Method, device and computer equipment for medium and long term electricity price forecasting suitable for electricity market

Abstract

The application relates to the field of electricity price prediction, and discloses a medium and long term electricity price prediction method, device and computer equipment suitable for a power market, which comprises the following steps: acquiring operation related time sequence data of the power market; obtaining first time sequence data on a first feature dimension from the operation related time sequence data according to the first feature dimension; wherein the first feature dimension is a feature dimension related to a bidding space; performing time sequence prediction on feature performance on the first feature dimension according to the first time sequence data to obtain bidding related time sequence data; and performing medium and long term electricity price prediction based on the bidding related time sequence data and the operation related time sequence data to obtain a predicted electricity price. Thus, by respectively performing stage-by-stage refined prediction on the bidding space related features and other operation related features, various factors influencing the electricity price can be more comprehensively captured, the accumulation and amplification of prediction errors are reduced, and therefore the final electricity price prediction result is more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electricity price prediction, and in particular to a medium and long term electricity price prediction method, device and computer equipment suitable for a power market. BACKGROUND

[0002] With the integration and development of the power market, the market transaction electricity is increasing, and the development of the power spot market is also accelerating. Spot trading helps to reveal the electricity price, while medium and long term trading can provide price protection for market participants, so it is crucial to predict the floating electricity price in the market in the future.

[0003] However, the method for predicting the floating electricity price in the market in the related art still has limitations, and a method for more accurately predicting the electricity price, especially the electricity price in medium and long term trading, is needed. SUMMARY

[0004] The present application aims to at least partially solve one of the technical problems in the related art. To this end, the present application proposes a medium and long term electricity price prediction method, device and computer equipment suitable for a power market. The main technical solution adopted by the present application includes:

[0005] In a first aspect, the present application provides a medium and long term electricity price prediction method suitable for a power market, which comprises: obtaining operation related time series data of the power market; obtaining first time series data on a first feature dimension from the operation related time series data according to the first feature dimension; wherein the first feature dimension is a feature dimension related to a bidding space; performing time series prediction on the feature performance on the first feature dimension according to the first time series data to obtain bidding related time series data; and performing medium and long term electricity price prediction based on the bidding related time series data and the operation related time series data to obtain a predicted electricity price.

[0006] Optionally, second time series data on a second feature dimension is obtained from the operation related time series data according to the second feature dimension; wherein the second feature dimension is a feature dimension unrelated to the bidding space. Correspondingly, performing medium and long term electricity price prediction based on the bidding related time series data and the operation related time series data to obtain the predicted electricity price comprises: performing medium and long term electricity price prediction based on the bidding related time series data, the first time series data and the second time series data to obtain the predicted electricity price.

[0007] Optionally, performing time series prediction on the feature performance on the first feature dimension according to the first time series data to obtain the bidding related time series data comprises: inputting the first time series data into an LSTM model; and performing time series prediction on the feature performance on the first feature dimension by the LSTM model to obtain the bidding related time series data.

[0008] Optionally, the second type of feature dimension includes any one of economic development, coal price, liquefied natural gas price, and climate change.

[0009] Optionally, the long-term and medium-term electricity price prediction based on the bidding-related time series data and the operation-related time series data comprises: performing feature dynamic enhancement processing on the bidding-related time series data and the operation-related time series data to obtain a prediction feature vector corresponding to the bidding-related time series data and a history feature vector corresponding to the operation-related time series data; and performing long-term and medium-term electricity price prediction based on the prediction feature vector, the history feature vector, the bidding-related time series data, and the operation-related time series data to obtain the predicted electricity price.

[0010] Optionally, the operation-related time series data further includes second time series data in a second type of feature dimension; the second type of feature dimension is a feature dimension irrelevant to the bidding space; and the feature dynamic enhancement processing on the bidding-related time series data and the operation-related time series data to obtain the prediction feature vector corresponding to the bidding-related time series data and the history feature vector corresponding to the operation-related time series data comprises: performing feature dynamic enhancement processing on the bidding-related time series data, the first time series data, and the second time series data to obtain a first feature vector corresponding to the first time series data, a second feature vector corresponding to the second time series data, and the prediction feature vector; and correspondingly, the long-term and medium-term electricity price prediction based on the prediction feature vector, the history feature vector, the bidding-related time series data, and the operation-related time series data to obtain the predicted electricity price comprises: performing long-term and medium-term electricity price prediction based on the first feature vector, the second feature vector, the prediction feature vector, the first time series data, the second time series data, and the bidding-related time series data to obtain the predicted electricity price.

[0011] Optionally, the first type of feature dimension includes any one of power load, wind power output, photovoltaic output, and tie line.

[0012] In a second aspect, an embodiment of the present application provides a long-term and medium-term electricity price prediction device suitable for a power market, which comprises: a time series data acquisition module configured to acquire operation-related time series data of the power market; a dimension data acquisition module configured to obtain first time series data in a first type of feature dimension from the operation-related time series data according to the first type of feature dimension; the first type of feature dimension is a feature dimension related to a bidding space; a time series prediction module configured to perform time series prediction on feature performance in the first type of feature dimension according to the first time series data to obtain bidding-related time series data; and an electricity price prediction module configured to perform long-term and medium-term electricity price prediction based on the bidding-related time series data and the operation-related time series data to obtain a predicted electricity price.

