CN117910657A - Prediction method, model training method, computing device, storage medium, and program product for carbon shift factor - Google Patents
Prediction method, model training method, computing device, storage medium, and program product for carbon shift factor Download PDFInfo
- Publication number
- CN117910657A CN117910657A CN202410295614.0A CN202410295614A CN117910657A CN 117910657 A CN117910657 A CN 117910657A CN 202410295614 A CN202410295614 A CN 202410295614A CN 117910657 A CN117910657 A CN 117910657A
- Authority
- CN
- China
- Prior art keywords
- data
- energy consumption
- prediction
- target
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 title claims abstract description 180
- 229910052799 carbon Inorganic materials 0.000 title claims abstract description 167
- 238000012549 training Methods 0.000 title claims abstract description 140
- 238000000034 method Methods 0.000 title claims abstract description 80
- 238000003860 storage Methods 0.000 title claims abstract description 18
- 238000005265 energy consumption Methods 0.000 claims abstract description 213
- 238000012216 screening Methods 0.000 claims description 35
- 238000004364 calculation method Methods 0.000 claims description 26
- 238000004590 computer program Methods 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 15
- 239000006229 carbon black Substances 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000004519 manufacturing process Methods 0.000 claims description 7
- 230000005611 electricity Effects 0.000 description 11
- 238000003066 decision tree Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 230000015654 memory Effects 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 5
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000004134 energy conservation Methods 0.000 description 4
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012417 linear regression Methods 0.000 description 3
- 230000005291 magnetic effect Effects 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000003646 Spearman's rank correlation coefficient Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 229910002092 carbon dioxide Inorganic materials 0.000 description 2
- 239000001569 carbon dioxide Substances 0.000 description 2
- 239000003245 coal Substances 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- -1 for example Substances 0.000 description 2
- 239000005431 greenhouse gas Substances 0.000 description 2
- 239000003345 natural gas Substances 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000003208 petroleum Substances 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 238000004378 air conditioning Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000017525 heat dissipation Effects 0.000 description 1
- 230000007787 long-term memory Effects 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
- 230000007935 neutral effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000004078 waterproofing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2123/00—Data types
- G06F2123/02—Data types in the time domain, e.g. time-series data
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (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)
Abstract
The embodiment of the invention provides a prediction method, a model training method, computing equipment, a storage medium and a program product of carbon rank factors. The prediction method of the carbon removal factor comprises the following steps: acquiring first energy consumption data generated by at least one energy source in a target area in a historical time period; predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source; acquiring a first carbon number factor released by the target area in the historical time period; and predicting a second carbon deposit factor corresponding to the target area in the prediction time period based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon deposit factor. The technical scheme provided by the embodiment of the invention realizes the prediction of the carbon rejection factor and ensures the accuracy of the carbon rejection factor.
Description
Technical Field
The embodiment of the invention relates to the technical field of artificial intelligence, in particular to a carbon-shift factor prediction method, a model training method, computing equipment, a storage medium and a program product.
Background
Carbon rejection factor refers to the amount of carbon emissions released per unit energy source during production, transportation, use, or processing. In the field of electricity, i.e. the amount of carbon emissions produced per unit of electrical energy.
The carbon emission factor can be used as an influence factor to calculate the green degree of the energy source and is used for guiding energy conservation and emission reduction measures. In the electricity utilization field, because electric energy may relate to various electric power sources, such as wind energy, heat energy, solar energy and other energy sources, carbon emission factors are usually obtained by an authority to synthesize various energy consumption in a measuring and calculating manner, an electricity consumer can only obtain carbon emission factors issued by the authority at present, and data calculation is performed based on the carbon emission factors so as to realize energy conservation and emission reduction measures such as future electricity utilization planning.
However, the carbon emission factors issued by the authorities cannot accurately represent the future carbon emission factors, and thus, the future energy saving and emission reduction measures formulated according to the carbon emission factors are not accurate.
Disclosure of Invention
The embodiment of the invention provides a carbon factor prediction method, a model training method, computing equipment, a storage medium and a program product, which are used for solving the technical problem of accuracy of carbon factor in the prior art.
In a first aspect, an embodiment of the present invention provides a method for predicting a carbon skeleton factor, including:
Acquiring first energy consumption data generated by at least one energy source in a target area in a historical time period;
predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source;
acquiring a first carbon number factor released by the target area in the historical time period;
And predicting a second carbon deposit factor corresponding to the target area in the prediction time period based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon deposit factor.
In a second aspect, an embodiment of the present invention provides a model training method, including:
Determining at least one sample characteristic parameter and a training label corresponding to the at least one sample characteristic parameter; the at least one sample characteristic parameter comprises at least one of first energy consumption sample data, time sample data and at least one weather sample data, and the training label comprises energy consumption prediction sample data; or the at least one sample characteristic parameter comprises at least one of predicted energy consumption sample data, time sample data, carbon footprint factor history sample data, and at least one weather sample data, the training tag comprising carbon footprint factor predicted sample data;
constructing a target feature set from the at least one sample feature parameter;
training a predictive model using the target feature set and the training tag;
Screening at least one key feature parameter from the target feature set;
Performing operation on any two key characteristic parameters to generate candidate characteristic parameters;
Calculating the correlation between the candidate characteristic parameter and any characteristic sample parameter in the target characteristic set;
And adding the candidate characteristic parameters with the correlation not meeting the correlation requirement into the target characteristic set, and returning to training the prediction model by utilizing the target characteristic set and the training label, wherein the step of training the prediction model is continuously executed until the prediction model reaches a training condition.
In a third aspect, an embodiment of the present invention provides a carbon factor prediction apparatus, including:
the first acquisition module is used for acquiring first energy consumption data generated by at least one energy source in a target area in a historical time period;
a first prediction module for predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source;
the second acquisition module is used for acquiring a first carbon number factor released by the target area in the historical time period;
and the second prediction module is used for predicting a second carbon factor corresponding to the target area in the prediction time period based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon factor.
In a fourth aspect, an embodiment of the present invention provides a model training apparatus, including:
The first determining module is used for determining at least one sample characteristic parameter and a training label corresponding to the at least one sample characteristic parameter; the at least one sample characteristic parameter comprises at least one of first energy consumption sample data, time sample data and at least one weather sample data, and the training label comprises energy consumption prediction sample data; or the at least one sample characteristic parameter comprises at least one of predicted energy consumption sample data, time sample data, carbon footprint factor history sample data, and at least one weather sample data, the training tag comprising carbon footprint factor predicted sample data;
a first feature set construction module for constructing a target feature set from the at least one sample feature parameter;
the first training module is used for training a prediction model by utilizing the target feature set and the training label;
A first screening module, configured to screen at least one key feature parameter from the target feature set;
The first parameter generation module is used for executing operation on any two key characteristic parameters to generate candidate characteristic parameters;
A first calculation module, configured to calculate a correlation between the candidate feature parameter and any feature sample parameter in the target feature set;
and the second training module is used for adding the candidate characteristic parameters of which the correlation does not meet the correlation requirement into the target characteristic set and triggering the first training module to continue to execute until the prediction model reaches the training condition.
In a fifth aspect, embodiments of the present invention provide a computing device comprising a processing component and a storage component;
The storage component stores one or more computer instructions; the one or more computer instructions are used for being invoked and executed by the processing component to realize the prediction method of the carbon skeleton factor provided by the embodiment of the invention or realize the model training method provided by the embodiment of the invention.
In a sixth aspect, an embodiment of the present invention provides a computer storage medium storing a computer program, where the computer program when executed by a computer implements a method for predicting a carbon skeleton factor provided by the embodiment of the present invention, or implements a model training method provided by the embodiment of the present invention.
In a seventh aspect, an embodiment of the present invention provides a computer program product, where the computer program product includes computer program code, and when the computer program code is executed by a computer device, the computer device executes the method for predicting a carbon factor provided by the embodiment of the present invention, or executes the method for training a model provided by the embodiment of the present invention.
The embodiment of the application acquires first energy consumption data generated by at least one energy source in a target area in a historical time period; predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source; acquiring a first carbon number factor released by the target area in the historical time period; based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon emission factor, predicting a second carbon emission factor corresponding to the target area in the predicted time period, and because of certain regularity of energy consumption, the embodiment of the application adopts a double-layer prediction framework by analyzing big data of the carbon emission factors released by the energy source consumption and the history, and predicts the carbon emission factor in the predicted time period of the target area based on the consumption of the energy source in the history time period of the target area, so that the predicted second carbon emission factor can accurately reflect the carbon emission level of the target area in the predicted time period, thereby more accurately guiding the energy saving and emission reduction measures of the target area.
These and other aspects of the invention will be more readily apparent from the following description of the embodiments.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 schematically illustrates a flow chart of a method for predicting carbon rejection factors according to one embodiment of the present invention;
FIG. 2 schematically illustrates a schematic diagram of a carbon number factor prediction method provided by an embodiment of the present invention;
FIG. 3 schematically illustrates a training process of a first predictive model provided by an embodiment of the invention;
FIG. 4 schematically illustrates a flow chart of a model training method provided by one embodiment of the present invention;
Fig. 5 schematically illustrates an application scenario of a method for predicting carbon-shift factors according to an embodiment of the present invention;
FIG. 6 schematically illustrates a block diagram of a carbon rejection factor prediction apparatus provided by an embodiment of the present invention;
FIG. 7 schematically illustrates a block diagram of a carbon number factor prediction apparatus provided by an embodiment of the present invention;
FIG. 8 schematically illustrates a block diagram of a computing device provided by an embodiment of the invention.
