples by a smote algorithm and adding the new mk samples into a training set.
4. The wind power generation power prediction device based on data migration according to claim 3, wherein the data preprocessing module further comprises:
and the characteristic engineering unit is used for carrying out characteristic engineering processing on the training set data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111008007.4A CN113449934B (en) | 2021-08-31 | 2021-08-31 | Wind power generation power prediction method and device based on data migration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111008007.4A CN113449934B (en) | 2021-08-31 | 2021-08-31 | Wind power generation power prediction method and device based on data migration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113449934A CN113449934A (en) | 2021-09-28 |
| CN113449934B true CN113449934B (en) | 2021-11-30 |
Family
ID=77819024
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111008007.4A Active CN113449934B (en) | 2021-08-31 | 2021-08-31 | Wind power generation power prediction method and device based on data migration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113449934B (en) |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108416695B (en) * | 2018-02-24 | 2020-07-07 | 合肥工业大学 | Power load probability density prediction method, system and medium based on deep learning |
| CN110084412A (en) * | 2019-04-12 | 2019-08-02 | 重庆邮电大学 | A kind of photovoltaic power generation big data prediction technique based on the study of Feature Conversion multi-tag |
| CN110969197B (en) * | 2019-11-22 | 2022-01-04 | 上海交通大学 | Quantile prediction method for wind power generation based on instance migration |
| CN111414717A (en) * | 2020-03-02 | 2020-07-14 | 浙江大学 | XGboost-L ightGBM-based unit power prediction method |
| CN111353653B (en) * | 2020-03-13 | 2020-12-11 | 大连理工大学 | Photovoltaic output short-term interval prediction method |
| CN117977568A (en) * | 2020-10-13 | 2024-05-03 | 三峡大学 | Power load prediction method based on nested LSTM and quantile calculation |
-
2021
- 2021-08-31 CN CN202111008007.4A patent/CN113449934B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113449934A (en) | 2021-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110298663B (en) | Fraud transaction detection method based on sequence wide and deep learning | |
| RU2007116053A (en) | METHOD FOR COMPUTERIZED TRAINING ONE OR MORE NEURAL NETWORKS | |
| CN105786860A (en) | Data processing method and device in data modeling | |
| CN110363354B (en) | Wind power prediction method for wind farm, electronic device and storage medium | |
| KR20170078256A (en) | Method and apparatus for time series data prediction | |
| CN107609141A (en) | It is a kind of that quick modelling method of probabilistic is carried out to extensive renewable energy source data | |
| CN110826618A (en) | Personal credit risk assessment method based on random forest | |
| WO2022134586A1 (en) | Meta-learning-based target classification method and apparatus, device and storage medium | |
| CN113343427B (en) | Structural topology configuration prediction method based on convolutional neural network | |
| CN105117326A (en) | Test case set generation method based on combination chaotic sequence | |
| Li et al. | Deep spatio-temporal wind power forecasting | |
| CN106485030B (en) | A kind of symmetrical border processing method for SPH algorithm | |
| CN114266421B (en) | New energy power prediction method based on composite meteorological feature construction and selection | |
| CN114299305A (en) | Salient object detection algorithm for aggregating dense and attention multi-scale features | |
| CN113449934B (en) | Wind power generation power prediction method and device based on data migration | |
| CN117556369A (en) | Power theft detection method and system for dynamically generated residual error graph convolution neural network | |
| CN113761026A (en) | Feature selection method, device, equipment and storage medium based on conditional mutual information | |
| CN104573331A (en) | K neighbor data prediction method based on MapReduce | |
| CN111738086A (en) | Composition method and system for point cloud segmentation and point cloud segmentation system and device | |
| CN114997360B (en) | Evolution parameter optimization method, system and storage medium of neural architecture search algorithm | |
| CN106445960A (en) | Data clustering method and device | |
| CN112395272B (en) | Communication algorithm database construction method, distributed machine device, and storage medium | |
| CN112768081B (en) | Common-control biological network motif discovery method and device based on subgraphs and nodes | |
| JPWO2019167240A1 (en) | Information processing equipment, control methods, and programs | |
| CN108764514B (en) | Photovoltaic power generation power prediction method based on parallel operation |
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 |