[0013] In a third aspect, the present application further provides a computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the method of any one of the above aspects when executing the computer program.

[0014] In a fourth aspect, the present application provides a computer readable storage medium, having stored thereon a computer program, which when executed by a processor implements the steps of the method of any one of the above.

[0015] In a fifth aspect, the present application provides a computer program product comprising a computer program which when executed by a processor implements the steps of the method of any one of the above.

[0016] In the above embodiments, firstly, the operation-related time series data in the electricity market is dimensionally divided, and the feature data of the bidding space-related feature dimension is time series predicted. Then, according to the bidding-related time series data obtained by time series predicting the feature performance on the first type of feature dimension, and other operation-related time series data in the electricity market, a secondary medium and long-term electricity price prediction is performed, and finally the predicted electricity price is obtained. At this point, through the phased and refined prediction method of firstly accurately predicting the key bidding space-related features, and then combining these prediction results with other operation-related features, not only various factors affecting the electricity price can be more comprehensively captured, but also the accumulation and amplification of prediction errors are reduced, so that the final electricity price prediction result is more accurate, and better meets the demand for medium and long-term electricity price prediction in the electricity market. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the specific embodiments or prior art of the present application, the drawings needed to be used in the description of the specific embodiments or prior art will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0018] Figure 1 Flow chart of the medium and long-term electricity price prediction method according to an embodiment of the present application;

[0019] Figure 2 Flow chart of the time series prediction method according to an embodiment of the present application;

[0020] Figure 3 Flow chart of the electricity price prediction method according to another embodiment of the present application;

[0021] Figure 4 Flow chart of the electricity price prediction method according to another embodiment of the present application;

[0022] Figure 5 Structural block diagram of a medium and long-term electricity price prediction device according to an embodiment of the present application;

[0023] Figure 6Fig. 1 is a diagram of an internal structure of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort should fall within the protection scope of the present application.

[0025] With the integration and development of the electricity market, the electricity traded in the market is increasing, and the development of the electricity spot market is accelerating. Spot trading helps to reveal the electricity price, while long-term trading can provide price protection for market participants. In this way, the electricity market trading system can be more perfect. Long-term trading has many advantages: it can stabilize market prices to some extent and enhance the predictability of future market changes for electricity sellers. At the same time, electricity buyers can ensure a certain electricity supply in advance through long-term contracts, thereby effectively controlling costs and reducing operational risks. In addition, multiple trading windows of long-term trading also provide arbitrage opportunities for market participants. Therefore, it is crucial to predict long-term electricity prices, because it can help electricity market participants make more reasonable investment and operation decisions, reduce market risks, and optimize resource allocation. Accurate electricity price prediction enables power producers to accurately grasp market trends and construct optimal electricity and price bidding strategies to maximize profits, while electricity consumers can develop reasonable electricity plans to reduce costs. At the same time, electricity price prediction provides a scientific basis for market regulators to promote healthy, stable, and orderly competition in the market.

[0026] However, compared with spot price prediction, long-term electricity price prediction faces more challenges due to greater price fluctuations and more complex influencing factors, including changes in fuel costs, increases in new energy installed capacity, and the impact of economic development on industrial electricity demand.

[0027] Therefore, according to an embodiment of the present application, a long-term electricity price prediction method suitable for the electricity market is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown.

[0028] In the present embodiment, a long-term electricity price prediction method suitable for the electricity market is provided, Figure 1is a flowchart of a method for medium and long-term electricity price prediction suitable for electricity markets according to an embodiment of the present application, as shown in Figure 1 The flow includes the following steps:

[0029] S110, obtaining operation-related time series data of the electricity market.

[0030] The operation-related time series data can refer to various time series data directly related to the operation and management of the power system in the electricity market. The operation-related time series data can be used to reflect the operation status and change trend of the electricity market at different time points. For example, the operation-related time series data can include power load, wind power output, photovoltaic output, tie-line, weather-related features, installed capacity, economic development, coal price, liquefied natural gas price, climate change, relevant policy updates, etc. Specifically, the operation-related time series data of the electricity market can be obtained from power companies and grid companies in real time through an automatic data acquisition system to obtain power load, power generation, market transactions, etc.

[0031] S120, obtaining first time series data on the first feature dimension from the operation-related time series data according to the first feature dimension.

[0032] The first feature dimension is a feature dimension related to the bidding space. Specifically, the first feature dimension can refer to a dimension mainly related to factors affecting the bidding process of the electricity market and the formation of electricity prices. For example, the first feature dimension can include power load reflecting the total power demand of all users in the power system within a certain time period; it can also include the output of various types of generating units such as wind power output and photovoltaic output that affect the supply capacity of the electricity market; it can also include tie-line transmission data reflecting the flow of electricity between different regions.