Detailed Description
In order to enable those skilled in the art to better understand the present invention, the following description will make clear and complete descriptions of the technical solutions according to the embodiments of the present invention with reference to the accompanying drawings.
In some of the flows described in the specification and claims of the present invention and in the foregoing figures, a plurality of operations occurring in a particular order are included, but it should be understood that the operations may be performed out of order or performed in parallel, with the order of operations such as 101, 102, etc., being merely used to distinguish between the various operations, the order of the operations themselves not representing any order of execution. In addition, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first" and "second" herein are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, and are not limited to the "first" and the "second" being different types.
It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present invention are information and data authorized by the user or fully authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region, and provide corresponding operation entries for the user to select authorization or rejection.
As the description of the background technology shows, the carbon emission factor is usually obtained by measuring and calculating by an authority to synthesize the consumption of various energy sources, and the electricity consumer can only obtain the carbon emission factor issued by the authority at present and calculate data based on the carbon emission factor so as to realize energy conservation and emission reduction measures such as future electricity planning.
However, in implementing the inventive concept, it was found that the carbon number factor issued by the authority does not accurately represent the carbon number factor in the future, resulting in unreliable data calculations based on the current carbon number factor. Thus, future energy saving and emission reduction measures formulated based on carbon emission factors issued by authorities are inaccurate.
In order to improve the accuracy of the carbon number factor, a series of researches have put forward the technical scheme of the embodiment of the application, in the embodiment of the application, obtain the first energy consumption data that at least one energy source produces in the goal area in the historical time quantum; predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source; acquiring a first carbon number factor released by the target area in the historical time period; based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon emission factor, predicting a second carbon emission factor corresponding to the target area in the predicted time period, and because of certain regularity of energy consumption, the embodiment of the application adopts a double-layer prediction framework by analyzing big data of the carbon emission factors released by the energy source consumption and the history, and predicts the carbon emission factor in the predicted time period of the target area based on the consumption of the energy source in the history time period of the target area, so that the predicted second carbon emission factor can accurately reflect the carbon emission level of the target area in the predicted time period, thereby more accurately guiding the energy saving and emission reduction measures of the target area.
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to fall within the scope of the invention.
Fig. 1 schematically illustrates a flowchart of a method for predicting carbon rejection factors according to an embodiment of the present application, where the technical solution of the present embodiment may be implemented by a server, and in practical application, the server may be implemented as a distributed server cluster formed by a plurality of servers, or may be implemented as a single server. The server may also be a server of a distributed system or a server that incorporates a blockchain. The server may also be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (ContentDeliveryNetwork, CDN), and basic cloud computing services such as big data and artificial intelligence platforms, or an intelligent cloud computing server or an intelligent cloud host with artificial intelligence technology, which is not limited in this application.
The method may comprise the steps of:
and 101, acquiring first energy consumption data generated by at least one energy source in a target area in a historical time period.
Wherein the historical time period may refer to a time period prior to the current time. The history period may be, for example, a history period constituted by a history time of a certain length from the current time and the current time.
In the embodiment of the present invention, the length of the historical time period may be flexibly selected according to actual application requirements, for example, one day, one week, one month, one year, etc., and the length of the historical time period is not limited herein.
In the electricity field, the energy source may refer to an energy source capable of generating electric energy, including a traditional energy source and a novel energy source, where the traditional energy source may refer to an energy source capable of generating greenhouse gases such as carbon dioxide, for example, coal, petroleum, natural gas, etc., in a use or production process, and the novel energy source may refer to solar energy, wind energy, etc.
The target area may refer to a relatively independent geographic area, such as a country, province, city, school, industrial park, etc.
The first energy consumption data of each energy source may refer to an energy consumption amount of electric energy used in a range of the target area for each energy source in a history period.
The first energy consumption data may be obtained according to the published data of the authority, and in some cases, the first energy consumption data of a certain energy source may not be obtained, and the corresponding historical energy source data is empty.
102, Predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source.
Wherein the predicted time period may refer to a time period after the current time. The predicted time period may be, for example, a predicted time period formed by a predicted time of a certain market from the current time and the current time.
In the embodiment of the present invention, the length of the predicted time period may be flexibly selected according to actual application requirements, for example, one day, one week, one month, one year, etc., and the length of the predicted time period is not limited herein. The length of the predicted time period may be the same as or different from the length of the history time period, which is not limited by the present invention.
103, Acquiring a first carbon number factor issued by the target area in a historical time period.
The first carbon factor may refer to a value of the carbon factor in a historical time period of a target area issued by the authority.
104, Predicting a second carbon factor corresponding to the target area in the prediction time period based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon factor.
In practical applications, the predicted energy data or the first carbon skeleton factor may be empty, so that when any one of the feature parameters is empty, the prediction may be performed by using the remaining feature parameters. Thus, in one possible implementation of the invention, the second energy consumption data or the first carbon footprint factor may be used to predict the second carbon footprint factor.
In another possible implementation of the present invention, the second energy consumption data and the first carbon deposit factor may be predicted to obtain the second carbon deposit factor.
In an embodiment of the present invention, the second carbon number factor may be used to indicate energy saving and emission reduction treatment of the target area within the predicted time period.
According to the embodiment, large data analysis is performed on the carbon emission factors released by energy consumption and history by utilizing the regularity of energy consumption, and a double-layer prediction framework is adopted, so that the carbon emission factor prediction is realized, the predicted second carbon emission factor can accurately reflect the carbon emission level of the target area in a prediction time period, and energy conservation and emission reduction treatment of the target area can be guided more accurately.
In some embodiments, the method may further comprise:
Acquiring at least one weather data corresponding to a target area in a predicted time period:
predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source includes:
Second energy consumption data of the at least one energy source within a predicted time period of the target area is predicted based on the first energy consumption data of the at least one energy source and at least one characteristic parameter of the at least one weather data.
In the concept of implementing the present invention, it is found that the energy consumption of the target area is generally affected by the weather condition in the target area, and thus, when the energy consumption of the target area is predicted, the weather condition of the target area can be quantified to obtain weather data, and the weather data is used as auxiliary data to perform the energy consumption prediction.
Weather data may refer to information describing atmospheric conditions and meteorological elements, and may include, for example, one or more of temperature, humidity, barometric pressure, precipitation, wind speed, wind direction, visibility, cloud cover, ultraviolet index.
Weather data is typically collected and recorded by weather observation stations, satellites, radars, and other weather equipment, and consolidated and distributed by weather institutions. Thus, weather data may be obtained from a weather facility.
The influence of weather conditions on the power consumption of a computing center will be schematically described below by taking a target area as a computing center and energy as electric energy.
First, temperature may have an impact on the energy consumption of the computing center. The operation of the computing center equipment generates a significant amount of heat and therefore requires heat dissipation to maintain the normal operating temperature of the equipment. When the outdoor temperature is high, the computing center needs to increase the use of the air conditioner to lower the indoor temperature to ensure the stability and performance of the device.
Second, humidity may have an impact on the energy consumption of the computing center. Humidity can have an impact on the reliability and durability of a computing device. Too high a humidity may lead to damage and corrosion of the equipment, while too low a humidity may lead to electrostatic discharge and equipment failure. Therefore, the computing center typically needs humidity control to ensure that the equipment is operating within a suitable humidity range.
Furthermore, precipitation may have an impact on the energy consumption of the computing center. Computing centers often need to take waterproofing measures to prevent rain intrusion into equipment and machine rooms. Such measures may include waterproof layers, drainage systems, etc.
And wind speed may have an impact on the energy consumption of the computing center. The computing center may employ an air-cooled system to replace a conventional air-conditioning system to reduce energy consumption. These systems utilize natural wind forces to dissipate heat and maintain the device temperature. Therefore, when the wind speed is higher, the effect of the air cooling system is better, and the energy consumption is correspondingly reduced.
In some embodiments, since electricity consumption may have a certain regularity at different times, such as more electricity consumption during the day and less electricity consumption at night, the predicting the second energy consumption data of the at least one energy source in the predicted period of the target area based on the at least one characteristic parameter in the first energy consumption data of the at least one energy source and the at least one weather data may include:
And predicting second energy consumption data of the at least one energy source in the predicted time period of the target area based on at least one characteristic parameter in the first energy consumption data of the at least one energy source, the at least one weather data and the time data corresponding to the predicted time period.
Of course, as yet another embodiment, the second energy consumption data of the at least one energy source in the predicted period of the target region may be predicted by using at least one characteristic parameter of the first energy consumption data of the at least one energy source and the time data corresponding to the predicted period.
In practical application, because the first energy consumption data of any energy source, any weather data and the like may be empty and cannot be obtained, by adopting the technical scheme of the embodiment of the invention, the second energy consumption data of at least one energy source in the prediction time period of the target area can be predicted under the condition that any characteristic parameter is missing. Thus, in one possible implementation of the present invention, the first energy consumption data of any energy source, any weather data or time number may be used as a characteristic parameter to predict and obtain the second energy consumption data.
In another possible implementation of the present invention, the second energy consumption data may be obtained using a plurality of characteristic parameters in at least one first energy consumption data, at least one weather data, and time data.
