CN111899123A - Electric quantity prediction method, electric quantity prediction device and computer readable storage medium - Google Patents
Electric quantity prediction method, electric quantity prediction device and computer readable storage medium Download PDFInfo
- Publication number
- CN111899123A CN111899123A CN202010741215.4A CN202010741215A CN111899123A CN 111899123 A CN111899123 A CN 111899123A CN 202010741215 A CN202010741215 A CN 202010741215A CN 111899123 A CN111899123 A CN 111899123A
- Authority
- CN
- China
- Prior art keywords
- training
- information
- electric quantity
- determining
- predicted value
- 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 56
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 230000005611 electricity Effects 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims description 459
- 230000002159 abnormal effect Effects 0.000 claims description 92
- 238000013528 artificial neural network Methods 0.000 claims description 36
- 238000012545 processing Methods 0.000 claims description 24
- 239000011159 matrix material Substances 0.000 claims description 23
- 230000015654 memory Effects 0.000 claims description 19
- 238000011423 initialization method Methods 0.000 claims description 10
- 238000010845 search algorithm Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000006403 short-term memory Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001932 seasonal effect 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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Software Systems (AREA)
- Marketing (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Public Health (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)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses an electric quantity prediction method, which comprises the following steps: acquiring electric quantity time sequence information, temperature information, holiday information, wind power information and weather information; determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information; determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value; and outputting the predicted electric quantity value. The predicted electric quantity value is predicted by comprehensively considering a plurality of factors of electric quantity time sequence information, temperature information, holiday information, wind power information and weather information, rather than only considering the factor of electric quantity data, so that the comprehensiveness and the accuracy of the electric quantity prediction result are improved. Also disclosed are an electricity amount prediction apparatus and a computer-readable storage medium.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a method and an apparatus for predicting power consumption, and a computer-readable storage medium.
Background
The power consumption data of the enterprise can be used for reflecting the overall production and development level of the enterprise. Because the power consumption is basically not influenced by artificial interference or statistical deletion and the like, the electricity utilization data can reflect the economic development condition of an enterprise more accurately. The accurate power consumption prediction can preview the electric energy consumption condition of the enterprise in time, and a certain indication effect is made for the future development planning of the enterprise.
In the prior art, the prediction of the power consumption data is mainly performed on the future power consumption of an enterprise through the historical power consumption data of the enterprise, but due to the fact that other influence factors such as seasonal factors are not considered, the power consumption requirements in different seasons are different, and if only the power consumption data is predicted, the prediction result is inaccurate.
Disclosure of Invention
The invention mainly aims to provide a power quantity prediction method, a power quantity prediction device and a computer readable storage medium, aiming at solving the problem that power quantity data is not predicted correctly.
In order to achieve the above object, the present invention provides a power prediction method, a power prediction apparatus and a computer-readable storage medium, wherein the power prediction method comprises the following steps:
acquiring electric quantity time sequence information, temperature information, holiday information, wind power information and weather information;
determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information;
determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value;
and outputting the predicted electric quantity value.
In one embodiment, the determining a first predicted value based on the power schedule information, a second predicted value based on the temperature information, the holiday information, and the wind power information, and a third predicted value based on the weather information, the determining a predicted power value based on the first predicted value, the second predicted value, and the third predicted value comprises:
inputting the electric quantity time sequence information, the temperature information, the holiday information, the wind power information and the weather information into a preset electric quantity prediction model;
and acquiring the predicted electric quantity value output by the electric quantity prediction model, wherein the electric quantity prediction model determines a first predicted value according to the electric quantity time sequence information, determines a second predicted value according to the air temperature information, the holiday information and the wind power information, determines a third predicted value according to the weather information, and outputs the predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value.
In one embodiment, before the step of inputting the power sequence information, the temperature information, the holiday information, the wind power information and the weather information into a preset power prediction model, the method further includes:
acquiring a first training set, a second training set and a third training set, wherein the first training set comprises a plurality of electric quantity time sequence training information, the second training set comprises a plurality of groups of training information, the training information comprises air temperature training information, holiday training information and wind power training information, and the third training set comprises a plurality of weather training information;
inputting the first training set into a first training layer of a neural network in a preset training model, inputting the second training set into a second training layer of the neural network, and inputting the third training set into a third training layer of the neural network;
and stopping training when the convergence value of a fourth training layer of the neural network is smaller than a preset threshold value, and storing a training model which stops training as an electric quantity prediction model, wherein the fourth training layer trains according to the output values of the first training layer, the second training layer and the third training layer.
In one embodiment, the step of inputting the first training set into a first training layer of a neural network in a preset training model comprises:
determining abnormal electric quantity values in the electric quantity time sequence training information, wherein the electric quantity time sequence training information comprises a plurality of electric quantity values which are sequentially adjacent in time;
determining an average value corresponding to the electric quantity values adjacent to the abnormal electric quantity value;
replacing the abnormal electric quantity value with an average value corresponding to the abnormal electric quantity value to obtain processed electric quantity time sequence training information;
inputting the processed electric quantity time sequence training information to the first training layer, wherein the first training set comprises the processed electric quantity time sequence training information.
