CN113837815A - Quotation prediction model construction method, prediction method, system, device and medium - Google Patents
Quotation prediction model construction method, prediction method, system, device and medium Download PDFInfo
- Publication number
- CN113837815A CN113837815A CN202111300944.7A CN202111300944A CN113837815A CN 113837815 A CN113837815 A CN 113837815A CN 202111300944 A CN202111300944 A CN 202111300944A CN 113837815 A CN113837815 A CN 113837815A
- Authority
- CN
- China
- Prior art keywords
- sample
- historical
- disclosure information
- matrix
- market
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000010276 construction Methods 0.000 title claims abstract description 21
- 239000011159 matrix material Substances 0.000 claims abstract description 148
- 238000012549 training Methods 0.000 claims abstract description 60
- 238000012360 testing method Methods 0.000 claims abstract description 37
- 238000013136 deep learning model Methods 0.000 claims abstract description 18
- 238000012795 verification Methods 0.000 claims abstract description 16
- 238000004590 computer program Methods 0.000 claims description 16
- 238000010606 normalization Methods 0.000 claims description 15
- 230000007246 mechanism Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 12
- 238000013135 deep learning Methods 0.000 description 7
- 230000006399 behavior Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 230000006872 improvement Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention provides a quotation prediction model construction method, a prediction method, a system, equipment and a medium, wherein the method comprises the following steps: acquiring historical disclosure information of the power market, and selecting historical disclosure information corresponding to similar day data before a forecast day to form a first sample; taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set; aiming at each market member, a multilayer deep learning model is constructed, and a prediction model is obtained by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification. The method can predict the electric power market quotation in the market environment, and can provide help for the electric power mechanism to supervise the electric power market operation. The method needs less workload of personnel and has the characteristics of high efficiency and convenience.
Description
Technical Field
The invention relates to a quotation prediction model construction method, a prediction system, equipment and a medium, and relates to the technical field of electric power market quotation.
Background
With the development of the power market, the behaviors of market members are full of autonomy, and quotation is a subjective factor of the market members. It is difficult for the regulatory authorities to obtain their quotation policies through manual analysis to achieve the purpose of supervision. Deep learning may provide a relatively objective way of analyzing predictions. In the centralized spot market, the forecast analysis of the quotation behaviors of market participants provides reference for the adjustment of the market rule organization form and the supervision of the market.
In the prior art, the quotation strategy is mainly analyzed manually, and the main effect is large.
Disclosure of Invention
The invention aims to provide a quotation prediction model construction method, a prediction system, equipment and a medium, and aims to predict quotation behaviors of market participants by a power grid under the background of electric power market reformation, make measures for guaranteeing the operation safety of the power grid in advance and guide the market to develop healthily and quickly.
In order to achieve the purpose, the invention adopts the following technical scheme:
a quotation prediction model construction method comprises the following steps:
acquiring historical disclosure information of the power market, and selecting historical disclosure information corresponding to similar day data before a forecast day to form a first sample;
taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
aiming at each market member, a multilayer deep learning model is constructed, and a prediction model is obtained by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
Optionally, the historical disclosure information of the power market is obtained, and the historical disclosure information corresponding to the data of the similar day before the forecast day is selected to form a first sample; the method specifically comprises the following steps:
according to the historical disclosure information of the power market, selecting similar day data before a forecast day, obtaining future market disclosure information according to market space in the similar day data, dividing each day into a plurality of points, calculating the variance between the historical disclosure information of each point and the future market disclosure information, obtaining a sample of the historical disclosure information corresponding to the day with the minimum sum of the variances of each point, and forming a first sample.
Optionally, the constructing an input matrix by using the first sample and the second sample as inputs and constructing an output matrix by using historical offers of the first sample and the second sample as outputs specifically includes:
obtaining historical disclosure information of the power market as a second sample, copying the first sample for multiple times until the number of the first sample is not less than the number of the second sample to obtain a first sample after the weight is improved, and then taking the first sample and the second sample after the weight is improved as input and constructing the first sample and the second sample into an input matrix; and taking the historical quotations in the first sample and the second sample after the weight lifting as outputs to construct an output matrix.