As can be seen from the foregoing description, the weather data affects the energy consumption and thus the carbon emission factor, and in some embodiments, based on the second energy consumption data of at least one energy source and at least one characteristic parameter of the first carbon emission factor, the prediction target region corresponding to the second carbon emission factor in the prediction time period may be specifically implemented as follows:
and predicting a second carbon emission factor corresponding to the target area in the predicted time period based on at least one characteristic parameter in the second energy consumption data, the at least one weather data and the first carbon emission factor of the at least one energy source.
In addition, the time data may be added to make the prediction, so in some embodiments, the predicting the second carbon footprint factor corresponding to the target area in the predicted time period based on the at least one characteristic parameter of the at least one energy source in the second energy consumption data, the at least one weather data, and the first carbon footprint factor may include:
And predicting a second carbon factor corresponding to the target area in the predicted time period based on at least one characteristic parameter in the second energy consumption data, the at least one weather data, the time data and the first carbon factor of the at least one energy source.
Of course, the second carbon factor corresponding to the target region in the predicted time period may be predicted based on at least one characteristic parameter of the second energy consumption data, the time data, and the first carbon factor of the at least one energy source.
In addition to the data analysis, the prediction accuracy may be further improved by using a machine learning model, so in some embodiments, predicting the second energy consumption data of the at least one energy source in the predicted time period of the target area based on the at least one feature parameter in the first energy consumption data, the at least one weather data, and the time data corresponding to the predicted time period may be specifically implemented as:
For any energy source, based on at least one characteristic parameter of first energy consumption data, at least one weather data and time data corresponding to a prediction time period of the energy source, second energy consumption data of the energy source in the prediction time period is predicted by using a first prediction model.
The first prediction model may be implemented as one of a linear regression model, a decision tree model, a random forest model, a support vector machine model, and a neural network model. In one possible implementation, the first prediction model may be one of ANN (ARTIFICIAL NEUTRAL NETWORK, artificial neural network), LSTM Long-term memory (Long short-term memory).
Specifically, predicting the second energy consumption data using the first prediction model may be implemented as:
At least one characteristic parameter is input to the first predictive model such that the first predictive model outputs second energy consumption data.
In practical application, in order to realize fine prediction, the first energy consumption data, the weather data and the time data may be time series data, and are composed of a plurality of time step data elements, where each time step may be set in combination with a practical requirement, for example, may be in an hour level, and each time step represents 1 hour.
The first energy consumption data is composed of the current time step and the energy consumption data corresponding to a plurality of time steps before the current time step respectively; the weather data are composed of weather data corresponding to a plurality of time steps respectively;
The above time data is composed of time information corresponding to each of a plurality of time steps, and may refer to a calendar variable corresponding to L hours in the future, alternatively, the time data may include at least one time sequence, for example, a month, a date, and a time sequence corresponding to each of the hours, if the current time is 14 hours, the time sequence corresponding to the hours is (10, 11, 12, 13), and since the month and the date corresponding to each of the time steps are the same, the time sequence corresponding to the month may be a value such as 5, and the time sequence corresponding to the date may be a value such as 26, and the time data is: 5, 26, (10, 11, 12, 13).
The second energy consumption data is composed of prediction data corresponding to a plurality of time steps after the current time step.
In one practical application, the prediction of the second energy consumption data using the first prediction model may be represented by the following equation (1), for example.
；（1）
Wherein,E may represent first energy consumption data,/>Can represent hour data,/>Month data,/>, can be representedCan represent date data,/>Can represent temperature data,/>Can represent humidity data,/>Can represent wind speed data,/>Dew point data can be represented,/>Network parameters that may represent the first predictive model; /(I)Representing second energy consumption data.
According to an embodiment of the present invention, based on at least one characteristic parameter of at least one of the second energy consumption data, at least one of the weather data, the time data, and the first carbon deposit factor of the at least one energy source, the second carbon deposit factor corresponding to the predicted target area in the predicted time period may be implemented as:
And predicting a second carbon factor of the target area in the predicted time period by using a second prediction model based on at least one characteristic parameter of at least one of second energy consumption data, at least one weather data, time data and the first carbon factor of the at least one energy source.
The second prediction model may be implemented as one of a linear regression model, a decision tree model, a random forest model, a support vector machine model, and a neural network model. In one possible implementation of the invention, the second predictive model may be implemented as XGBoost (Extreme Gradient Boosting, extreme gradient lifting).
Specifically, predicting the second energy consumption data using the second prediction model may be implemented as:
At least one characteristic parameter is input to the second predictive model such that the second predictive model outputs a second carbon number factor.
Predicting the second carbon number factor using the second prediction model may be embodied using equation (2) below.
Wherein,Can represent second energy consumption data,/>Can represent hour data,/>Month data,/>, can be representedCan represent date data,/>Can represent temperature data,/>The humidity data may be represented as such,Can represent wind speed data,/>Dew point data may be represented.
The second predictive model may be generated by training with a loss function. In the embodiment of the present invention, taking the second prediction model XGBoost as an example, the loss function may be a joint loss of a plurality of weak learners, and specifically may be the following formula (3).
Wherein,Can refer to loss value,/>May refer to the output value of the second predictive model,/>Tag values that may refer to the second predictive model,/>May refer to the model complexity of a weak learner decision tree.
For convenience of understanding, fig. 2 schematically shows a schematic diagram of a carbon number factor prediction method provided by an embodiment of the present invention.
As shown in fig. 2, the carbon factor prediction method provided by the embodiment of the invention adopts a double-layer architecture. When the carbon number factor value in the prediction time period is predicted by using the first energy consumption data, the weather data and the time data are firstly input into a first prediction model, and the second energy consumption data in the prediction time period are predicted. And then, inputting second energy consumption data, weather data and time data into a second prediction model, and predicting to obtain a second carbon skeleton factor in a prediction time period.
When the second energy consumption data is generated by using the first prediction model, the same or different first prediction models may be used for prediction with respect to the first energy consumption data of different types of energy sources. The plurality of first prediction models may be the same type of first prediction model, or may be different types of first prediction models.
According to the embodiment of the invention, the first prediction model may be obtained by pre-training based on the first energy consumption sample data, the time sample data, at least one sample characteristic parameter of at least one weather sample data, and the energy consumption prediction sample data corresponding to the at least one sample characteristic parameter.
The first energy consumption sample data, the time sample data, the weather sample data are similar to the data content pointed by the first energy consumption data, the time data and the weather data, and are not described herein again.
In practical applications, the first energy consumption sample data, the time sample data, the weather sample data and the like of any energy source may be empty and cannot be obtained, so that when any one of the characteristic parameters is empty, the rest of the characteristic parameters can be used for prediction. The sample characteristic parameters can be obtained by using any one or more of the first energy consumption sample data, the time sample data and the weather sample data, and the sample characteristic parameters are used as a training data set of the first prediction model to train the first prediction model.
Specifically, the training data set may be input to a first prediction model to be trained, and the first prediction model may predict energy consumption according to the input training data set and output an energy consumption prediction value.
The energy consumption prediction samples may be the expected output of the first prediction model during the present round of training, i.e. the values expected to be predicted by the first prediction model from the training data set.
After the energy consumption predicted value output by the first prediction model is obtained, the network parameters of the first prediction model can be adjusted based on the deviation of the energy consumption predicted value and the energy consumption predicted sample data, so that the training of the first prediction model is realized.
In one possible implementation of the present invention, the deviation of the energy consumption prediction value and the energy consumption prediction sample data may be calculated using the following equation (4).
From the foregoing description, it is known that some sample feature parameters may be empty and not available, so as to ensure that in the absence of feature parameters, the first prediction model still performs accurate prediction, and in some embodiments, the first prediction model is specifically trained to be obtained as follows:
determining at least one sample characteristic parameter in the first energy consumption sample data, the time sample data and the at least one weather sample data;
Constructing a target feature set from at least one sample feature parameter;
Inputting a target feature set as a model, taking energy consumption prediction sample data as a training label, and training a first prediction model;
screening at least one key feature parameter from the target feature set;
Performing operation on any two key characteristic parameters to generate candidate characteristic parameters;
calculating the correlation between the candidate characteristic parameter and any characteristic sample parameter in the target characteristic set;
And adding the candidate characteristic parameters with the correlation not meeting the correlation requirement into the target characteristic set, returning to input the target characteristic set as a model, taking the energy consumption prediction sample data as a training label, and continuously executing the step of training the first prediction model until the first prediction model reaches the training condition.
In practical applications, in order to implement the fine prediction, the first energy consumption sample data, the time sample data, and the at least one weather sample data may be time series data, and are composed of a plurality of time step data elements, where each time step may be set in combination with an actual requirement, for example, may be in an hour level, and each time step represents 1 hour.
The first energy consumption sample data may be composed of a plurality of time steps and energy consumption values corresponding to each time step; the weather sample data may be composed of a plurality of time steps and weather data values corresponding to each time step.
For the target feature set, sample feature parameters contained in the target feature set can be analyzed, the importance degree of each sample feature parameter on model prediction is determined, and the sample feature parameters with higher importance degree are determined to be key feature parameters.
In one possible implementation of the present invention, analysis methods such as residual decision tree, saproliferation and interpretation (SHAPLEY ADDITIVE exPlanations, SHAP) may be used to analyze the sample feature parameters contained in the target feature set to measure the contribution of each sample feature parameter to model prediction.
After the key feature parameters are obtained through screening, more feature parameters can be generated based on at least one key feature parameter obtained through screening.
In another implementation manner of the present invention, in the case of screening a plurality of key feature parameters from a target feature set, a data pair may be generated from two optional key feature parameters in the plurality of key feature parameters, and an operation is performed on the data pair to generate a candidate feature parameter.