In one embodiment, the step of inputting the second training set to the second training layer of the neural network is preceded by:
determining abnormal air temperature training information and abnormal wind power training information in the training information, wherein the air temperature training information comprises a first preset number of air temperature training parameters, the number of the air temperature training parameters in the abnormal air temperature training information is less than the first preset number, the wind power training information comprises a second preset number of wind power training parameters, and the number of the wind power training parameters in the abnormal wind power training information is less than the second preset number;
determining temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information, and determining the temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information;
determining wind training parameters missing from the abnormal wind training information according to each wind training parameter in the abnormal wind training information, and determining wind training parameters missing from the abnormal wind training information according to each wind training parameter in the abnormal wind training information;
adding missing air temperature training parameters to the abnormal air temperature training information and missing wind power training parameters to the abnormal wind power training information.
In one embodiment, the step of inputting the first training set to the first training layer of the neural network in the preset training model is preceded by:
determining at least one of a learning rate, a weight initialization method and a regularization coefficient respectively corresponding to the first training layer, the second training layer, the third training layer and the fourth training layer in the training model according to a grid search algorithm;
and setting training parameters of a training model according to at least one of the learning rate, the weight initialization method and the regularization coefficient.
In one embodiment, before the step of determining the third predicted value according to the weather information, the method further includes:
determining a word matrix corresponding to the weather information according to the weather information;
determining a word vector corresponding to the weather information according to the corresponding relation between the word matrix and the word vector and the word matrix;
and determining a third predicted value according to the word vector.
In order to achieve the above object, the present invention further provides a power prediction apparatus, which includes a memory, a processor, and a power prediction program stored in the memory and executable on the processor, wherein the power prediction program, when executed by the processor, implements the steps of the power prediction method as described above.
In one embodiment, the power prediction device comprises a power prediction model, and a network of the power prediction model comprises a first processing layer, a second processing layer, a third processing layer and a fourth processing layer; and the fourth processing layer outputs a predicted electric quantity value according to the output values of the first processing layer, the second processing layer and the third processing layer.
To achieve the above object, the present invention also provides a computer readable storage medium storing a power amount prediction program, which when executed by a processor implements the steps of the power amount prediction method as described above.
The invention provides an electric quantity prediction method, an electric quantity prediction device and a computer readable storage medium, which are characterized in that electric quantity time sequence information, temperature information, holiday information, wind power information and weather information are firstly acquired; then determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information; and finally, determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value and outputting the predicted electric quantity value. The predicted electric quantity value is predicted by comprehensively considering a plurality of factors of electric quantity time sequence information, temperature information, holiday information, wind power information and weather information, rather than only considering the factor of electric quantity data, so that the comprehensiveness and the accuracy of the electric quantity prediction result are improved.
Drawings
Fig. 1 is a schematic diagram of a hardware structure of an electric quantity predicting apparatus according to an embodiment of the present invention;
fig. 2 is a schematic flow chart illustrating a power prediction method according to a first embodiment of the present invention;
fig. 3 is a flowchart illustrating a power prediction method according to a second embodiment of the present invention;
fig. 4 is a flowchart illustrating a power prediction method according to a third embodiment of the present invention;
fig. 5 is a schematic flow chart illustrating a power prediction method according to a fourth embodiment of the present invention;
fig. 6 is a schematic flowchart illustrating a detailed process of step S60 of inputting the first training set into a first training layer of a neural network in a preset training model according to the power prediction method of the present invention;
fig. 7 is a flowchart illustrating a power prediction method according to a sixth embodiment of the invention.
Fig. 8 is a detailed flowchart illustrating the step S20 of determining the third predicted value according to the weather information in the seventh embodiment of the power prediction method according to the present invention;
the implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The main solution of the embodiment of the invention is as follows: acquiring electric quantity time sequence information, temperature information, holiday information, wind power information and weather information; determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information; determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value; and outputting the predicted electric quantity value.
The predicted electric quantity value is predicted by comprehensively considering a plurality of factors of electric quantity time sequence information, temperature information, holiday information, wind power information and weather information, rather than only considering the factor of electric quantity data, so that the comprehensiveness and the accuracy of the electric quantity prediction result are improved.
As an implementation, the electric quantity prediction device may be as shown in fig. 1.
The embodiment scheme of the invention relates to an electric quantity prediction device, which comprises: a processor 101, e.g. a CPU, a memory 102, a communication bus 103. Wherein a communication bus 103 is used for enabling the connection communication between these components.
The memory 102 may be a high-speed RAM memory or a non-volatile memory (e.g., a disk memory). As shown in fig. 1, a memory 102, which is a kind of computer-readable storage medium, may include therein a power amount prediction program; and the processor 101 may be configured to call the power prediction program stored in the memory 102, and perform the following operations:
acquiring electric quantity time sequence information, temperature information, holiday information, wind power information and weather information;
determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information;
determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value;
and outputting the predicted electric quantity value.