[0033] Further, the first time series data can refer to time series data corresponding to the first feature dimension extracted from the operation-related time series data. Specifically, the first time series data can reflect the change trend and regularity of the first feature over time. For example, taking power load as an example, the first time series data of the power load can show the load change at different time periods in a week, such as gradual increase in load during the week, peak at the weekend, and decrease to a low point the next week, etc.

[0034] Specifically, the feature dimension directly related to the bidding space can be determined from a large amount of operation-related time series data by focusing on key features, so as to further extract and screen the first time series data on the first feature dimension.

[0035] S130, performing time series prediction on the feature performance on the first feature dimension according to the first time series data to obtain bidding-related time series data.

[0036] The time series prediction refers to a data processing method for predicting data values at a certain time point or time period in the future according to historical time series data. Specifically, the time series prediction can predict the feature representation of a specific value or state of a specific feature in the first feature dimension at different time points. For example, taking the power load as an example, if the historical power load data is obtained, the time series prediction can be performed based on the autocorrelation, trend and seasonality of the time series data of the power load in the last week by combining the historical meteorological related features and historical data such as installed capacity, so as to obtain the future power load time series data at the power load feature dimension at the present time or in the future period. The power load time series data is the bidding related time series data. That is, the bidding related time series data can be time series data for reflecting the predicted performance of the bidding space related features in the future period.

[0037] Optionally, a time series prediction model can be trained in advance when performing the time series prediction. The time series prediction model can be an ARIMA model, a deep learning model, a long short-term memory network (LSTM) or the like. Data analysis and machine learning tools such as Pandas, NumPy, Scikit-learn, TensorFlow or Keras of Python can be used for data processing, model training and prediction. Further, a part of the historical data can be used as a validation set in the training process to evaluate the prediction performance of the constructed time series prediction model. By comparing the difference between the model prediction value and the actual value, the error index (such as mean square error, mean absolute error, etc.) is calculated, and the training and validation are repeatedly performed to ensure that the model has good prediction ability and generalization performance.

[0038] S140, performing medium and long term electricity price prediction based on the bidding related time series data and the operation related time series data to obtain a predicted electricity price.

[0039] Optionally, before performing the medium and long term electricity price prediction, the bidding related time series data and the operation related time series data can be integrated by alignment and synchronization to form a comprehensive data set, so as to ensure the consistency of different data sources in time. Specifically, in the data fusion process, the features that are helpful for electricity price prediction can be extracted and constructed. For example, the interaction features or derived features between different features can be calculated, such as the interaction feature of power load and temperature (for analyzing the influence of air conditioner power consumption on load), or the moving average of fuel price.

[0040] Further, after preprocessing all the bidding-related time series data and operation-related time series data, a pre-trained prediction model can be used to output the predicted value of the future electricity price by comprehensively considering the influence of various factors, i.e., to perform medium and long-term electricity price prediction based on the bidding-related time series data and the operation-related time series data, and to obtain the predicted electricity price.

[0041] It can be understood that the pre-trained prediction model can be a machine learning model (such as a random forest, a gradient boosting tree, or a support vector machine, etc.) and a deep learning model (such as a fully connected neural network, a convolutional neural network, or a recurrent neural network, etc.). Exemplarily, the prediction model can include a gated residual network, a Self-Attention (self-attention mechanism), and a fully connected layer, in full consideration of the interpretability, prediction ability, and computing resources of the model. In the training, the relationship between different features and the electricity price is learned constantly, so that the prediction model can capture the interaction and influence between the features, and the model parameters and structure are constantly adjusted in combination with the training results, so that the prediction model can accurately predict the future medium and long-term electricity price.

[0042] wherein the gating structure can control the degree of nonlinear contribution in the model, and the formula is as follows:

[0043]

[0044] In the formula, GLU represents the output of the gating linear unit; σ represents a nonlinear activation function; W and V represent different weight matrices obtained based on multiple training, which are used for linear transformation of input features; b and c represent different bias terms obtained based on multiple training, which are also used for linear transformation.

[0045] The residual network solves the gradient vanishing problem of the deep network through residual connection, and does not directly map the input to the output, but adds the input across part of the network layer to the output. Specifically, the input features can be first linearly transformed. In order to alleviate the gradient vanishing problem, the ELU activation function is applied, which maintains linearity in the positive domain and presents exponential growth nonlinearity in the negative domain. Then, the GLU activation function in the gating structure is applied to the data processed by the ELU activation function. Most importantly, the input features of the original input are added to the features processed by the activation function to form a residual connection. Finally, the result of the residual connection is subjected to layer normalization, and the output features processed by the residual network layer are output. The formula is as follows:

[0046] Y = LayerNorm(a + GLU((ELU(W'a + b')))

[0047] In the formula, Y represents the output of the residual network layer; a represents the input feature vector of the residual network layer; W' represents the weight matrix for linear transformation of the input features; b' represents the bias term, also for linear transformation; ELU represents a nonlinear activation function; GLU represents an activation function of a gating mechanism; and LayerNorm represents layer normalization.