The arithmetic operations may include, for example, addition, subtraction, multiplication, division, and the like. The specific manner of operation performed on each pair of data pairs may be determined randomly.
In order to improve the accuracy of model prediction, in a training data set, low correlation between each parameter is often required to be ensured, so that the influence of multiple collinearity on a model can be reduced, and the condition that an excessively high linear dependency relationship exists between the parameters is avoided.
Therefore, after obtaining the candidate feature parameters, the correlation between the candidate feature parameters and any feature sample parameter in the target feature set may be first determined, and a correlation result corresponding to each candidate feature parameter may be obtained. In embodiments of the present invention, the pearson correlation coefficient (Pearson correlation coefficient) and the spearman correlation coefficient (spearman's rank correlation coefficient) may be used to derive a correlation of the candidate feature parameter with any of the feature sample parameters in the target feature set.
The correlation requirement may for example comprise that the correlation is greater than a preset threshold. That is, only candidate feature parameters having a low correlation with any one of the feature sample parameters in the target feature set, if the correlation result is smaller than the preset threshold, can be added to the target feature set.
The target feature set with the newly added candidate feature parameters can be used for the next round of model training.
Further, after the model training of the next round is completed, the generating operation of screening the key feature parameters and the candidate feature parameters based on the key feature parameters from the target feature set may be continued until the first prediction model training is completed.
The training condition may include, for example, that the number of training times reaches the number of prediction times, or that a deviation value between the energy consumption predicted value and the energy consumption predicted sample data output by the first prediction model is smaller than a preset deviation threshold.
According to an embodiment of the present invention, screening at least one key feature parameter from the target feature set may be implemented as:
Determining a first prediction result generated by the first prediction model based on the target feature set;
at least one key feature parameter is screened from the target feature set based on the first prediction result.
The feature parameters with higher association degree with the first prediction result in the target feature set can be screened as key feature parameters.
The degree of association of the feature parameter with the first prediction result may be achieved by calculating a feature importance index based on a calculation method such as feature importance of a decision tree, a coefficient size of LASSO (The Least Absolute SHRINKAGE AND Selection Operator, least absolute shrinkage and selection operator) regression, or using a feature selection algorithm.
According to an embodiment of the present invention, the carbon number factor prediction method further includes:
Acquiring first energy consumption data generated in a first time period before a target historical time from historical production data of a target area as first energy consumption sample data and first energy consumption data generated in a second time period after the target historical time as energy consumption prediction sample data;
And taking the weather data generated in the second time period as weather sample data and the time data corresponding to the second time period as time sample data.
In some embodiments of the invention, the carbon number factor prediction method further comprises:
Training a plurality of first candidate models based on the first energy consumption sample data, the time sample data, at least one sample characteristic parameter in the at least one weather sample data, and the energy consumption prediction sample data corresponding to the at least one sample characteristic parameter;
performing model evaluation on the plurality of first candidate models;
And selecting a first candidate model with the model evaluation result meeting the performance requirement as a first prediction model.
In the process of realizing the inventive concept, it is found that different models have different dependency and sensitivity on characteristic parameters in the data set, so that the accuracy of the different models in energy consumption prediction may be different. Based on the above, the embodiment of the application can select a plurality of first candidate models, uniformly train the plurality of first candidate models, and determine the first prediction model meeting the new performance requirement from the plurality of candidate models through evaluating the plurality of first candidate models after training is completed.
Wherein the performance requirement may for example comprise that the prediction accuracy is higher than a preset threshold.
The plurality of first candidate models may be models constructed by using different ideas and methods, for example, models constructed based on a linear regression ideas, models constructed based on a neural network, models constructed based on a support vector machine, and the like.
According to the embodiment of the invention, the second prediction model is obtained based on the predicted energy consumption sample data, the second time sample data, the carbon emission factor sample data, at least one sample characteristic parameter in the at least one first atmospheric sample data, and the carbon emission factor predicted sample data corresponding to the at least one sample characteristic parameter.
According to an embodiment of the invention, the second predictive model is specifically trained to be obtained as follows:
determining at least one sample characteristic parameter in the predicted energy consumption sample data, the time sample data, the carbon array factor sample data and the at least one first atmospheric sample data;
Constructing a target feature set from at least one sample feature parameter;
Inputting a target feature set as a model, taking carbon skeleton factor prediction sample data as a training label, and training a second prediction model;
screening at least one key feature parameter from the target feature set;
Performing operation on any two key characteristic parameters to generate candidate characteristic parameters;
calculating the correlation between the candidate characteristic parameter and any characteristic sample parameter in the target characteristic set;
And adding the candidate characteristic parameters with the correlation not meeting the correlation requirement into a target characteristic set, returning to input the target characteristic set as a model, taking carbon skeleton factor prediction sample data as a training label, and continuously executing the step of training the second prediction model until the second prediction model reaches a training condition.
In practical application, in order to realize fine prediction, the above-mentioned prediction energy consumption sample data, time sample data, and carbon factor sample data may be time series data, and are composed of a plurality of time step data elements, where each time step may be set in combination with an actual requirement, for example, may be in an hour level, and each time step represents 1 hour.
The predicted energy consumption sample data may be composed of a plurality of time steps and energy consumption values corresponding to each time step; the weather sample data may be composed of a plurality of time steps and weather data values corresponding to each time step.
For the target feature set, sample feature parameters contained in the target feature set can be analyzed, the importance degree of each sample feature parameter on model prediction is determined, and the sample feature parameters with higher importance degree are determined to be key feature parameters.
In one possible implementation manner of the present invention, analysis methods such as residual decision tree, saprolidine addition and interpretation may be used to analyze sample feature parameters contained in the target feature set to measure the contribution degree of each sample feature parameter to model prediction. A larger index value may indicate that the feature has a greater impact on the prediction result.
After the key feature parameters are obtained through screening, more feature parameters can be generated based on at least one key feature parameter obtained through screening.
In one implementation manner of the present invention, under the condition that a key feature parameter is obtained by screening from a target feature set, the key feature parameter may be preprocessed to generate a first key feature parameter, then the key feature parameter and the first key feature parameter form a data pair, and an operation is performed on the data pair to generate a candidate feature parameter.
The preprocessing of the key feature parameters may, for example, comprise obtaining a weight factor, and calculating the weight factor and the key feature parameters to obtain the first key parameter.
In another implementation manner of the present invention, in the case of screening a plurality of key feature parameters from a target feature set, a data pair may be generated from two optional key feature parameters in the plurality of key feature parameters, and an operation is performed on the data pair to generate a candidate feature parameter.
The arithmetic operations may include, for example, addition, subtraction, multiplication, division, and the like. The specific manner of operation performed on each pair of data pairs may be determined randomly.
Since the key feature parameters are parameters that contribute more to the model prediction, candidate feature parameters generated using the key feature parameters may contribute more to the model prediction.
In order to improve the accuracy of model prediction, in a training data set, low correlation between each parameter is often required to be ensured, so that the influence of multiple collinearity on a model can be reduced, and the condition that an excessively high linear dependency relationship exists between the parameters is avoided.
Therefore, after obtaining the candidate feature parameters, the correlation between the candidate feature parameters and any feature sample parameter in the target feature set may be first determined, and a correlation result corresponding to each candidate feature parameter may be obtained.
The correlation requirement may for example comprise that the correlation is greater than a preset threshold. That is, only candidate feature parameters having a low correlation with any one of the feature sample parameters in the target feature set, if the correlation result is smaller than the preset threshold, can be added to the target feature set.
The target feature set with the newly added candidate feature parameters can be used for the next round of model training.
Further, after the model training of the next round is completed, the generating operation of screening the key feature parameters and the candidate feature parameters based on the key feature parameters from the target feature set may be continued until the first prediction model training is completed.
The training condition may include, for example, that the number of training times reaches the number of prediction times, or that a deviation value between the energy consumption predicted value and the energy consumption predicted sample data output by the first prediction model is smaller than a preset deviation threshold.
For ease of understanding, the training process of the first predictive model will be described below with reference to the model training schematic shown in fig. 3.
As shown in fig. 3, input data x of first energy consumption data of at least one energy source in a history period, weather data in a prediction period, time data, and the like may be input into the first prediction model, to obtain a first prediction result.
After the training of the first prediction model for one round is finished, a first prediction result is obtained, and then, based on the first prediction result, the key feature extraction 301 may be performed to obtain at least one key feature parameter.
After deriving the key feature parameters, a feature generation operation 302 may be performed. For example, candidate feature parameters may be generated by performing an arithmetic process on a data pair composed of any two key feature parameters.
Further, feature selection processing 303 may be performed for candidate feature parameters. For example, the correlation between the candidate feature parameter and the feature parameter in the target feature set may be determined, and the candidate feature parameter with lower correlation may be added to the input data x, for the training process of the first prediction model of the next round until the first prediction model meets the training condition.
According to an embodiment of the present invention, the first energy consumption data, the weather data, the time data, and the second energy consumption data are time series data, respectively;
According to an embodiment of the present invention, the method may further include:
And calculating the corresponding carbon emission quantity in any time range in the predicted time period according to the second carbon emission factor corresponding to the target area.
Wherein the carbon emission amount may be calculated by multiplying the predicted value of the carbon emission factor by the energy consumption of the target region in the selected time range.