Optionally, after the training set is trained, if the test set verification requirement is not met, adjusting parameters of the multi-layer deep learning model for retraining.
Optionally, the proportion of the number of the first samples to the total number of the second samples is not less than 30%, and the data of the similar day before the prediction day is the data of the similar day within 7 days before the prediction day.
Optionally, the training set and test set ratio is 9: 1.
A method of price quote prediction comprising the steps of:
acquiring historical disclosure information of the power market on a forecast day to form an input matrix;
inputting the input matrix into the electric power market quotation prediction model established by the method for prediction to obtain an output matrix;
and carrying out inverse normalization on the output matrix to obtain quotation prediction data.
A quote forecasting model construction system comprising:
the similar day data acquisition unit is used for acquiring historical disclosure information of the power market, and selecting the historical disclosure information corresponding to the similar day data before the forecast day to form a first sample;
the historical data acquisition unit is used for taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
the model training unit is used for constructing a multilayer deep learning model aiming at each market member and obtaining a prediction model by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
As a further improvement of the invention, the model training unit is further used for adjusting the parameters of the multi-layer deep learning model for retraining if the test set verification requirements are not met after the training of the training set.
As a further improvement of the present invention, the similar day data acquiring unit is specifically configured to:
according to the historical disclosure information of the power market, selecting similar day data before a forecast day, obtaining future market disclosure information according to market space in the similar day data, dividing each day into a plurality of points, calculating the variance between the historical disclosure information of each point and the future market disclosure information, obtaining a sample of the historical disclosure information corresponding to the day with the minimum sum of the variances of each point, and forming a first sample.
As a further improvement of the present invention, in the historical data obtaining unit, the first sample and the second sample are used as inputs and are constructed as an input matrix, and the historical quotations of the first sample and the second sample are used as outputs to construct an output matrix, which is specifically used for:
obtaining historical disclosure information of the power market as a second sample, copying the first sample for multiple times until the number of the first sample is not less than the number of the second sample to obtain a first sample after the weight is improved, and then taking the first sample and the second sample after the weight is improved as input and constructing the first sample and the second sample into an input matrix; and taking the historical quotations in the first sample and the second sample after the weight lifting as outputs to construct an output matrix.
An offer prediction system comprising:
the acquisition module is used for acquiring historical disclosure information of the power market on the forecast day and forming an input matrix;
the prediction module is used for inputting the input matrix into the electric power market quotation prediction model constructed according to the quotation prediction model construction system for prediction to obtain an output matrix;
and the normalization module is used for carrying out inverse normalization on the output matrix to obtain the quotation prediction.
An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the offer prediction model construction method when executing the computer program.
A computer-readable storage medium, storing a computer program which, when executed by a processor, implements the steps of the offer prediction model construction method.
The invention has the beneficial effects that:
the quotation prediction model construction method provided by the invention constructs the electric power market quotation prediction model through deep learning, can directly carry out quotation prediction, and is more accurate and faster than the traditional manual prediction. The construction method of the electric power market quotation prediction model in the method can be used for quantitatively processing complex nonlinear relations and associating strong association variables and weak association variables. The electric power market quotation prediction model can be used for predicting behaviors from historical quotation and power grid market disclosure information from an objective angle, and has small main observation influence and small personnel workload.
Furthermore, especially by calculating the variance of the historical disclosure information and the future market disclosure information of each point, the day with the minimum sum of variances can be obtained, the prediction is closer to the real prediction situation than the prediction directly through the historical disclosure information, and the prediction is more accurate and faster than the traditional manual prediction.
According to the prediction method, the input matrix is input into the pre-established electric power market quotation prediction model for prediction, and the quotation prediction data obtained by prediction can be used for predicting the electric power market quotation in the market environment, so that the power mechanism is helped to supervise the electric power market operation. The electric power market quotation is predicted through the pre-established electric power market quotation prediction model, the workload of personnel is reduced, and the method has the characteristics of high efficiency, convenience, rapidness and accuracy. The method can help the power grid to analyze the characteristics of the quotation strategy of market participants through deep learning, so as to predict the quotation behavior, make measures for guaranteeing the operation safety of the power grid in advance and guide the market to develop healthily and rapidly.