In one embodiment, the processor 101 may be configured to call the power prediction program stored in the memory 102 and perform the following operations:
inputting the electric quantity time sequence information, the temperature information, the holiday information, the wind power information and the weather information into a preset electric quantity prediction model;
and acquiring the predicted electric quantity value output by the electric quantity prediction model, wherein the electric quantity prediction model determines a first predicted value according to the electric quantity time sequence information, determines a second predicted value according to the air temperature information, the holiday information and the wind power information, determines a third predicted value according to the weather information, and outputs the predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value.
In one embodiment, the processor 101 may be configured to call the power prediction program stored in the memory 102 and perform the following operations:
acquiring a first training set, a second training set and a third training set, wherein the first training set comprises a plurality of electric quantity time sequence training information, the second training set comprises a plurality of groups of training information, the training information comprises air temperature training information, holiday training information and wind power training information, and the third training set comprises a plurality of weather training information;
inputting the first training set into a first training layer of a neural network in a preset training model; inputting the second training set to a second training layer of the neural network; and inputting the third training set to a third training layer of the neural network;
and stopping training when the convergence value of a fourth training layer of the neural network is smaller than a preset threshold value, and storing a training model which stops training as an electric quantity prediction model, wherein the fourth training layer trains according to the output values of the first training layer, the second training layer and the third training layer.
In one embodiment, the processor 101 may be configured to call the power prediction program stored in the memory 102 and perform the following operations:
determining abnormal electric quantity values in the electric quantity time sequence training information, wherein the electric quantity time sequence training information comprises a plurality of electric quantity values which are sequentially adjacent in time;
determining an average value corresponding to the electric quantity values adjacent to the abnormal electric quantity value;
replacing the abnormal electric quantity value with an average value corresponding to the abnormal electric quantity value to obtain processed electric quantity time sequence training information;
inputting the processed electric quantity time sequence training information to the first training layer, wherein the first training set comprises the processed electric quantity time sequence training information.
In one embodiment, the processor 101 may be configured to call the power prediction program stored in the memory 102 and perform the following operations:
determining abnormal air temperature training information and abnormal wind power training information in the training information, wherein the air temperature training information comprises a first preset number of air temperature training parameters, the number of the air temperature training parameters in the abnormal air temperature training information is less than the first preset number, the wind power training information comprises a second preset number of wind power training parameters, and the number of the wind power training parameters in the abnormal wind power training information is less than the second preset number;
determining temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information, and determining the temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information;
determining wind training parameters missing from the abnormal wind training information according to each wind training parameter in the abnormal wind training information, and determining wind training parameters missing from the abnormal wind training information according to each wind training parameter in the abnormal wind training information;
adding missing air temperature training parameters to the abnormal air temperature training information and missing wind power training parameters to the abnormal wind power training information.
In one embodiment, the processor 101 may be configured to call the power prediction program stored in the memory 102 and perform the following operations:
determining at least one of a learning rate, a weight initialization method and a regularization coefficient respectively corresponding to the first training layer, the second training layer, the third training layer and the fourth training layer in the training model according to a grid search algorithm;
and setting training parameters of a training model according to at least one of the learning rate, the weight initialization method and the regularization coefficient.
In one embodiment, the processor 101 may be configured to call the power prediction program stored in the memory 102 and perform the following operations:
determining a word matrix corresponding to the weather information according to the weather information;
determining a word vector corresponding to the weather information according to the corresponding relation between the word matrix and the word vector and the word matrix;
and determining a third predicted value according to the word vector.
Based on the hardware architecture of the electric quantity prediction device, the embodiment of the electric connection prediction method is provided.
Referring to fig. 2, fig. 2 is a diagram illustrating a power prediction method according to a first embodiment of the present invention, the power prediction method includes the following steps:
step S10: acquiring electric quantity time sequence information, temperature information, holiday information, wind power information and weather information.
Specifically, the electricity quantity time sequence information may be electricity quantity parameters arranged according to a time sequence, or the electricity quantity time sequence information may be an electricity quantity parameter sequence obtained by converting the electricity quantity parameters into an electricity quantity parameter sequence with a certain window length, for example, the electricity quantity sequence may be sequence data with a window of 7, and represents an electricity quantity parameter with a time step every 7 days. The power consumption parameter in the power sequence information can be expressed by kilowatt-hour or thousands of kilowatt-hour.
The temperature information is weather temperature information, may be a temperature parameter of weather including date information, or may be a temperature parameter sequence of weather in which the temperature parameter is converted into a window length. The temperature parameter in the temperature information may be represented by an average of the highest temperature and the lowest temperature on the day.
The holiday information is the holiday date corresponding to the national legal holiday or other holidays.
The wind power information is information indicating the degree of wind power, and may be a wind power parameter including date information, or a wind power parameter sequence in which the wind power parameter is converted into a certain window length. Wherein, the degree information of the wind power can be represented by wind power grade; specific numbers are extracted from the wind power information by using a regular expression, for example, the ranges of 3-4 grades are used for representing the wind power information, and the mean value of the ranges is calculated to be used as the current day wind power parameter in the wind power information.
The weather information is weather text information used for representing weather conditions, and can be a weather brief report, wherein the weather information can be cloudy, clear and the like. The weather information can be weather data containing date information, and can also be a weather data sequence which converts the weather data into a certain window length.