According to an embodiment of the present invention, the target area is deployed with a data center for providing a computing service, and the carbon skeleton factor prediction method further includes:
based on the carbon emission quantity, generating recommendation prompt information of the data center;
and sending the recommendation prompt information to the target user.
Among the plurality of time ranges of the target area, the recommended prompt information may include a time range in which the carbon emission amount is minimum. The prompt message is used for prompting a user to schedule the computing task of the data center to be performed in the time range.
According to an embodiment of the present invention, the target area is deployed with a data center, and the carbon skeleton factor prediction method further includes:
Calculating the corresponding carbon emission quantity in any time range in the predicted time period according to the second carbon emission factor corresponding to the target area;
calculating the corresponding calculation cost of the data center in any time range in the prediction time period by combining the second carbon-shift factor corresponding to the target area;
and distributing calculation tasks according to the calculation costs respectively corresponding to the data centers of different target areas in different time ranges.
The calculation cost of different time ranges is calculated through the predicted value of the carbon number factor, and the calculation task can be distributed to the time range with lower calculation cost to operate, so that the operation cost of the data center can be reduced.
According to an embodiment of the present invention, based on at least one characteristic parameter of the first energy consumption data, at least one weather data, and time data representing a predicted time period, the second energy consumption data corresponding to at least one energy source in the predicted time period of the prediction target area may be specifically implemented as:
Determining at least one initial characteristic parameter constituted by the first energy consumption data, the at least one weather data, and time data representing the predicted time period;
forming a first feature set from the at least one initial feature parameter;
Performing operation on any two characteristic parameters in the first characteristic set to generate target characteristic parameters;
calculating the correlation between the target characteristic parameter and any characteristic parameter in the initial characteristic set;
Adding target feature parameters with correlation not meeting correlation requirements into the first feature set, returning any two feature parameters in the first feature set to execute operation, and continuously executing the step of generating the target feature parameters until the first feature set meets the feature requirements to obtain a second feature set;
And predicting second energy consumption data corresponding to the at least one energy source in the predicted time period by using the second feature set.
In the actual application process of the carbon skeleton factor prediction, there may be a case where there is a data loss in the first energy consumption data, at least one weather data, or time data representing the predicted time period. The lack of data may affect the accuracy of the carbon number factor predictions.
Thus, before the prediction of the carbon number factor is performed, the feature generation operation may be performed first.
For the first feature set, the initial feature parameters contained in the first feature set may be analyzed, the importance degree of each initial feature parameter on model prediction is determined, and the sample feature parameters with higher importance degree are determined as key feature parameters.
In one possible implementation manner of the present invention, the initial feature parameters included in the first feature set may be analyzed by using analysis methods such as a residual decision tree, saprolidine addition, interpretation, etc., so as to measure the contribution degree of each initial feature parameter to model prediction. A larger index value may indicate that the feature has a greater impact on the prediction result.
After the initial key feature parameters are obtained through screening, more feature parameters can be generated based on at least one initial key feature parameter obtained through screening.
In one implementation manner of the present invention, in the case of screening from the first feature set to obtain a plurality of initial key feature parameters, a data pair may be generated from two optional initial key feature parameters in the plurality of initial key feature parameters, and an operation is performed on the data pair to generate the target feature parameter.
The arithmetic operations may include, for example, addition, subtraction, multiplication, division, and the like. The specific manner of operation performed on each pair of data pairs may be determined randomly.
Since the initial key feature parameters are parameters that contribute more to the model prediction, the target feature parameters generated using the initial key feature parameters may contribute more to the model prediction.
In order to improve the accuracy of model prediction, in a training data set, low correlation between each parameter is often required to be ensured, so that the influence of multiple collinearity on a model can be reduced, and the condition that an excessively high linear dependency relationship exists between the parameters is avoided.
Therefore, after the target feature parameters are obtained, the correlation between the target feature parameters and any one of the initial feature sample parameters in the initial feature set can be judged first, and a correlation result corresponding to each target feature parameter can be obtained. In embodiments of the present invention, the pearson correlation coefficient (Pearson correlation coefficient) and the spearman correlation coefficient (spearman's rank correlation coefficient) may be used to derive a correlation of the candidate feature parameter with any of the feature sample parameters in the target feature set.
The correlation requirement may for example comprise that the correlation is greater than a preset threshold. That is, only target feature parameters having a correlation result smaller than a preset threshold, i.e., having a low correlation with any one of the initial feature sample parameters in the initial feature set, can be added to the initial feature set.
FIG. 4 schematically illustrates a flowchart of a model training method according to an embodiment of the present invention, and as shown in FIG. 4, the model training method may specifically include the following steps:
401, determining at least one sample characteristic parameter and a training label corresponding to the at least one sample characteristic parameter;
402, constructing a target feature set from at least one sample feature parameter;
403, training a prediction model by using the target feature set and the training label;
As an alternative, the at least one sample characteristic parameter comprises at least one of first energy consumption sample data, time sample data, and at least one weather sample data, and the training tag comprises energy consumption prediction sample data; the prediction model is the first prediction model.
As another alternative, the at least one sample characteristic parameter comprises at least one of predicted energy consumption sample data, time sample data, carbon black historical sample data, and at least one weather sample data, and the training tag comprises carbon black predicted sample data; the prediction model is the second prediction model.
404, Screening at least one key feature parameter from the target feature set;
405, performing operation on any two key feature parameters to generate candidate feature parameters;
406, calculating the correlation between the candidate feature parameters and any feature sample parameter in the target feature set;
And 407, adding the candidate characteristic parameters with correlation which do not meet the correlation requirement into the target characteristic set, and returning to use the target characteristic set and the training label to train the prediction model, wherein the step of training the prediction model is continuously executed until the prediction model reaches the training condition.
In practical application, in order to realize fine prediction, the first energy consumption sample data, the weather sample data and the time sample data may be time series data, and are composed of a plurality of time step data elements, where each time step may be set in combination with a practical requirement, for example, may be in an hour level, and each time step represents 1 hour.
The historical sample energy consumption data can be composed of a plurality of time steps and energy consumption values corresponding to each time step; the weather sample data may be composed of a plurality of time steps and weather data values corresponding to each time step.
The first energy consumption sample data may include first energy consumption data generated from a first period of time before the acquired target historical time from the historical production data of the target area. The first energy consumption data generated in the second period after the target history time may be used as the energy consumption sample data.
The weather sample data may be weather data occurring for a second period of time; the time sample data may be time data corresponding to the second time period.
The target history time may refer to a time target history time before the current time point, for example, a history time period formed by a history time point before the current time point and the current time point, or a history time period formed by a first history time point before the current time point and a second history time point before the current time point.
In the embodiment of the present invention, the length of the historical time period may be flexibly selected according to actual application requirements, for example, one day, one week, one month, one year, etc., and the length of the historical time period is not limited herein.
In the electricity field, the energy source may refer to an energy source capable of generating electric energy, including a traditional energy source and a novel energy source, where the traditional energy source may refer to an energy source capable of generating greenhouse gases such as carbon dioxide, for example, coal, petroleum, natural gas, etc., in a use or production process, and the novel energy source may refer to solar energy, wind energy, etc.
The target area may refer to a relatively independent geographic area, such as a country, province, city, school, industrial park, etc.
For the target feature set, sample feature parameters contained in the target feature set can be analyzed, the importance degree of each sample feature parameter on model prediction is determined, and the sample feature parameters with higher importance degree are determined to be key feature parameters.
In one possible implementation manner of the present invention, analysis methods such as residual decision tree, saprolidine addition and interpretation may be used to analyze sample feature parameters contained in the target feature set to measure the contribution degree of each sample feature parameter to model prediction. A larger index value may indicate that the feature has a greater impact on the prediction result.
After the key feature parameters are obtained through screening, more feature parameters can be generated based on at least one key feature parameter obtained through screening.
In one implementation manner of the present invention, under the condition that a key feature parameter is obtained by screening from a target feature set, the key feature parameter may be preprocessed to generate a first key feature parameter, then the key feature parameter and the first key feature parameter form a data pair, and an operation is performed on the data pair to generate a candidate feature parameter.
The preprocessing of the key feature parameters may, for example, comprise obtaining a weight factor, and calculating the weight factor and the key feature parameters to obtain the first key parameter.
In another implementation manner of the present invention, in the case of screening a plurality of key feature parameters from a target feature set, a data pair may be generated from two optional key feature parameters in the plurality of key feature parameters, and an operation is performed on the data pair to generate a candidate feature parameter.
The arithmetic operations may include, for example, addition, subtraction, multiplication, division, and the like. The specific manner of operation performed on each pair of data pairs may be determined randomly.
Since the key feature parameters are parameters that contribute more to the model prediction, candidate feature parameters generated using the key feature parameters may contribute more to the model prediction.
In order to improve the accuracy of model prediction, in a training data set, low correlation between each parameter is often required to be ensured, so that the influence of multiple collinearity on a model can be reduced, and the condition that an excessively high linear dependency relationship exists between the parameters is avoided.
Therefore, after obtaining the candidate feature parameters, the correlation between the candidate feature parameters and any feature sample parameter in the target feature set may be first determined, and a correlation result corresponding to each candidate feature parameter may be obtained.
The correlation requirement may for example comprise that the correlation is greater than a preset threshold. That is, only candidate feature parameters having a low correlation with any one of the feature sample parameters in the target feature set, if the correlation result is smaller than the preset threshold, can be added to the target feature set.
The target feature set with the newly added candidate feature parameters can be used for the next round of model training.