[0048] Further, after passing through the gated residual network, it can also be input into a Self-Attention (self-attention mechanism), which can learn global dependencies and filter the importance of different features. Specifically, in the calculation process, for each token (i.e., each element in the input sequence), three vectors query (query vector Q), key (key vector K) and value (value vector V) are first generated, all of which are derived from the product of the input matrix X and a matrix (these vectors are obtained by multiplying the input matrix X with different weight matrices). It can be understood that they are essentially linear transformations of the input matrix X, but contain different data information. Among them, the query vector is analogous to a query, the key vector is analogous to an index, and the value vector is analogous to an answer. The Attention calculation formula in the form of key-value pair is as follows:

[0049]

[0050] In the formula, Attention(Q, K, V) represents the output of the self-attention mechanism; Q represents the query vector for retrieving relevant information from all key vectors; K represents the key vector for comparing with the query vector to determine the relevance of each value vector; V represents the value vector, which contains the actual information; softmax represents a probability distribution function; d represents the dimension of the key vector, represents a scaling factor to prevent the dot product result from being too large, thereby avoiding gradient disappearance when applying the softmax function.

[0051] Specifically, in the calculation process of Self-Attention, the query vector of each token is dot multiplied with the key vector of all tokens to obtain multiple attention scores representing the contribution degree of each token to the current token. Then these scores are standardized by a scaling factor and a softmax function to obtain the attention weights of different features. Further, these attention weights are also used to weight the corresponding value vectors to finally obtain the comprehensive representation of the current token. It can be understood that through the Self-Attention mechanism, the prediction model can learn the mutual relationship and importance between different time steps or features. For example, the prediction model may find that during the winter peak period, some features (such as heating demand) have a greater impact on electricity prices, and therefore give these features higher attention weights. In this way, the prediction model can dynamically adjust the attention degree to different time steps and features to more accurately predict future electricity price changes.

[0052] Optionally, during the training process, a part of the first time series data can also be used as a validation set to evaluate the prediction performance of the prediction model. By comparing the difference between the model prediction value and the actual value for multiple times, the error index (such as mean square error or mean absolute error, etc.) is calculated, so that the model is continuously optimized according to the evaluation results. At the same time, cross-validation or hyperparameter tuning methods can also be used to improve the generalization ability and prediction accuracy of the model.

[0053] In the above embodiments, first, the operation-related time series data in the electricity market is dimensionally divided, and the feature data of the feature dimension related to the bidding space is time series predicted. Then, according to the bidding-related time series data obtained by time series predicting the feature performance on the first type of feature dimension, and other operation-related time series data in the electricity market, a secondary medium and long-term electricity price prediction is performed, and finally a predicted electricity price is obtained. At this point, through the phased and refined prediction method of first accurately predicting the key features related to the bidding space and then combining these prediction results with other operation-related features, various factors affecting the electricity price can be more comprehensively captured, and the accumulation and amplification of prediction errors are reduced, so that the final electricity price prediction result is more accurate, and better meets the demand for medium and long-term electricity price prediction in the electricity market.

[0054] In some embodiments, the second time series data on the second type of feature dimension is obtained from the operation-related time series data according to the second type of feature dimension.

[0055] The second type of feature dimension is a feature dimension irrelevant to the bidding space. Specifically, the second type of feature dimension is opposite to the first type of feature dimension, and can refer to a dimension that does not directly participate in the bidding process but has an indirect impact on the electricity price. Exemplarily, it can be economic development, fuel price (such as coal price, liquefied natural gas price), climate change or related policy updates, etc. Correspondingly, the second time series data can refer to time series data corresponding to the second type of feature dimension extracted from the operation-related time series data. Exemplarily, the second time series data of economic development can be GDP growth rate or industrial output value, etc. changing over time; the second time series data of fuel price can be historical data of coal price, liquefied natural gas price, etc.

[0056] Further, the feature dimension irrelevant to the bidding space can also be determined from a large amount of operation-related time series data by focusing on key features, so as to extract and screen the second time series data on the second type of feature dimension.

[0057] Correspondingly, the long-term and medium-term electricity price prediction based on the bidding-related time series data and the operation-related time series data obtains a predicted electricity price, including:

[0058] The long-term and medium-term electricity price prediction based on the bidding-related time series data, the first time series data and the second time series data obtains a predicted electricity price.

[0059] Similarly, before the long-term and medium-term electricity price prediction, the bidding-related time series data, the first time series data and the second time series data can be aligned and synchronized for integration operation to form a comprehensive data set, and a prediction model can be used to comprehensively consider the influence of various factors to output a predicted value of the future electricity price, i.e., to obtain the predicted electricity price.

[0060] In the above embodiments, by dividing the entire feature dimension of the operation-related time series data into the second type of feature dimension and extracting the second time series data corresponding thereto, factors irrelevant to the bidding space but having an indirect impact on the electricity price can be more comprehensively captured, so as to provide more abundant and more accurate input data for the long-term and medium-term electricity price prediction, which helps to improve the capturing ability of the prediction model to market complex dynamics and enhance the reliability and accuracy of the prediction result.

[0061] In some embodiments, the performance of the features on the first type of feature dimension is time series predicted according to the first time series data to obtain the bidding-related time series data, please refer to the accompanying drawings Figure 2 , including:

[0062] S210, input the first time series data into the LSTM model.