Step S20: and determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information.
Specifically, the first predicted value is a reference value used for determining a predicted electric quantity value, which is obtained by calculating electric quantity time sequence information; the second predicted value is a reference value for determining a predicted electric quantity value obtained by calculating temperature information, holiday information and wind power information; the third predicted value is a reference value used for determining the predicted electric quantity value after the weather information is calculated. The method comprises the steps of determining a first predicted value according to electric quantity time sequence information, determining a second predicted value according to air temperature information, holiday information and wind power information, and determining a third predicted value according to weather information, wherein the first predicted value, the second predicted value and the third predicted value can be obtained through calculation of a function formula or calculation of an input neural network.
Step S30: and determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value.
Specifically, the predicted electric quantity value is data indicating a future electric quantity used. And calculating the electric quantity time sequence information, the temperature information, the holiday information, the wind power information and the weather information in the past period of time to obtain the electric quantity data in the future period of time. For example, the electricity consumption data of the future 1 day can be calculated by using the electricity quantity time series information, the temperature information, the holiday information, the wind power information and the weather information of the past 7 days.
The predicted electric quantity value is determined according to the first predicted value, the second predicted value and the third predicted value, and the predicted electric quantity value can be obtained by setting weights for the first predicted value, the second predicted value and the third predicted value and then calculating. Or inputting the first predicted value, the second predicted value and the third predicted value into the neural network model for calculation to obtain the predicted electric quantity value.
Step S40: and outputting the predicted electric quantity value.
In the technical scheme provided by the embodiment, electric quantity time sequence information, air temperature information, holiday information, wind power information and weather information are obtained firstly; then determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information; and finally, determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value and outputting the predicted electric quantity value. The predicted electric quantity value is predicted by comprehensively considering a plurality of factors of electric quantity time sequence information, temperature information, holiday information, wind power information and weather information, rather than only considering the factor of electric quantity data, so that the comprehensiveness and the accuracy of the electric quantity prediction result are improved.
Referring to fig. 3, fig. 3 is a diagram illustrating a power prediction method according to a second embodiment of the present invention, wherein the steps S20 and S30 include the following steps:
step S21: and inputting the electric quantity time sequence information, the temperature information, the holiday information, the wind power information and the weather information into a preset electric quantity prediction model.
Specifically, the preset electric quantity prediction model is a functional module for electric quantity prediction, and may be a neural network model including an LSTM (Long Short-Term Memory) cyclic neural network. The electric quantity prediction model can be a multi-input neural network, and the multi-input neural network is a network structure capable of accepting multiple input data types, including multiple data types such as numbers, categories and images. The multi-input neural network is a neural network with a plurality of input layers, a neural network structure is established by a plurality of input data independently, and after model training is carried out relatively independently, the respective output layers or intermediate layers are combined into a combination layer and put into the same neural network for fusion training.
Step S31: and acquiring the predicted electric quantity value output by the electric quantity prediction model, wherein the electric quantity prediction model determines a first predicted value according to the electric quantity time sequence information, determines a second predicted value according to the air temperature information, the holiday information and the wind power information, determines a third predicted value according to the weather information, and outputs the predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value.
Specifically, the electric quantity prediction model calculates electric quantity time sequence information, air temperature information, holiday information, wind power information and weather information to obtain a predicted electric quantity value. And the predicted electric quantity value is the future electric quantity predicted by the electric quantity prediction model.
In the technical scheme provided by the embodiment, the electric quantity prediction model is used for analyzing the input data to obtain the predicted electric quantity value, and the electric quantity prediction model has higher reaction speed, so that the electric quantity prediction efficiency is improved. And various data are input into the electric quantity prediction model to obtain a predicted electric quantity value, so that the accuracy of electric quantity prediction is improved.
Referring to fig. 4, fig. 4 is a third embodiment of the power prediction method of the present invention, and based on the second embodiment, before the step 21, the method further includes the following steps:
step S50: the method comprises the steps of obtaining a first training set, a second training set and a third training set, wherein the first training set comprises a plurality of electric quantity time sequence training information, the second training set comprises a plurality of groups of training information, the training information comprises air temperature training information, holiday training information and wind power training information, and the third training set comprises a plurality of weather training information.
Specifically, the first training set is a training sample set including a plurality of electric quantity timing training information, and is used for training the training model. The electric quantity time sequence training information is electric quantity time sequence information used for training the training model.
The second training set is a training sample set including a plurality of sets of training information, including air temperature training information, holiday training information, and wind power training information, for training the training model. The air temperature training information is air temperature information used for training a training model, the holiday training information is holiday information used for training the training model, and the wind power training information is wind power information used for training the training model.
The third training set is a training sample set including a plurality of weather training information, and is used for training the training model. The weather training information is weather information used for training the training model.
Step S60: inputting the first training set into a first training layer of a neural network in a preset training model, inputting the second training set into a second training layer of the neural network, and inputting the third training set into a third training layer of the neural network.
Specifically, the preset training model is a functional module for learning and analyzing information in a training sample set, and becomes an electric quantity prediction model after the training of the training model is stopped. The network of the training model comprises a first training layer, a second training layer, a third training layer and a fourth training layer; and the fourth training layer outputs a predicted electric quantity value according to the output values of the first training layer, the second training layer and the third training layer.