Further, after the model training of the next round is completed, the generating operation of screening the key feature parameters and the candidate feature parameters based on the key feature parameters from the target feature set can be continued until the prediction model training is completed.
The training condition may include, for example, that the number of training times reaches the number of prediction times, or that a deviation value between an energy consumption prediction value output by the prediction model and energy consumption prediction sample data is smaller than a preset deviation threshold.
Taking a computing task allocation scenario of a data center as an example, a technical scheme of the embodiment of the present application is described with reference to a scenario diagram shown in fig. 5.
Assuming that the data centers 500 are deployed in a plurality of regions, the server 501 may obtain, for any region (target region), first energy consumption data of the target region in a historical time period of T-L, and predict, using a first prediction model, second energy consumption data of the target region in a predicted time period based on the first energy consumption data. Then, a second prediction model can be utilized to predict and obtain a second carbon footprint factor of the target area in a prediction time period based on the second energy consumption data and the first carbon footprint factor of the target area in the history time period issued by the authority.
Wherein T represents the current time step, L can be any natural number, and T-L can represent L time steps before the current time step.
The service end 501 is configured to task M hours of t+1 to t+m according to the carbon emission factor prediction sequence of the t+1 to t+l time period, where M is smaller than L, and may calculate the number of carbon emissions in each of M hours of t+1 to t+m by combining the power of the t+1 to t+m time period. Where t+l may represent L time steps after the current time step and t+m may represent M time steps after the current time step.
The server 501 may migrate the calculation task of the first data center with a larger number of carbon rows to the second data center with a smaller number of carbon rows according to the carbon emission numbers corresponding to the multiple regions in the time period t+1 to t+m.
Or distributing the calculation task to be processed to the T+N hours with less carbon emission of the first data center for calculation processing.
Fig. 6 is a schematic structural diagram of an embodiment of a device for predicting carbon factor according to an embodiment of the present application, where the device 600 for predicting carbon factor may include:
a first obtaining module 601, configured to obtain first energy consumption data generated by at least one energy source in a target area during a historical period;
A first prediction module 602, configured to predict second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source;
A second obtaining module 603, configured to obtain a first carbon rank factor issued by the target area in the historical time period;
a second prediction module 604, configured to predict a second carbon factor corresponding to the target area in the prediction time period based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon factor.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
A third obtaining module, configured to obtain at least one weather data corresponding to the target area in the predicted time period:
According to an embodiment of the invention, the first prediction module 602 comprises:
And the first prediction sub-module is used for predicting second energy consumption data of the at least one energy source in the predicted time period of the target area based on at least one characteristic parameter in the first energy consumption data of the at least one energy source, the at least one weather data and the time data corresponding to the predicted time period.
According to an embodiment of the present invention, the second prediction module 604 includes:
And the second prediction sub-module is used for predicting a second carbon bank factor corresponding to the target area in the prediction time period based on at least one characteristic parameter in the second energy consumption data of the at least one energy source, the at least one weather data, the time data and the first carbon bank factor.
According to an embodiment of the invention, the first prediction submodule comprises:
A first prediction unit, configured to predict, for any energy source, second energy consumption data of the energy source within a prediction time period by using a first prediction model based on at least one feature parameter of first energy consumption data of the energy source, the at least one weather data, and time data corresponding to the prediction time period;
according to an embodiment of the invention, the second prediction sub-module comprises:
and the second prediction unit is used for predicting a second carbon black factor of the target area in the prediction time period by using a second prediction model based on at least one characteristic parameter in the second energy consumption data, the at least one weather data, the time data and the first carbon black factor of the at least one energy source.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
The second determining module is used for determining at least one sample characteristic parameter in the first energy consumption sample data, the time sample data and the at least one weather sample data;
A second feature set construction module for constructing a target feature set from the at least one sample feature parameter;
The third training module is used for inputting the target feature set as a model, taking the energy consumption prediction sample data as a training label and training the first prediction model;
The second screening module is used for screening at least one key characteristic parameter from the target characteristic set;
The second parameter generation module is used for executing operation on any two key characteristic parameters to generate candidate characteristic parameters;
a second calculation module, configured to calculate a correlation between the candidate feature parameter and any feature sample parameter in the target feature set;
And the fourth training module is used for adding the candidate characteristic parameters of which the correlation does not meet the correlation requirement into the target characteristic set and triggering the third training module to continue to execute until the first prediction model reaches the training condition.
According to an embodiment of the invention, the second screening module comprises:
a first result determining unit configured to determine a first prediction result generated by the first prediction model based on the target feature set;
And the screening unit is used for screening at least one key characteristic parameter from the target characteristic set based on the first prediction result.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
the first data determining module is used for acquiring first energy consumption data generated in a first time period before a target historical time from historical production data of the target area as first energy consumption sample data and first energy consumption data generated in a second time period after the target historical time as energy consumption prediction sample data;
and the second data determining module is used for taking the weather data generated in the second time period as weather sample data and taking the time data corresponding to the second time period as time sample data.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
The model training module is used for training a plurality of first candidate models based on at least one sample characteristic parameter in the first energy consumption sample data, the time sample data and at least one weather sample data and the energy consumption prediction sample data corresponding to the at least one sample characteristic parameter;
The model evaluation module is used for performing model evaluation on the plurality of first candidate models;
And the model determining module is used for selecting a first candidate model with the model evaluation result meeting the performance requirement as the first prediction model.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
a third determining module for determining at least one sample characteristic parameter of the second energy consumption sample data, the time sample data, the carbon black factor history sample data and the at least one weather sample data;
A third feature set construction module for constructing a target feature set from the at least one sample feature parameter;
the fifth training module is used for inputting the target feature set as a model, taking carbon rank factor prediction sample data as a training label and training the second prediction model;
A third screening module, configured to screen at least one key feature parameter from the target feature set;
the third parameter generation module is used for executing operation on any two key characteristic parameters to generate candidate characteristic parameters;
A third calculation module, configured to calculate a correlation between the candidate feature parameter and any feature sample parameter in the target feature set;
And the sixth training module is used for adding the candidate characteristic parameters with the correlation not meeting the correlation requirement into the target characteristic set and triggering the fifth training module to continue to execute until the second prediction model reaches the training condition.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
and the first carbon bank calculation module is used for calculating the corresponding carbon emission quantity in any time range in the prediction time period according to the second carbon bank factor corresponding to the target area.
According to an embodiment of the invention, the target area is deployed with a data center to provide computing services.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
The prompt information generation module is used for generating recommendation prompt information of the data center based on the carbon emission quantity;
and the prompt information sending module is used for sending the recommended prompt information to the target user.
According to an embodiment of the invention, the target area is deployed with a data center.
According to an embodiment of the present invention, the prediction apparatus of a carbon number factor further includes:
a second carbon number calculation module, configured to calculate, according to a second carbon number factor corresponding to the target area, a corresponding carbon emission number in an arbitrary time range in the predicted time period;
the cost calculation module is used for calculating the corresponding calculation cost of the data center in any time range in the prediction time period by combining the second carbon-shift factor corresponding to the target area;
the task allocation module is used for allocating the calculation tasks according to the calculation costs respectively corresponding to the data centers of different target areas in different time ranges.
According to an embodiment of the invention, the first prediction submodule comprises:
A fourth determining module configured to determine at least one initial characteristic parameter constituted by the first energy consumption data, the at least one weather data, and time data representing the predicted time period;
a fifth feature set construction module for constructing a first feature set from the at least one initial feature parameter;
a fourth parameter generating module, configured to perform an operation on any two feature parameters in the first feature set, to generate a target feature parameter;
A fourth calculation module, configured to calculate a correlation between the target feature parameter and any feature parameter in the initial feature set;
The feature merging module is used for adding target feature parameters of which the correlation does not meet the correlation requirement into the first feature set, returning to execute operation on any two feature parameters in the first feature set, and continuously executing the step of generating the target feature parameters until the first feature set meets the feature requirement to obtain a second feature set;
And the energy consumption prediction module is used for predicting second energy consumption data corresponding to the at least one energy source in the prediction time period by utilizing the second characteristic set.
The model training apparatus shown in fig. 6 may perform the method for predicting carbon rejection factors according to the embodiment shown in fig. 1, and its implementation principle and technical effects will not be described again. The specific manner in which the respective modules and units perform the operations in the apparatus for predicting carbon black in the above embodiment has been described in detail in the embodiments related to the method, and will not be described in detail herein.
Fig. 7 is a schematic structural diagram of an embodiment of a model training apparatus according to an embodiment of the present application, where a model training apparatus 700 may include:
A first determining module 701, configured to determine at least one sample feature parameter and a training label corresponding to the at least one sample feature parameter; the at least one sample characteristic parameter comprises at least one of first energy consumption sample data, time sample data and at least one weather sample data, and the training label comprises energy consumption prediction sample data; or the at least one sample characteristic parameter comprises at least one of predicted energy consumption sample data, time sample data, carbon footprint factor history sample data, and at least one weather sample data, the training tag comprising carbon footprint factor predicted sample data;
a first feature set construction module 702 for constructing a target feature set from the at least one sample feature parameter;
a first training module 703, configured to train a prediction model using the target feature set and the training label;
a first screening module 704, configured to screen at least one key feature parameter from the target feature set;
The first parameter generating module 705 is configured to perform an operation on any two key feature parameters to generate candidate feature parameters;
A first calculation module 706, configured to calculate a correlation between the candidate feature parameter and any feature sample parameter in the target feature set;
And the second training module 707 is configured to add the candidate feature parameters whose correlation does not meet the correlation requirement to the target feature set, and trigger the first training module to continue executing until the prediction model reaches a training condition.