[0063] Specifically, a part of the first time series data can be pre-divided as a training set, and the parameters of the LSTM model can be initialized according to the specific task requirements, including the number of network layers, the number of neurons in each layer, and the activation function, etc. At the same time, the training parameters of the model are set, such as learning rate and optimizer, etc., to train an LSTM model that can realize time series prediction.

[0064] It should be understood that LSTM (Long Short-Term Memory, i.e. long short-term memory network) is a special recurrent neural network structure that can be used to process and predict long-term dependencies in time series data.

[0065] Specifically, there are three gate mechanisms in the network structure of the LSTM (Long Short-Term Memory Network) model, namely the forget gate, the input gate and the output gate, which respectively determine which output information in the last time step needs to be discarded, which input information in the current time needs to be retained and combined, and which needs to be output.

[0066] Among them, the forget gate can be used to determine which information to retain from the hidden state of the last time step. Exemplarily, it can be analogized to determine which first time series data (such as historical power load information) is important and needs to be retained, and which information is outdated and can be discarded. The output calculation formula is:

[0067] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0068] In the formula, f t represents the output of the forget gate at the t-th time step; σ represents a nonlinear activation function, W f represents different weight matrices based on multiple training; b f represents a bias term based on multiple training; [h t-1 , xt ] represents the hidden state of the previous time step and the input feature (first time series data) of the current time step.

[0069] The input gate is composed of two parts, one of which determines which new information needs to be stored in the memory unit, and the other updates the memory unit state. Exemplarily, it can be analogized to determine which new information (such as the latest power demand or supply data contained in the first time series data) in the current time step is important and needs to be included in the model's memory. The output calculation formula is:

[0070] i t =σ(Wi ·[h t-1 , x t ]+b i )

[0071]

[0072] where i t represents the output of the input gate; is the candidate memory cell state; W i represents different weight matrices obtained based on multiple training; and b i represents bias terms obtained based on multiple training.

[0073] It can be understood that i t can be regarded as a decision factor that determines how much of the new information (such as the current power demand, supply situation or weather conditions, etc.) at the current time step should be incorporated into the memory of the model, and the decision is calculated based on the current input features and the previous hidden state (i.e., the power load prediction result at the previous time step). represents the new information that can be updated to the memory of the model at the current time step, which can include the current power market supply and demand situation, changes in new energy generation capacity or fluctuations in fuel costs, etc.

[0074] Further, in combination with the concepts of the input gate and the forget gate, the update of the memory cell of the LSTM at each time step can be expressed as follows:

[0075]

[0076] where represents the memory cell state at the current time step; f t represents the output of the input gate; represents the memory cell state at the previous time step; i t represents the output of the input gate; represents the candidate memory cell state.

[0077] The network structure also includes an output gate for determining how the information in the memory cell contributes to the hidden state at the next time step. Illustratively, it can be analogous to determining how to output the information in the memory cell (i.e., the combination of historical power load information and current power load information) to the next step of the model. The calculation formula is:

[0078] o t = σ(W o ·[h t-1 , x t ]+b o )

[0079] h t= o t tanh(C t )

[0080] wherein, o t represents the output of the output gate sigmoid function; C t represents the updated memory cell state; h t represents the final hidden state output.

[0081] It can be understood that the sigmoid function output of the output gate can be regarded as a "valve" that determines how much information in the memory cell state (i.e., the memory containing historical power load information and current market conditions) at each time step should be output to the next step of the model. t As the updated memory cell state, it is the core component of LSTM that stores long-term information, which is updated at each time step according to the output of the input gate and the forget gate. Exemplarily, it can be regarded as a "knowledge base" maintained by the LSTM model, which contains useful historical information and the integration of current market conditions, reflecting the model's memory of past power load changes and the possible impact of current market conditions on power load. It is updated at each time step.

[0082] And the final hidden state output is the output produced by LSTM at the end of each time step, which not only contains the information of the current time step, but also contains the information passed down by the memory cell from all previous time steps. It can be regarded as the "comprehensive understanding" of the model for the power load prediction of the current time step, which will be used as one of the input features for the next time step to help the model capture long-term dependencies in time series data. That is, the hidden state output is the result of the model's comprehensive analysis of current market conditions and historical power load information.

[0083] Further, after the LSTM model is trained, the first time series data also needs to be preprocessed to arrange into a three-dimensional array format containing sample quantity, time step length and feature dimension suitable for input of the LSTM model, and input into the LSTM model.

[0084] S220, performing time series prediction on the performance of the features in the first feature dimension through the LSTM model to obtain bidding-related time series data.

[0085] Specifically, through this pre-trained LSTM model, the performance of the features in the first feature dimension is predicted in time series to obtain bidding-related time series data reflecting the trend and regularity of each feature in the first feature dimension over time.

[0086] Optionally, the predicted bidding-related time series data can also be analyzed to evaluate the prediction performance of the model. Based on the analysis results, the parameters and structure of the LSTM model can be further adjusted and optimized to improve the accuracy and reliability of future predictions.