The first training layer is used to train a first training set, which may include a LSTM (long short-Term Memory) recurrent neural network and a fully-connected layer.
The second training layer is used for training a second training set, and may include an Embedding layer, an LSTM (Long Short-Term Memory) recurrent neural network, and a full connection layer.
The third training layer is used to train a third training set, which may include a fully-connected layer.
Batch normalization layers can be added in the full-connection layer in the first training layer and the full-connection layer in the third training layer respectively, so that overfitting can be effectively reduced, and the average absolute percentage error value is reduced. The batch normalization layer is used for compressing the data in the range of [ -1,1], enabling the data to be subjected to normal distribution with the mean value of 0 and the variance of 1, and therefore the problem that the gradient of the neural network disappears is effectively solved, and the convergence speed of the neural network is improved.
Step S70: and stopping training when the convergence value of a fourth training layer of the neural network is smaller than a preset threshold value, and storing a training model which stops training as an electric quantity prediction model, wherein the fourth training layer trains according to the output values of the first training layer, the second training layer and the third training layer.
Specifically, the fourth training layer is used to train the output values of the first training layer, the second training layer, and the third training layer, and the fourth training layer may include a fully-connected layer. The training model is trained through a large amount of electric quantity time sequence information, air temperature information, holiday information, wind power information and weather information, the convergence value of the training model represents the difference situation of the output result after iterative computation and the expected result, and the preset convergence threshold value refers to the expected difference situation reached by the training of the training model. And when the convergence value is smaller than the preset convergence threshold value, representing that the output result of the training model data reaches the expected result, stopping training the training model, and storing the training model which stops training as the electric quantity prediction model.
In the technical scheme provided by this embodiment, because a technical means of training a training model to obtain an electric quantity prediction model is adopted, and the training model is trained through a large number of first training sets, second training sets and third training sets, various data are comprehensively considered by the generated electric quantity prediction model, and therefore the accuracy and comprehensiveness of diagnosis results are ensured.
Referring to fig. 5, fig. 5 is a fourth embodiment of the electric quantity prediction method according to the present invention, and based on the third embodiment, before the step S60, the method further includes the following steps:
step S80: and determining at least one of a learning rate, a weight initialization method and a regularization coefficient respectively corresponding to the first training layer, the second training layer, the third training layer and the fourth training layer in the training model according to a grid search algorithm.
Specifically, the grid search algorithm is an exhaustive search method for determining the training parameters, and arranges and combines possible values of each training parameter, lists all possible combination results, and generates a grid. Each combination was then trained and performance was evaluated using cross-validation. After all parameter combinations are tried, the optimal parameter combination is automatically adjusted. Training parameters such as a learning rate, a weight initialization method and a regularization coefficient can be determined through grid search, wherein the learning rate, the weight initialization method and the regularization coefficient are super parameters pre-adjusted by a training model, and an optimal super parameter combination is found through adjustment of the super parameters, wherein the optimal super parameter combination is a super parameter combination which enables a mean square error corresponding to a training sample set to be minimum, the optimal parameter combination enables the training model to be converged more quickly, and the phenomenon of overfitting of the training model is prevented. The test set is a sample set for performing error analysis on the training model, 10% of the training set can be divided to be used as the test set, and the remaining 90% of the training set is used for training the training model.
Step S90: and setting training parameters of a training model according to at least one of the learning rate, the weight initialization method and the regularization coefficient.
Specifically, the training parameters are hyper-parameters set by the training model before training.
In the technical scheme provided by this embodiment, by searching and determining the training parameters of the training model, the training time of the training model can be saved by selecting appropriate training parameters, so that the training model converges faster, and the error corresponding to the converged training model is smaller.
Referring to fig. 6, fig. 6 is a fifth embodiment of the power prediction method according to the present invention, and based on the third embodiment, the step S60 includes the following steps:
step S61: and determining abnormal electric quantity values in the electric quantity time sequence information, wherein the electric quantity time sequence information comprises a plurality of electric quantity values which are sequentially adjacent in time.
Specifically, the abnormal electricity amount value is an abnormally large or small electricity amount value in the electricity amount sequence information, for example, an electricity amount of a holiday enterprise is an abnormally small electricity amount value with respect to an electricity amount of a non-holiday. The existence of the abnormal electric quantity value can cause the deviation of electric quantity prediction, or the situation that the model can not be converged when the training model is trained.
Step S62: and determining the average value corresponding to the electric quantity values adjacent to the abnormal electric quantity value.
Specifically, since the electric quantity time series information includes a plurality of electric quantity values which are adjacent in time sequence, the electric quantity value adjacent to the abnormal electric quantity value can be determined, and the adjacent electric quantity value can be a plurality of adjacent electric quantity values. The average value of the adjacent electric quantity values is an average value of the plurality of electric quantity values calculated.
Step S63: and replacing the abnormal electric quantity value with an average value corresponding to the abnormal electric quantity value to obtain processed electric quantity time sequence information, and inputting the processed electric quantity time sequence training information to the first training layer, wherein the first training set comprises the processed electric quantity time sequence training information.