Drawings
FIG. 1 is a flow chart of a method of constructing a forecast model;
FIG. 2 is a flow chart of a method for forecasting electric power market quotes in accordance with the present invention;
FIG. 3 is a schematic structural diagram of a system for constructing a forecast model of an electric power market price according to an alternative embodiment of the present invention;
FIG. 4 is a schematic structural diagram of a quotation prediction model construction system according to an alternative embodiment of the present invention;
FIG. 5 is a schematic diagram of an electric power market quotation prediction system according to an alternative embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to an alternative embodiment of the invention.
Detailed Description
The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
The following detailed description is exemplary in nature and is intended to provide further details of the invention. Unless otherwise defined, all technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention.
Interpretation of terms
Deep Learning (Deep Learning): is a collection of algorithms that model high complexity data through multi-layer nonlinear transformations.
Electric power market: referred to herein as the power spot market and the ancillary services market.
And (3) quotation strategy: the market participant carries out the strategy of reporting the price according to the market rule.
Normalization: the method is a processing method aiming at the dimension, and simplifies the physical dimension with the unit into the relative relation without the dimension.
As shown in fig. 1, the invention provides a method for constructing a price quotation prediction model, which comprises the following steps:
acquiring historical disclosure information of the power market, and selecting historical disclosure information corresponding to similar day data before a forecast day to form a first sample;
taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
aiming at each market member, a multilayer deep learning model is constructed, and a prediction model is obtained by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
The concrete description is as follows:
s1, according to the historical disclosure information of the power market, selecting the data of the similar days before the forecast day, obtaining the future market disclosure information from the market space in the data of the similar days, dividing each day into a plurality of points, calculating the variance between the historical disclosure information of each point and the future market disclosure information, obtaining the sample of the historical disclosure information corresponding to the day with the minimum sum of the variances of each point, and forming a first sample.
The daily scale may be divided into a plurality of points, for example, 24 points and 96 points, according to different criteria.
The weight of the data of the similar days is improved: specifically, data close to the 7 th predicted day is selected according to the electric power market disclosure information, a sample with the minimum sum of the variances at each point is obtained according to the market space in the disclosure information, and the sample is copied according to the proportion of 30% of the total samples and added into a training sample.
S2, obtaining historical disclosure information as a second sample, obtaining historical disclosure information of the power market as the second sample, copying the first sample for multiple times until the number of the first sample is not less than the number of the second sample to obtain a first sample after weight lifting, and then taking the first sample and the second sample after weight lifting as input and constructing the first sample and the second sample into an input matrix; and taking the historical quotations in the first sample and the second sample after the weight lifting as outputs to construct an output matrix.
In actual calculation, copying the multiple of 2 to the first sample until the number of the first sample is not less than the number of the second samples so as to increase the weight of the first sample, and taking the first sample and the second sample after the weight is improved as input and constructing the input matrix; historical quotations in the first sample and the second sample after the weight is lifted are used as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
e.g., data collection, data for training the model is obtained and the market disclosure information is constructed as an input matrix.
Carrying out deep learning training model: and aiming at each market member, a multi-layer deep learning model is constructed, historical data is used for training, historical disclosure information is input, and historical report value is output. And respectively constructing input and output matrixes for input and output data, normalizing the input and output matrixes, dividing the input and output matrixes into a training set and a testing set, wherein the ratio is approximately 9:1, after training, according to the performance of the testing set, the electric power market quotation prediction model of the deep learning model is obtained when the performance meets the requirements, the input matrix formed by input predicted values is retrained by adjusting model parameters when the performance does not meet the requirements, and the output matrix is obtained by using the previously trained model to obtain the final electric power market quotation prediction model.
The proportion of the number of the first samples to the total number of the first samples and the second samples is not less than 30%, and the data of the similar days before the prediction day is the data of the similar days in 7 days before the prediction day. The proportion of the first sample is improved, the proportion of the data of the similar days is increased, and the authenticity and the accuracy of the training sample are improved.
The ratio of the training set to the test set is 9:1, the proportion of the training set is improved to increase the training effect, the training set can be verified by testing the segmented test set, and training errors caused by samples in the training set are avoided.