[0087] In the above embodiments, LSTM is used as a time series prediction model, which can effectively capture the time series variation and long-term dependence of the feature representation in the first type of feature dimension, thereby accurately predicting the bidding-related time series data. LSTM effectively reduces the gradient vanishing and gradient explosion problems when processing long sequence data through its unique gating mechanism and memory unit, enabling the model to better utilize historical information and current input features for comprehensive analysis and generate prediction results with high accuracy and reliability, providing a solid foundation for subsequent medium and long-term electricity price prediction.

[0088] In some embodiments, the second type of feature dimension includes any one of economic development, coal price, liquefied natural gas price, and climate change.

[0089] It can be understood that, since the second type of feature dimension refers to those dimensions that do not directly participate in the bidding process but have an indirect impact on electricity prices, it can include economic development, coal price, liquefied natural gas price, or climate change, etc. Similarly, the second type of feature dimension can also include technological progress factors such as power generation technology innovation or energy storage technology development, social factors such as population growth and population structure or changes in consumption habits, and international factors such as changes in supply and demand or price fluctuations in the international energy market, etc.

[0090] Taking economic development as an example, illustratively, in terms of industrial electricity demand changes, rapid economic development will be accompanied by an increase in industrial production and commercial activities, leading to an increase in industrial electricity demand. For example, during the economic growth period, manufacturing expansion or factory operation rate increases, leading to a significant increase in industrial electricity consumption, further increasing the total demand of the electricity market, thereby driving up electricity prices. Conversely, during economic recession, industrial production slows down, electricity demand decreases, and electricity prices may face downward pressure. Similarly, it is also reflected in changes in consumer spending levels, as economic development increases household income levels and consumer spending capacity, the demand for electricity will also increase. For example, each household may purchase more electrical equipment such as air conditioners or washing machines, thereby increasing residential electricity consumption and indirectly affecting electricity prices.

[0091] In the above embodiments, by including the second type of feature dimension such as economic development in the electricity price prediction model, factors that do not directly participate in the bidding process but have an important indirect impact on electricity prices can be more comprehensively captured. This helps to improve the accuracy and reliability of medium and long-term electricity price prediction, making the prediction results more reflective of the true dynamics and future trends of the market.

[0092] In some embodiments, the mid-long term electricity price prediction is performed based on the bidding-related time series data and the operation-related time series data, to obtain a predicted electricity price, please refer to the accompanying drawings Figure 3 , comprising:

[0093] S310, performing feature dynamic enhancement processing on the bidding-related time series data and the operation-related time series data, to obtain a predicted feature vector corresponding to the bidding-related time series data and a historical feature vector corresponding to the operation-related time series data.

[0094] S320, performing mid-long term electricity price prediction using the predicted feature vector, the historical feature vector, the bidding-related time series data and the operation-related time series data, to obtain a predicted electricity price.

[0095] Wherein, the feature dynamic enhancement processing can be a method of feature extraction and optimization processing for time series data. The feature dynamic enhancement processing can strengthen the capture ability of the prediction model for key features by dynamically adjusting the expression ability and importance of the features, so as to obtain a feature vector containing important information. Similarly, the predicted feature vector can be a feature vector containing key feature information in the bidding-related time series data after model extraction and enhancement processing. Similarly, the historical feature vector can be a feature vector containing key feature information in the operation-related time series data after model extraction and enhancement processing.

[0096] Specifically, the feature dynamic enhancement processing on the bidding-related time series data and the operation-related time series data can be realized by using a gated residual network and a Self-Attention mechanism. Illustratively, the obtained bidding-related time series data and operation-related time series data can be preprocessed first to ensure the integrity and consistency of the data. Then the feature dynamic enhancement processing is performed by the gated residual network in the prediction model. The gating mechanism controls the degree of nonlinear transformation by introducing a gating unit (such as a gated linear unit, GLU), which flexibly adjusts the expression of the features. The residual connection allows the network to combine features at different levels, i.e. not directly mapping the input features to the output, but adding the input directly to the output across part of the network layers, thereby enhancing the processing ability of the prediction model for time series information, while avoiding the gradient vanishing problem in deep network training. Further, based on the characteristics of the input bidding-related time series data and operation-related time series data, the gated residual network can dynamically adjust the strength of feature extraction, thereby better capturing the dynamic changes in the time series data, and finally screening out the feature vector that contributes most to the predicted electricity price,

[0097] It can be understood that the Self-Attention (self-attention mechanism) is further connected after the gated residual network. Since the Self-Attention (self-attention mechanism) can capture the dependency between any two positions in the input sequence, not just the local or adjacent positions, it can identify the features that have the greatest impact on the predicted electricity price in the entire time series. Based on the interrelationship between the features and the degree of contribution to the prediction target, the attention score between each feature and other features is calculated, thereby giving different weights to each feature, and finally obtaining the final attention weight. Finally, the weights are used to weight the sum of the value vectors to obtain the comprehensive representation of each feature.

[0098] Further, after obtaining the comprehensive representation of each feature, it can be used for medium and long term electricity price prediction to obtain the predicted electricity price. Specifically, first, the bidding-related time series data and the operation-related time series data processed by the gated residual network and the Self-Attention are integrated. At this time, each feature has obtained its comprehensive representation, which contains the global importance in the time series and the weight value of the contribution degree to the prediction target. Further, using the fully connected layer of the prediction model, based on the mapping relationship between the input features and the predicted electricity price, the final prediction result is obtained, that is, the future electricity price is predicted using the model, and the final predicted electricity price result is output.