Specifically, the average value replaces an abnormal electric quantity value in the electric quantity time sequence information, so that new electric quantity time sequence information is obtained.
In the technical solution provided in this embodiment, since there may be a case where the power consumption data is abnormal in the power consumption time series information, for example, the power consumption data is abnormal little in holidays, the power consumption time series information needs to be preprocessed before determining the first predicted value.
Referring to fig. 7, fig. 7 is a sixth embodiment of the electric quantity prediction method according to the present invention, and based on the third embodiment, before the step S50, the method further includes the following steps:
step S100: determining abnormal air temperature training information and abnormal wind power training information in the training information, wherein the air temperature training information comprises a first preset number of air temperature training parameters, the number of the air temperature training parameters in the abnormal air temperature training information is less than the first preset number, the wind power training information comprises a second preset number of wind power training parameters, and the number of the wind power training parameters in the abnormal wind power training information is less than the second preset number.
Specifically, the abnormal air temperature training information is the air temperature training information with a missing value, and the number of the air temperature training parameters in the abnormal air temperature training information is less than a first preset number corresponding to the air temperature training parameters in the normal air temperature training information. Wherein the temperature training parameter is a temperature parameter of weather in the temperature training information.
The abnormal wind power training information is wind power training information with missing values, and the number of the wind power training parameters in the abnormal wind power training information is less than a second preset number corresponding to the wind power training parameters in the normal wind power training information. Wherein the wind training parameters are wind parameters in the wind training information.
Step S110: determining temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information, and determining the temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information; determining wind training parameters missing from the abnormal wind training information according to the wind training parameters in the abnormal wind training information, and determining wind training parameters missing from the abnormal wind training information according to the wind training parameters in the abnormal wind training information.
Specifically, the missing air temperature training parameter is an air temperature training parameter that does not exist in the air temperature training information. And determining the missing air temperature training parameters in the abnormal air temperature parameters, and determining the missing air temperature training parameters according to the air temperature training parameters in the abnormal air temperature training information. The average value of the individual air temperature training parameters may be used as the missing air temperature training parameter.
The missing wind training parameters are wind training parameters that do not exist in the wind training information. And determining missing wind power training parameters in the abnormal wind power parameters, and determining the missing wind power training parameters according to all the wind power training parameters in the abnormal wind power training information. The average of the individual wind training parameters may be used as the missing wind training parameter.
Step S120: adding missing air temperature training parameters to the abnormal air temperature training information and missing wind power training parameters to the abnormal wind power training information.
Specifically, after the missing air temperature training information is added to the abnormal air temperature training information, the number of air temperature training parameters in the air temperature training information satisfies a first preset number. And after the missing wind power training information is added to the abnormal wind power training information, the number of the wind power training parameters in the wind power training information meets a second preset number.
For convenience of description, the present embodiment shows that the steps S100 to S120 precede the step S50 in fig. 7, but the steps S100 to S120 may precede the step S60.
In the technical solution provided in this embodiment, due to the existence of the abnormal air temperature training information and the abnormal wind power training information, a large error may occur when the training model is trained. And data replacement is carried out on the abnormal air temperature training information and the abnormal wind power training information, the integrity of data in the air temperature information and the wind power information is ensured, so that the model can be converged more quickly during model training, and the prediction result is more accurate during electric quantity prediction.
Referring to fig. 8, fig. 8 is a seventh embodiment of the electric quantity prediction method according to the present invention, and based on any one of the first to sixth embodiments, the determining a third predicted value according to the weather information in step S20 includes:
step S22: and determining a word matrix corresponding to the weather information according to the weather information.
Specifically, the word matrix is a matrix used for representing information corresponding to words, text data in weather information is converted into a word matrix form, and the word matrix is used for training a training model or inputting an electric quantity prediction model for electric quantity prediction.
Step S23: determining a word vector corresponding to the weather information according to the corresponding relation between the word matrix and the word vector and the word matrix; and determining a third predicted value according to the word vector.
Specifically, the corresponding relationship between the word matrix and the word vector may be a simple mapping from a hot one matrix (one-hot) to the word vector, and the corresponding relationship between the word matrix and the word vector is obtained after training through a neural network (e.g., an Embedding layer). Through the corresponding relation between the word matrix and the word vector and the word matrix, the word vector corresponding to each word in the compression dimension can be determined, so that the dimension reduction and the feature extraction of the text data are realized. And performing feature extraction and calculation on information contained in the word vector to determine a third predicted value.
In the technical scheme provided by this embodiment, the text data in the weather information is converted into a word matrix, a word vector is determined by the word matrix, and the text data is converted into digital information, so that the weather information is quantized and calculated to obtain a third predicted value.
The present invention also provides an electric quantity prediction apparatus, which includes a memory, a processor, and an electric quantity prediction program stored in the memory and executable on the processor, wherein the electric quantity prediction program, when executed by the processor, implements the steps of the electric quantity prediction method according to the above embodiment.