S3, aiming at each market member, constructing a multilayer deep learning model, and training by using an input matrix and an output matrix of a training set to obtain a prediction model; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
And verifying that the input matrix of the test set is input into the prediction model to obtain a corresponding output matrix, and comparing the obtained output matrix with the output matrix of the test set, wherein if the obtained output matrix is the same as the output matrix of the test set, the requirement is met, and if the obtained output matrix is different from the output matrix of the test set, the requirement is not met.
And if the test set verification requirements are not met after the training set is trained, adjusting parameters of the multi-layer deep learning model for retraining until the requirements are met.
As shown in fig. 2 and 3, a method for predicting an electric power market quotation includes the following steps:
acquiring historical disclosure information of the power market on a forecast day, and forming an input matrix;
inputting the input matrix into a prediction model according to the power market quotation for prediction to obtain an output matrix;
and carrying out inverse normalization on the output matrix to obtain the price prediction.
And finally obtaining a predicted quotation result: and predicting by using the trained electric power market quotation prediction model, and performing inverse normalization on the output matrix to obtain quotation.
Through deep learning, the method can predict the electric power market quotation in the market environment, and can provide help for the electric power mechanism to supervise the electric power market operation. The method needs less workload of personnel and has the characteristics of high efficiency and convenience. Compared with the traditional manual prediction, the method can quantize and process the complex nonlinear relation and simultaneously associate the strong association variable and the weak association variable.
As shown in fig. 4, the present invention provides a quotation prediction model building system, including:
the similar day data acquisition unit is used for acquiring historical disclosure information of the power market, and selecting the historical disclosure information corresponding to the similar day data before the forecast day to form a first sample;
the historical data acquisition unit is used for taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
the model training unit is used for constructing a multilayer deep learning model aiming at each market member and obtaining a prediction model by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
And the model training unit is also used for adjusting parameters of the multilayer deep learning model for retraining if the test set verification requirements are not met after the training of the training set.
The similar day data acquisition unit is specifically configured to:
according to the historical disclosure information of the power market, selecting similar day data before a forecast day, obtaining future market disclosure information according to market space in the similar day data, dividing each day into a plurality of points, calculating the variance between the historical disclosure information of each point and the future market disclosure information, obtaining a sample of the historical disclosure information corresponding to the day with the minimum sum of the variances of each point, and forming a first sample.
In the historical data obtaining unit, the first sample and the second sample are used as inputs and are constructed into an input matrix, and the historical quotations of the first sample and the second sample are used as outputs to construct an output matrix, which is specifically used for:
obtaining historical disclosure information of the power market as a second sample, copying the first sample for multiple times until the number of the first sample is not less than the number of the second sample to obtain a first sample after the weight is improved, and then taking the first sample and the second sample after the weight is improved as input and constructing the first sample and the second sample into an input matrix; and taking the historical quotations in the first sample and the second sample after the weight lifting as outputs to construct an output matrix.
As shown in fig. 5, another objective of the present invention is to provide an electric power market quotation prediction system, comprising:
the acquisition module is used for acquiring historical disclosure information of the power market on the forecast day and forming an input matrix;
the prediction module is used for inputting the input matrix into a prediction model according to the power market quotation for prediction to obtain an output matrix;
and the normalization module is used for carrying out inverse normalization on the output matrix to obtain the quotation prediction.
A third object of the present invention is to provide an electronic device, as shown in fig. 6, including a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the offer prediction model building method or the steps of the offer prediction method when executing the computer program.
The construction method of the electric power market quotation prediction model comprises the following steps:
acquiring historical disclosure information of the power market, and selecting historical disclosure information corresponding to similar day data before a forecast day to form a first sample;
taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
aiming at each market member, a multilayer deep learning model is constructed, and a prediction model is obtained by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
The electric power market quotation prediction method comprises the following steps:
acquiring historical disclosure information of the power market on a forecast day, and forming an input matrix;
inputting the input matrix into a prediction model according to the power market quotation for prediction to obtain an output matrix;
and carrying out inverse normalization on the output matrix to obtain the price prediction.
And finally obtaining a predicted quotation result: and predicting by using the trained electric power market quotation prediction model, and performing inverse normalization on the output matrix to obtain quotation.