[0099] In the above embodiment, by combining the gated residual network and Self-Attention for feature dynamic enhancement processing and feature importance evaluation, the features that most contribute to the prediction of electricity price in the bidding-related time series data and the operation-related time series data are accurately identified and utilized, and the weights of the features are dynamically adjusted to enhance the expression ability of the key features, effectively capture the dynamic changes and long-term dependencies in the time series data, thereby fully playing the role of these features in medium and long term electricity price prediction, and finally improving the accuracy and stability of the prediction.

[0100] In some embodiments, the operation-related time series data further includes second time series data in a second type of feature dimension.

[0101] The second type of feature dimension is a feature dimension that is irrelevant to the bidding space.

[0102] The bidding-related time series data and the operation-related time series data are subjected to feature dynamic enhancement processing to obtain a predicted feature vector corresponding to the bidding-related time series data and a historical feature vector corresponding to the operation-related time series data. Please refer to Figure 4 , including:

[0103] S410, performing feature dynamic enhancement processing on the bidding related time series data, the first time series data and the second time series data to obtain a first feature vector corresponding to the first time series data, a second feature vector corresponding to the second time series data, and a prediction feature vector.

[0104] The first feature vector refers to a feature vector obtained by performing feature dynamic enhancement processing on the first time series data extracted from the first type of feature dimension (a feature dimension related to the bidding space), which reflects the importance of the features in the first type of feature dimension in the prediction process. The second feature vector refers to a feature vector obtained by performing feature dynamic enhancement processing on the second time series data extracted from the second type of feature dimension (a feature dimension unrelated to the bidding space). It reflects the importance of the features in the second type of feature dimension in the prediction process.

[0105] Similarly, performing feature dynamic enhancement processing on the bidding related time series data, the first time series data and the second time series data can also be realized by using a gated residual network and a Self-Attention (self-attention mechanism). Using the gated residual network in the prediction model, the flow of information is automatically controlled to flexibly adjust the expression of features, dynamically adjust the weight of input data, and obtain a comprehensive representation of each feature. That is, a first feature vector corresponding to the first time series data, a second feature vector corresponding to the second time series data, and a prediction feature vector are obtained.

[0106] Correspondingly, using the prediction feature vector, the historical feature vector, the bidding related time series data and the operation related time series data to perform medium and long term electricity price prediction to obtain a predicted electricity price, comprising:

[0107] S420, performing medium and long term electricity price prediction according to the first feature vector, the second feature vector, the prediction feature vector, the first time series data, the second time series data and the bidding related time series data to obtain a predicted electricity price.

[0108] Similarly, after obtaining the comprehensive representation of each feature, the fully connected layer of the prediction model can be used to predict the future electricity price based on the mapping relationship between the input features and the predicted electricity price, based on the first feature vector, the second feature vector, the prediction feature vector, the first time series data, the second time series data and the bidding related time series data, that is, the final predicted electricity price result can be output.

[0109] In the above embodiments, by subdividing the operation-related time series data into the first type of feature dimensions related to the bidding space and the second type of feature dimensions unrelated to the bidding space, and performing feature dynamic enhancement processing on the two types of feature data, the change trend and law of factors unrelated to the bidding space but indirectly affecting the electricity price can be more comprehensively captured. With the aid of the gated residual network and the self-attention mechanism, the model can dynamically adjust the feature weights to generate the first feature vector, the second feature vector and the predicted feature vector reflecting the importance of each feature. These enhanced feature vectors combined with the original time series data provide more abundant information for medium and long term electricity price prediction, thereby significantly improving the accuracy and reliability of the prediction and better meeting the demand for medium and long term electricity price prediction in the electricity market.

[0110] In some embodiments, the first type of feature dimensions includes any one of power load, wind power output, photovoltaic output, and tie-line.

[0111] It can be understood that, since the first type of feature dimensions represents dimensions mainly involving factors affecting the bidding process of the electricity market and the formation of electricity prices, it can include power load, wind power output, photovoltaic output, or tie-line, etc. Further, the first type of feature dimensions can also include fuel cost and operation and maintenance cost of thermal power generation, reservoir storage and river flow of hydropower generation, and a series of generation-side factors, can also include power grid transmission capacity factors such as transmission line capacity or transformer capacity of the power grid, and can also include market participant factors such as sensitivity of power users to price changes and demand elasticity.

[0112] Still taking power load as an example, it is exemplarily illustrated that, since power load is a direct embodiment of demand in the electricity market, the increase or decrease of power load directly affects the balance of supply and demand of electricity. When power load increases, industrial production expands or resident electricity consumption may rise, thereby the demand for electricity in the market rises. At this time, if the supply cannot keep up in time, it will lead to power shortage, thereby pushing up the electricity price. Conversely, when power load decreases, such as economic activity slowdown or implementation of energy-saving measures, etc., the power supply is relatively surplus at this time, and the electricity price may fall. In addition, the volatility of power load will also cause the volatility of electricity price. For example, in the peak electricity consumption period (such as summer high temperature or winter heating, etc.), power load may surge, leading to tight power supply, so the electricity price in these periods will usually rise; while in the valley electricity consumption period, the load decreases, leading to relatively low electricity price.