The model training apparatus shown in fig. 7 may perform the model training method described in the embodiment shown in fig. 5, and its implementation principle and technical effects will not be described again. The specific manner in which the respective modules and units of the model training apparatus in the above embodiment perform operations has been described in detail in the embodiments related to the method, and will not be described in detail here.
In one possible design, the prediction apparatus and the model training apparatus for carbon emission factors provided by the embodiments of the present invention may be implemented as a computing device, as shown in fig. 8, where the computing device may include a storage component 801 and a processing component 802;
The storage component 801 stores one or more computer instructions, where the one or more computer instructions are called by the processing component 802 to implement the prediction method and the model training method of the carbon skeleton factor provided by the embodiment of the present invention.
Of course, the computing device may necessarily include other components, such as input/output interfaces, communication components, and the like. The input/output interface provides an interface between the processing component and a peripheral interface module, which may be an output device, an input device, etc. The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, and the like.
The computing device may be a physical device or an elastic computing host provided by the cloud computing platform, and at this time, the computing device may be a cloud server, and the processing component, the storage component, and the like may be a base server resource rented or purchased from the cloud computing platform.
When the computing device is a physical device, the computing device may be implemented as a distributed cluster formed by a plurality of servers or terminal devices, or may be implemented as a single server or a single terminal device.
The embodiment of the invention also provides a computer readable storage medium which stores a computer program, and the computer program can realize the prediction method and the model training method of the carbon shift factor provided by the embodiment of the invention when being executed by a computer.
The embodiment of the invention also provides a computer program product, which comprises a computer program, wherein the computer program can realize the prediction method and the model training method of the carbon emission factor provided by the embodiment of the invention when being executed by a computer.
Wherein the processing components of the respective embodiments above may include one or more processors to execute computer instructions to perform all or part of the steps of the methods described above. Of course, the processing component may also be implemented as one or more Application Specific Integrated Circuits (ASICs), digital Signal Processors (DSPs), digital Signal Processing Devices (DSPDs), programmable Logic Devices (PLDs), field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic elements for executing the methods described above.
The storage component is configured to store various types of data to support operation in the device. The memory component may be implemented by any type or combination of volatile or nonvolatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (17)
1. A method for predicting carbon rejection factor, comprising:
Acquiring first energy consumption data generated by at least one energy source in a target area in a historical time period;
predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source;
acquiring a first carbon number factor released by the target area in the historical time period;
And predicting a second carbon deposit factor corresponding to the target area in the prediction time period based on the second energy consumption data of the at least one energy source and at least one characteristic parameter in the first carbon deposit factor.
2. The method as recited in claim 1, further comprising:
acquiring at least one weather data corresponding to the target area in the prediction time period:
The predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on the first energy consumption data of the at least one energy source includes:
And predicting second energy consumption data of the at least one energy source in the predicted time period of the target area based on at least one characteristic parameter in the first energy consumption data of the at least one energy source, the at least one weather data and the time data corresponding to the predicted time period.
3. The method of claim 2, wherein predicting a corresponding second carbon footprint factor for the target region over the predicted time period based on the second energy consumption data of the at least one energy source and at least one characteristic parameter of the first carbon footprint factor comprises:
And predicting a second carbon deposit factor corresponding to the target area in the predicted time period based on at least one characteristic parameter of the at least one energy source in the second energy consumption data, the at least one weather data, the time data and the first carbon deposit factor.
4. The method of claim 3, wherein predicting second energy consumption data of the at least one energy source within a predicted time period of the target area based on at least one characteristic parameter of the first energy consumption data, the at least one weather data, and the time data corresponding to the predicted time period comprises:
For any energy source, predicting second energy consumption data of the energy source in a prediction time period by using a first prediction model based on at least one characteristic parameter in first energy consumption data of the energy source, the at least one weather data and time data corresponding to the prediction time period;
The predicting a second carbon footprint factor corresponding to the target area within the predicted time period based on at least one characteristic parameter of the at least one energy source among the second energy consumption data, the at least one weather data, the time data, and the first carbon footprint factor includes:
And predicting a second carbon footprint factor of the target area in the prediction time period by using a second prediction model based on at least one characteristic parameter in the second energy consumption data, the at least one weather data, the time data and the first carbon footprint factor of the at least one energy source.
5. The method according to claim 4, characterized in that the first predictive model is specifically trained to be obtained in the following way:
determining at least one sample characteristic parameter in the first energy consumption sample data, the time sample data and the at least one weather sample data;
constructing a target feature set from the at least one sample feature parameter;
inputting the target feature set as a model, taking energy consumption prediction sample data as a training label, and training the first prediction model;
Screening at least one key feature parameter from the target feature set;
Performing operation on any two key characteristic parameters to generate candidate characteristic parameters;
Calculating the correlation between the candidate characteristic parameter and any characteristic sample parameter in the target characteristic set;
And adding the candidate characteristic parameters with the correlation not meeting the correlation requirement into the target characteristic set, returning to input the target characteristic set as a model, taking the energy consumption prediction sample data as a training label, and continuously executing the step of training the first prediction model until the first prediction model reaches a training condition.
6. The method of claim 5, wherein said screening at least one key feature parameter from said target feature set comprises:
determining a first prediction result generated by the first prediction model based on the target feature set;
and screening at least one key feature parameter from the target feature set based on the first prediction result.
7. The method as recited in claim 4, further comprising:
Acquiring first energy consumption data generated in a first time period before a target historical time from historical production data of the target area as first energy consumption sample data and first energy consumption data generated in a second time period after the target historical time as energy consumption prediction sample data;
And taking the weather data generated in the second time period as weather sample data and the time data corresponding to the second time period as time sample data.
8. The method as recited in claim 7, further comprising:
Training a plurality of first candidate models based on at least one sample feature parameter in the first energy consumption sample data, the time sample data and the at least one weather sample data, and the energy consumption prediction sample data corresponding to the at least one sample feature parameter;
performing model evaluation on the plurality of first candidate models;
And selecting a first candidate model with the model evaluation result meeting the performance requirement as the first prediction model.
9. The method according to claim 4, characterized in that the second predictive model is specifically trained to be obtained in the following way:
determining at least one sample characteristic parameter of the second energy consumption sample data, the time sample data, the carbon black historical sample data and the at least one weather sample data;
constructing a target feature set from the at least one sample feature parameter;
Inputting the target feature set as a model, taking carbon skeleton factor prediction sample data as a training label, and training the second prediction model;
Screening at least one key feature parameter from the target feature set;
Performing operation on any two key characteristic parameters to generate candidate characteristic parameters;
Calculating the correlation between the candidate characteristic parameter and any characteristic sample parameter in the target characteristic set;
And adding the candidate characteristic parameters with the correlation not meeting the correlation requirement into the target characteristic set, returning to input the target characteristic set as a model, taking the carbon skeleton factor prediction sample data as a training label, and continuously executing the step of training the second prediction model until the second prediction model reaches a training condition.
10. The method as recited in claim 1, further comprising:
And calculating the corresponding carbon emission quantity in any time range in the prediction time period according to the second carbon emission factor corresponding to the target area.
11. The method of claim 10, wherein the target area is deployed with a data center to provide computing services, the method further comprising:
Generating recommendation prompt information of the data center based on the carbon emission quantity;
And sending the recommendation prompt information to the target user.
12. The method of claim 1, wherein the target area is deployed with a data center, the method further comprising:
Calculating the corresponding carbon emission quantity in any time range in the predicted time period according to the second carbon emission factor corresponding to the target area;
Calculating the corresponding calculation cost of the data center in any time range in the prediction time period by combining the second carbon-shift factor corresponding to the target area;
and distributing calculation tasks according to the calculation costs respectively corresponding to the data centers of different target areas in different time ranges.
13. The method of claim 1, wherein predicting second energy consumption data for the target area for the at least one energy source over a predicted time period based on at least one characteristic parameter of the first energy consumption data, the at least one weather data, and time data representing the predicted time period comprises:
Determining at least one initial characteristic parameter constituted by the first energy consumption data, the at least one weather data, and time data representing the predicted time period;
forming a first feature set from the at least one initial feature parameter;
Performing operation on any two characteristic parameters in the first characteristic set to generate target characteristic parameters;
calculating the correlation between the target characteristic parameter and any characteristic parameter in the initial characteristic set;
Adding target feature parameters with correlation not meeting correlation requirements into the first feature set, returning any two feature parameters in the first feature set to execute operation, and continuously executing the step of generating the target feature parameters until the first feature set meets the feature requirements to obtain a second feature set;
And predicting second energy consumption data corresponding to the at least one energy source in the predicted time period by using the second feature set.
14. A method of model training, comprising:
Determining at least one sample characteristic parameter and a training label corresponding to the at least one sample characteristic parameter; the at least one sample characteristic parameter comprises at least one of first energy consumption sample data, time sample data and at least one weather sample data, and the training label comprises energy consumption prediction sample data; or the at least one sample characteristic parameter comprises at least one of predicted energy consumption sample data, time sample data, carbon footprint factor history sample data, and at least one weather sample data, the training tag comprising carbon footprint factor predicted sample data;
constructing a target feature set from the at least one sample feature parameter;
training a predictive model using the target feature set and the training tag;
Screening at least one key feature parameter from the target feature set;
Performing operation on any two key characteristic parameters to generate candidate characteristic parameters;
Calculating the correlation between the candidate characteristic parameter and any characteristic sample parameter in the target characteristic set;
And adding the candidate characteristic parameters with the correlation not meeting the correlation requirement into the target characteristic set, and returning to training the prediction model by utilizing the target characteristic set and the training label, wherein the step of training the prediction model is continuously executed until the prediction model reaches a training condition.