A fourth object of the present invention is to provide a computer-readable storage medium storing a computer program, which when executed by a processor implements the steps of the offer prediction model construction method or the steps of the offer prediction method.
The construction method of the electric power market quotation prediction model comprises the following steps:
acquiring historical disclosure information of the power market, and selecting historical disclosure information corresponding to similar day data before a forecast day to form a first sample;
taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
aiming at each market member, a multilayer deep learning model is constructed, and a prediction model is obtained by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
The electric power market quotation prediction method comprises the following steps:
acquiring historical disclosure information of the power market on a forecast day, and forming an input matrix;
inputting the input matrix into a prediction model according to the power market quotation for prediction to obtain an output matrix;
and carrying out inverse normalization on the output matrix to obtain the price prediction.
And finally obtaining a predicted quotation result: and predicting by using the trained electric power market quotation prediction model, and performing inverse normalization on the output matrix to obtain quotation.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (14)
1. A quotation prediction model construction method is characterized by comprising the following steps:
acquiring historical disclosure information of the power market, and selecting historical disclosure information corresponding to similar day data before a forecast day to form a first sample;
taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
aiming at each market member, a multilayer deep learning model is constructed, and a prediction model is obtained by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
2. The method of claim 1,
the method comprises the steps of obtaining historical disclosure information of the power market, and selecting historical disclosure information corresponding to similar day data before a forecast day to form a first sample; the method specifically comprises the following steps:
according to the historical disclosure information of the power market, selecting similar day data before a forecast day, obtaining future market disclosure information according to market space in the similar day data, dividing each day into a plurality of points, calculating the variance between the historical disclosure information of each point and the future market disclosure information, obtaining a sample of the historical disclosure information corresponding to the day with the minimum sum of the variances of each point, and forming a first sample.
3. The method of claim 1,
the method for constructing the output matrix by taking the first sample and the second sample as inputs and taking the historical quotations of the first sample and the second sample as outputs specifically comprises the following steps:
obtaining historical disclosure information of the power market as a second sample, copying the first sample for multiple times until the number of the first sample is not less than the number of the second sample to obtain a first sample after the weight is improved, and then taking the first sample and the second sample after the weight is improved as input and constructing the first sample and the second sample into an input matrix; and taking the historical quotations in the first sample and the second sample after the weight lifting as outputs to construct an output matrix.
4. The method of claim 1,
and after the training set is trained, if the test set verification requirement is not met, adjusting parameters of the multilayer deep learning model for retraining.
5. The method of claim 1,
the proportion of the number of the first samples to the total number of the second samples is not less than 30%, and the data of the similar days before the prediction day is the data of the similar days in 7 days before the prediction day.
6. The method of claim 1,
the training set and test set ratio was 9: 1.
7. A method for forecasting an offer, comprising the steps of:
acquiring historical disclosure information of the power market on a forecast day to form an input matrix;
inputting the input matrix into an electric power market quotation prediction model established by the method of any one of claims 1 to 6 for prediction to obtain an output matrix;
and carrying out inverse normalization on the output matrix to obtain quotation prediction data.
8. A quotation prediction model construction system, comprising:
the similar day data acquisition unit is used for acquiring historical disclosure information of the power market, and selecting the historical disclosure information corresponding to the similar day data before the forecast day to form a first sample;
the historical data acquisition unit is used for taking the historical disclosure information of the power market as a second sample, taking the first sample and the second sample as input and constructing an input matrix, and taking the historical quotations of the first sample and the second sample as output to construct an output matrix; normalizing the input matrix and the output matrix, and dividing to form a training set and a test set;
the model training unit is used for constructing a multilayer deep learning model aiming at each market member and obtaining a prediction model by training an input matrix and an output matrix of a training set; and verifying through an input matrix and an output matrix of the test set, and taking the prediction model as an electric power market quotation prediction model after verification.
9. The system of claim 8,
and the model training unit is also used for adjusting parameters of the multilayer deep learning model for retraining if the test set verification requirements are not met after the training of the training set.
10. The system of claim 8,
the similar day data acquisition unit is specifically configured to:
according to the historical disclosure information of the power market, selecting similar day data before a forecast day, obtaining future market disclosure information according to market space in the similar day data, dividing each day into a plurality of points, calculating the variance between the historical disclosure information of each point and the future market disclosure information, obtaining a sample of the historical disclosure information corresponding to the day with the minimum sum of the variances of each point, and forming a first sample.