The invention also provides an electric quantity prediction device, which comprises an electric quantity prediction model, wherein a network of the electric quantity prediction model comprises a first processing layer, a second processing layer, a third processing layer and a fourth processing layer; and the fourth processing layer outputs a predicted electric quantity value according to the output values of the first processing layer, the second processing layer and the third processing layer.
The present invention also provides a computer-readable storage medium storing a power amount prediction program, which when executed by a processor implements the steps of the power amount prediction method according to the above embodiment.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a computer-readable storage medium (such as ROM/RAM, magnetic disk, optical disk) as described above, and includes several instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. An electric quantity prediction method, characterized by comprising:
acquiring electric quantity time sequence information, temperature information, holiday information, wind power information and weather information;
determining a first predicted value according to the electric quantity time sequence information, determining a second predicted value according to the air temperature information, the holiday information and the wind power information, and determining a third predicted value according to the weather information;
determining a predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value;
and outputting the predicted electric quantity value.
2. The electricity amount prediction method according to claim 1, wherein the step of determining a first predicted value based on the electricity amount time series information, determining a second predicted value based on the temperature information, the holiday information, and the wind power information, and determining a third predicted value based on the weather information, and determining a predicted electricity amount value based on the first predicted value, the second predicted value, and the third predicted value comprises:
inputting the electric quantity time sequence information, the temperature information, the holiday information, the wind power information and the weather information into a preset electric quantity prediction model;
and acquiring the predicted electric quantity value output by the electric quantity prediction model, wherein the electric quantity prediction model determines a first predicted value according to the electric quantity time sequence information, determines a second predicted value according to the air temperature information, the holiday information and the wind power information, determines a third predicted value according to the weather information, and outputs the predicted electric quantity value according to the first predicted value, the second predicted value and the third predicted value.
3. The power prediction method according to claim 2, wherein before the step of inputting the power sequence information, the temperature information, the holiday information, the wind power information, and the weather information into a preset power prediction model, the method further comprises:
acquiring a first training set, a second training set and a third training set, wherein the first training set comprises a plurality of electric quantity time sequence training information, the second training set comprises a plurality of groups of training information, the training information comprises air temperature training information, holiday training information and wind power training information, and the third training set comprises a plurality of weather training information;
inputting the first training set into a first training layer of a neural network in a preset training model, inputting the second training set into a second training layer of the neural network, and inputting the third training set into a third training layer of the neural network;
and stopping training when the convergence value of a fourth training layer of the neural network is smaller than a preset threshold value, and storing a training model which stops training as an electric quantity prediction model, wherein the fourth training layer trains according to the output values of the first training layer, the second training layer and the third training layer.
4. The power prediction method of claim 3, wherein the step of inputting the first training set into a first training layer of a neural network in a preset training model comprises:
determining abnormal electric quantity values in the electric quantity time sequence training information, wherein the electric quantity time sequence training information comprises a plurality of electric quantity values which are sequentially adjacent in time;
determining an average value corresponding to the electric quantity values adjacent to the abnormal electric quantity value;
replacing the abnormal electric quantity value with an average value corresponding to the abnormal electric quantity value to obtain processed electric quantity time sequence training information;
inputting the processed electric quantity time sequence training information to the first training layer, wherein the first training set comprises the processed electric quantity time sequence training information.
5. The power prediction method of claim 3, wherein the step of inputting the second training set to a second training layer of the neural network is preceded by the step of:
determining abnormal air temperature training information and abnormal wind power training information in the training information, wherein the air temperature training information comprises a first preset number of air temperature training parameters, the number of the air temperature training parameters in the abnormal air temperature training information is less than the first preset number, the wind power training information comprises a second preset number of wind power training parameters, and the number of the wind power training parameters in the abnormal wind power training information is less than the second preset number;
determining temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information, and determining the temperature training parameters which are missing in the abnormal temperature training information according to all the temperature training parameters in the abnormal temperature training information;
determining wind training parameters missing from the abnormal wind training information according to each wind training parameter in the abnormal wind training information, and determining wind training parameters missing from the abnormal wind training information according to each wind training parameter in the abnormal wind training information;
adding missing air temperature training parameters to the abnormal air temperature training information and missing wind power training parameters to the abnormal wind power training information.
6. The method for predicting electric quantity according to claim 3, wherein the step of inputting the first training set to the first training layer of the neural network in the preset training model is preceded by the steps of:
determining at least one of a learning rate, a weight initialization method and a regularization coefficient respectively corresponding to the first training layer, the second training layer, the third training layer and the fourth training layer in the training model according to a grid search algorithm;
and setting training parameters of a training model according to at least one of the learning rate, the weight initialization method and the regularization coefficient.
7. The power prediction method of any of claims 1-6, wherein determining a third predicted value based on the weather information comprises:
determining a word matrix corresponding to the weather information according to the weather information;
determining a word vector corresponding to the weather information according to the corresponding relation between the word matrix and the word vector and the word matrix;
and determining a third predicted value according to the word vector.
8. A power prediction device, comprising a memory, a processor, and a power prediction program stored in the memory and executable on the processor, the power prediction program when executed by the processor implementing the steps of the power prediction method according to any one of claims 1-7.