15. A computing device comprising a processing component and a storage component;
The storage component stores one or more computer instructions; the one or more computer instructions are to be invoked by the processing component to implement a method of predicting carbon black as claimed in any one of claims 1 to 13, or to implement a model training method as claimed in claim 14.
16. A computer storage medium, characterized in that a computer program is stored, which, when being executed by a computer, implements the method of predicting carbon black according to any one of claims 1 to 13, or implements the model training method according to claim 14.
17. A computer program product, characterized in that the computer program product comprises computer program code which, when executed by a computer device, performs the method of predicting carbon black according to any one of the preceding claims 1 to 13 or the method of model training according to claim 14.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410295614.0A CN117910657B (en) | 2024-03-14 | 2024-03-14 | Prediction method, model training method, computing device, storage medium, and program product for carbon shift factor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410295614.0A CN117910657B (en) | 2024-03-14 | 2024-03-14 | Prediction method, model training method, computing device, storage medium, and program product for carbon shift factor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117910657A true CN117910657A (en) | 2024-04-19 |
| CN117910657B CN117910657B (en) | 2024-07-26 |
Family
ID=90695056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410295614.0A Active CN117910657B (en) | 2024-03-14 | 2024-03-14 | Prediction method, model training method, computing device, storage medium, and program product for carbon shift factor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117910657B (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113988489A (en) * | 2021-12-27 | 2022-01-28 | 国网浙江省电力有限公司信息通信分公司 | Prediction evaluation method and device of electrical carbon factor and readable storage medium |
| CN114841274A (en) * | 2022-05-12 | 2022-08-02 | 百度在线网络技术（北京）有限公司 | Language model training method and device, electronic equipment and storage medium |
| CN115600722A (en) * | 2022-09-02 | 2023-01-13 | 浙江大学（Cn） | Long-term prediction method for electric carbon factor of enterprise |
| CN116822997A (en) * | 2023-08-30 | 2023-09-29 | 国网区块链科技(北京)有限公司 | Real-time carbon emission accounting method and device considering green power |
| CN116862050A (en) * | 2023-06-09 | 2023-10-10 | 广东电网有限责任公司广州供电局 | Time sequence network-based daily prediction method, system, storage medium and equipment for carbon emission factors |
| CN117371622A (en) * | 2023-12-06 | 2024-01-09 | 国网湖北省电力有限公司经济技术研究院 | Prediction method, system and equipment for long-term dynamic carbon emission factor in region |
| CN117371610A (en) * | 2023-10-30 | 2024-01-09 | 国网山东省电力公司营销服务中心(计量中心) | Carbon emission monitoring method and system |
| CN117408394A (en) * | 2023-12-14 | 2024-01-16 | 国网天津市电力公司电力科学研究院 | Carbon emission factor prediction method and device for electric power system and electronic equipment |
| WO2024020801A1 (en) * | 2022-07-26 | 2024-02-01 | 西门子股份公司 | Carbon emission calculation method, electronic device and readable medium |
| CN117494931A (en) * | 2023-09-22 | 2024-02-02 | 国网湖北省电力有限公司经济技术研究院 | Marginal carbon emission factor determining method, system and equipment based on power grid node |
| CN117610737A (en) * | 2023-12-07 | 2024-02-27 | 广东技术师范大学 | ASTGCN-based carbon emission factor prediction method and system |
| CN117634678A (en) * | 2023-11-24 | 2024-03-01 | 国舜绿建科技有限公司 | Low-carbon park carbon emission prediction method based on actual operation scene |
| CN117670586A (en) * | 2023-11-28 | 2024-03-08 | 东北大学佛山研究生创新学院 | Power grid node carbon factor prediction method and system based on graph neural network |
-
2024
- 2024-03-14 CN CN202410295614.0A patent/CN117910657B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113988489A (en) * | 2021-12-27 | 2022-01-28 | 国网浙江省电力有限公司信息通信分公司 | Prediction evaluation method and device of electrical carbon factor and readable storage medium |
| CN114841274A (en) * | 2022-05-12 | 2022-08-02 | 百度在线网络技术（北京）有限公司 | Language model training method and device, electronic equipment and storage medium |
| WO2024020801A1 (en) * | 2022-07-26 | 2024-02-01 | 西门子股份公司 | Carbon emission calculation method, electronic device and readable medium |
| CN115600722A (en) * | 2022-09-02 | 2023-01-13 | 浙江大学（Cn） | Long-term prediction method for electric carbon factor of enterprise |
| CN116862050A (en) * | 2023-06-09 | 2023-10-10 | 广东电网有限责任公司广州供电局 | Time sequence network-based daily prediction method, system, storage medium and equipment for carbon emission factors |
| CN116822997A (en) * | 2023-08-30 | 2023-09-29 | 国网区块链科技(北京)有限公司 | Real-time carbon emission accounting method and device considering green power |
| CN117494931A (en) * | 2023-09-22 | 2024-02-02 | 国网湖北省电力有限公司经济技术研究院 | Marginal carbon emission factor determining method, system and equipment based on power grid node |
| CN117371610A (en) * | 2023-10-30 | 2024-01-09 | 国网山东省电力公司营销服务中心(计量中心) | Carbon emission monitoring method and system |
| CN117634678A (en) * | 2023-11-24 | 2024-03-01 | 国舜绿建科技有限公司 | Low-carbon park carbon emission prediction method based on actual operation scene |
| CN117670586A (en) * | 2023-11-28 | 2024-03-08 | 东北大学佛山研究生创新学院 | Power grid node carbon factor prediction method and system based on graph neural network |
| CN117371622A (en) * | 2023-12-06 | 2024-01-09 | 国网湖北省电力有限公司经济技术研究院 | Prediction method, system and equipment for long-term dynamic carbon emission factor in region |
| CN117610737A (en) * | 2023-12-07 | 2024-02-27 | 广东技术师范大学 | ASTGCN-based carbon emission factor prediction method and system |
| CN117408394A (en) * | 2023-12-14 | 2024-01-16 | 国网天津市电力公司电力科学研究院 | Carbon emission factor prediction method and device for electric power system and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| 杜强;陈乔;杨锐;: "基于Logistic模型的中国各省碳排放预测", 长江流域资源与环境, no. 02, 15 February 2013 (2013-02-15) * |
| 王志明;谭显胜;袁哲明;伍朝华;: "自调用支持向量回归优化支持向量机参数", 系统仿真学报, no. 02, 8 February 2010 (2010-02-08) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117910657B (en) | 2024-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hyandye et al. | A Markovian and cellular automata land-use change predictive model of the Usangu Catchment | |
| Wang et al. | LASSO and LSTM integrated temporal model for short-term solar intensity forecasting | |
| Salehin et al. | An artificial intelligence based rainfall prediction using LSTM and neural network | |
| Izadyar et al. | Appraisal of the support vector machine to forecast residential heating demand for the District Heating System based on the monthly overall natural gas consumption | |
| CN112285807B (en) | Meteorological information prediction method and device | |
| CN111667093B (en) | Medium-and-long-term wind power generation calculation method and device | |
| Goyal et al. | Evaluation of machine learning tools as a statistical downscaling tool: temperatures projections for multi-stations for Thames River Basin, Canada | |
| Shortridge et al. | Scenario discovery with multiple criteria: An evaluation of the robust decision‐making framework for climate change adaptation | |
| CN106919645A (en) | The sight spot meteorological element Intelligent fine Forecasting Methodology at the big scenic spot of complex landform | |
| CN114493052A (en) | Multi-model fusion self-adaptive new energy power prediction method and system | |
| CN117613883A (en) | Method and device for predicting generated power, computer equipment and storage medium | |
| CN112906757A (en) | Model training method, optical power prediction method, device, equipment and storage medium | |
| CN112990530A (en) | Regional population number prediction method and device, electronic equipment and storage medium | |
| Li et al. | Enhancing process-based hydrological models with embedded neural networks: A hybrid approach | |
| Khan et al. | Hybrid deep learning approach for multi-step-ahead prediction for daily maximum temperature and heatwaves | |
| Ghassemi et al. | Optimal surrogate and neural network modeling for day-ahead forecasting of the hourly energy consumption of university buildings | |
| CN117932470A (en) | Moon runoff forecasting method based on dynamic Bayesian network | |
| CN117391257A (en) | Road congestion condition prediction method and device | |
| CN117910657B (en) | Prediction method, model training method, computing device, storage medium, and program product for carbon shift factor | |
| Sharma et al. | Deep Learning Based Prediction Of Weather Using Hybrid_stacked Bi-Long Short Term Memory | |
| CN116646917A (en) | Campus multi-step probability power load prediction method and system | |
| CN114741976B (en) | Displacement prediction method, device and equipment and readable storage medium | |
| CN114418243B (en) | Distributed new energy cloud grid prediction method and system | |
| Majumdar et al. | Rainfall forecasting for Silchar city using stacked-LSTM | |
| Saha et al. | Predictor‐Year Subspace Clustering Based Ensemble Prediction of Indian Summer Monsoon |
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 |