11. The system of claim 10,
in the historical data obtaining unit, the first sample and the second sample are used as inputs and are constructed into an input matrix, and the historical quotations of the first sample and the second sample are used as outputs to construct an output matrix, which is specifically used for:
obtaining historical disclosure information of the power market as a second sample, copying the first sample for multiple times until the number of the first sample is not less than the number of the second sample to obtain a first sample after the weight is improved, and then taking the first sample and the second sample after the weight is improved as input and constructing the first sample and the second sample into an input matrix; and taking the historical quotations in the first sample and the second sample after the weight lifting as outputs to construct an output matrix.
12. An offer prediction system, comprising:
the acquisition module is used for acquiring historical disclosure information of the power market on the forecast day and forming an input matrix;
the prediction module is used for inputting the input matrix into the electric power market quotation prediction model constructed by the quotation prediction model construction system according to any one of claims 8 to 11 for prediction to obtain an output matrix;
and the normalization module is used for carrying out inverse normalization on the output matrix to obtain the quotation prediction.
13. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the offer prediction model construction method according to any of claims 1-6 when executing the computer program.
14. A computer-readable storage medium storing a computer program which, when executed by a processor, implements the steps of the offer prediction model construction method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111300944.7A CN113837815A (en) | 2021-11-04 | 2021-11-04 | Quotation prediction model construction method, prediction method, system, device and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111300944.7A CN113837815A (en) | 2021-11-04 | 2021-11-04 | Quotation prediction model construction method, prediction method, system, device and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113837815A true CN113837815A (en) | 2021-12-24 |
Family
ID=78967265
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111300944.7A Pending CN113837815A (en) | 2021-11-04 | 2021-11-04 | Quotation prediction model construction method, prediction method, system, device and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113837815A (en) |
-
2021
- 2021-11-04 CN CN202111300944.7A patent/CN113837815A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021004324A1 (en) | Resource data processing method and apparatus, and computer device and storage medium | |
| US20200372342A1 (en) | Systems and methods for predictive early stopping in neural network training | |
| CN110717671B (en) | Method and device for determining contribution degree of participants | |
| CN108334954A (en) | Construction method, device, storage medium and the terminal of Logic Regression Models | |
| CN112308273A (en) | Memory, petrochemical enterprise pollution discharge management method, device and equipment | |
| Azimi et al. | Applying basic control theory principles to project control: Case study of off-site construction shops | |
| CN105956722A (en) | Short-term wind power prediction method and apparatus | |
| CN115906954A (en) | Multivariate time sequence prediction method and device based on graph neural network | |
| Mylonas et al. | Conditional variational autoencoders for probabilistic wind turbine blade fatigue estimation using Supervisory, Control, and Data Acquisition data | |
| CN111967655A (en) | Short-term load prediction method and system | |
| CN111415027A (en) | Method and device for constructing component prediction model | |
| CN114970926A (en) | Model training method, enterprise operation risk prediction method and device | |
| CN114219177A (en) | Computer room environment regulation and control method and device, electronic equipment and storage medium | |
| CN112257958A (en) | Power saturation load prediction method and device | |
| CN115049019B (en) | Method and device for evaluating arsenic adsorption performance of metal organic framework and related equipment | |
| CN113837815A (en) | Quotation prediction model construction method, prediction method, system, device and medium | |
| CN114840591A (en) | Method and device for determining sectional switch power data | |
| CN114154415A (en) | Equipment life prediction method and device | |
| CN113657656A (en) | Loan data analysis and prediction method and device | |
| CN113987261A (en) | Video recommendation method and system based on dynamic trust perception | |
| CN113159419A (en) | Group feature portrait analysis method, device and equipment and readable storage medium | |
| CN113139332A (en) | Automatic model construction method, device and equipment | |
| Zha et al. | Container throughput time series forecasting using a hybrid approach | |
| CN111026661A (en) | Method and system for comprehensively testing usability of software | |
| CN112434839B (en) | Distribution transformer heavy overload risk prediction method and electronic equipment |
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 |