9. The power prediction apparatus of claim 8, wherein the power prediction apparatus comprises a power prediction model, and the network of power prediction models comprises a first processing layer, a second processing layer, a third processing layer, and a fourth processing layer; and the fourth processing layer outputs a predicted electric quantity value according to the output values of the first processing layer, the second processing layer and the third processing layer.
10. A computer-readable storage medium storing a power prediction program which, when executed by a processor, implements the steps of the power prediction method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010741215.4A CN111899123A (en) | 2020-07-28 | 2020-07-28 | Electric quantity prediction method, electric quantity prediction device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010741215.4A CN111899123A (en) | 2020-07-28 | 2020-07-28 | Electric quantity prediction method, electric quantity prediction device and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111899123A true CN111899123A (en) | 2020-11-06 |
Family
ID=73182379
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010741215.4A Pending CN111899123A (en) | 2020-07-28 | 2020-07-28 | Electric quantity prediction method, electric quantity prediction device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111899123A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116526479A (en) * | 2023-07-03 | 2023-08-01 | 国网北京市电力公司 | Method, device, equipment and medium for predicting power supply quantity |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20180060317A (en) * | 2016-11-28 | 2018-06-07 | 한국전력공사 | Apparatus and method for predicting electricity demand based on deep neural networks |
| CN109034500A (en) * | 2018-09-04 | 2018-12-18 | 湘潭大学 | A kind of mid-term electric load forecasting method of multiple timings collaboration |
| CN109325624A (en) * | 2018-09-28 | 2019-02-12 | 国网福建省电力有限公司 | A kind of monthly electric power demand forecasting method based on deep learning |
| CN109711620A (en) * | 2018-12-26 | 2019-05-03 | 浙江大学 | A kind of Short-Term Load Forecasting Method based on GRU neural network and transfer learning |
-
2020
- 2020-07-28 CN CN202010741215.4A patent/CN111899123A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20180060317A (en) * | 2016-11-28 | 2018-06-07 | 한국전력공사 | Apparatus and method for predicting electricity demand based on deep neural networks |
| CN109034500A (en) * | 2018-09-04 | 2018-12-18 | 湘潭大学 | A kind of mid-term electric load forecasting method of multiple timings collaboration |
| CN109325624A (en) * | 2018-09-28 | 2019-02-12 | 国网福建省电力有限公司 | A kind of monthly electric power demand forecasting method based on deep learning |
| CN109711620A (en) * | 2018-12-26 | 2019-05-03 | 浙江大学 | A kind of Short-Term Load Forecasting Method based on GRU neural network and transfer learning |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116526479A (en) * | 2023-07-03 | 2023-08-01 | 国网北京市电力公司 | Method, device, equipment and medium for predicting power supply quantity |
| CN116526479B (en) * | 2023-07-03 | 2023-10-13 | 国网北京市电力公司 | Method, device, equipment and medium for predicting power supply quantity |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111210093B (en) | Daily water consumption prediction method based on big data | |
| CN112085285B (en) | Bus load prediction method, device, computer equipment and storage medium | |
| CN113988477A (en) | Photovoltaic power short-term prediction method and device based on machine learning and storage medium | |
| CN112990587B (en) | Method, system, equipment and medium for accurately predicting power consumption of transformer area | |
| CN113343581B (en) | Transformer fault diagnosis method based on graph Markov neural network | |
| CN114792156A (en) | Photovoltaic output power prediction method and system based on curve characteristic index clustering | |
| CN113657668A (en) | Power load prediction method and system based on LSTM network | |
| CN111178585A (en) | Fault reporting amount prediction method based on multi-algorithm model fusion | |
| CN110909958A (en) | Short-term load prediction method considering photovoltaic grid-connected power | |
| CN114330934A (en) | Model parameter self-adaptive GRU new energy short-term power generation power prediction method | |
| CN116599050A (en) | Photovoltaic prediction method and related device based on self-attention mechanism | |
| CN111899123A (en) | Electric quantity prediction method, electric quantity prediction device and computer readable storage medium | |
| CN111612648B (en) | Training method and device for photovoltaic power generation prediction model and computer equipment | |
| CN116109007B (en) | Power generation power determination method, server and storage medium | |
| CN114611799B (en) | Time sequence neural network new energy output multi-step prediction method based on supervised learning | |
| CN116485443A (en) | Electric power spot market price information prediction optimization method and system | |
| CN114066079B (en) | Multi-tenant-based network appointment vehicle supply and demand difference prediction method and device | |
| CN116342077B (en) | New energy power transfer learning prediction method suitable for data-missing station | |
| CN117993570A (en) | Load prediction method, device, equipment and storage medium | |
| CN116011698B (en) | Method, device, computer equipment and storage medium for determining unit combination | |
| WO2022156743A1 (en) | Feature construction method and apparatus, model training method and apparatus, and device and medium | |
| CN117039870A (en) | Power load prediction method, device, equipment and storage medium | |
| CN117992789A (en) | Method, system, equipment and storage medium for testing switch of charging gun | |
| CN116780515A (en) | Power consumption prediction method and device, computer equipment and storage medium | |
| CN114519458A (en) | Prediction method and device for customer installation capacity of data center |
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 |