[0113] In the above embodiments, by refining the first type of feature dimension into specific factors such as power load, wind power output, photovoltaic output, and tie line, the key factors affecting the power market bidding process and power price formation can be more accurately captured and analyzed, enabling the prediction model to assess the specific impact of each factor on power supply and demand balance, market competitiveness, and price fluctuations, thereby improving the accuracy and reliability of the prediction results.

[0114] The embodiments of the present specification also provide a method for medium and long-term electricity price prediction suitable for a power market, comprising the following steps:

[0115] S802, obtaining operation-related time series data of the power market.

[0116] S804, obtaining first time series data on the first type of feature dimension from the operation-related time series data according to the first type of feature dimension.

[0117] The first type of feature dimension is a feature dimension related to the bidding space. The first type of feature dimension includes any one of power load, wind power output, photovoltaic output, and tie line.

[0118] S806, obtaining second time series data on the second type of feature dimension from the operation-related time series data according to the second type of feature dimension.

[0119] The second type of feature dimension is a feature dimension unrelated to the bidding space. The second type of feature dimension includes any one of economic development, coal price, liquefied natural gas price, and climate change.

[0120] S808, inputting the first time series data into an LSTM model.

[0121] S810, performing time series prediction of the feature performance on the first type of feature dimension by the LSTM model to obtain bidding-related time series data.

[0122] S812, performing feature dynamic enhancement processing on the bidding-related time series data, the first time series data, and the second time series data to obtain a first feature vector corresponding to the first time series data, a second feature vector corresponding to the second time series data, and a predicted feature vector.

[0123] S814, performing medium and long-term electricity price prediction according to the first feature vector, the second feature vector, the predicted feature vector, the first time series data, the second time series data, and the bidding-related time series data to obtain a predicted electricity price.

[0124] It should be understood that although the steps in the above flowcharts are shown in sequence according to the directions of the arrows, the steps are not necessarily executed in the order of the directions of the arrows. Unless otherwise specified herein, the execution of the steps is not strictly limited in sequence, and the steps can be executed in other sequences. Moreover, at least some of the steps of the above flowcharts can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of the steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0125] The embodiments of the present specification also provide a mid-long term electricity price prediction device 500 suitable for a power market, as shown in the figure, comprising: a time series data acquisition module 510, a dimension data acquisition module 520, a time series prediction module 530 and an electricity price prediction module 540, wherein: Figure 5

[0126] The time series data acquisition module 510 is configured to acquire operation-related time series data of the power market.

[0127] The dimension data acquisition module 520 is configured to obtain first time series data on a first feature dimension from the operation-related time series data according to the first feature dimension; wherein the first feature dimension is a feature dimension related to a bidding space.

[0128] The time series prediction module 530 is configured to perform time series prediction on the feature performance on the first feature dimension according to the first time series data, to obtain bidding-related time series data.

[0129] The electricity price prediction module 540 is configured to perform mid-long term electricity price prediction based on the bidding-related time series data and the operation-related time series data, to obtain a predicted electricity price.

[0130] In some embodiments, second time series data on a second feature dimension is obtained from the operation-related time series data according to the second feature dimension; wherein the second feature dimension is a feature dimension irrelevant to the bidding space. The dimension data acquisition module 520 is further configured to correspondingly perform mid-long term electricity price prediction based on the bidding-related time series data and the operation-related time series data, to obtain the predicted electricity price, comprising: performing mid-long term electricity price prediction based on the bidding-related time series data, the first time series data and the second time series data, to obtain the predicted electricity price.

[0131] In some embodiments, the time series prediction module 530 is further configured to input the first time series data into an LSTM model; and perform time series prediction on the feature performance on the first feature dimension through the LSTM model, to obtain the bidding-related time series data.

[0132] ​In some embodiments, the dimension data obtaining module 520 is further configured to determine that the second type of feature dimension comprises any one of economic development, coal price, liquefied natural gas price, and climate change.

[0133] In some embodiments, the electricity price prediction module 540 is further configured to perform feature dynamic enhancement processing on the bidding-related time series data and the operation-related time series data to obtain a predicted feature vector corresponding to the bidding-related time series data and a historical feature vector corresponding to the operation-related time series data; and perform medium and long term electricity price prediction by using the predicted feature vector, the historical feature vector, the bidding-related time series data, and the operation-related time series data to obtain the predicted electricity price.

[0134] In some embodiments, the operation-related time series data further comprises second time series data in a second type of feature dimension; the second type of feature dimension is a feature dimension irrelevant to the bidding space; and the feature dynamic enhancement processing on the bidding-related time series data and the operation-related time series data to obtain the predicted feature vector corresponding to the bidding-related time series data and the historical feature vector corresponding to the operation-related time series data comprises: performing feature dynamic enhancement processing based on the bidding-related time series data, the first time series data, and the second time series data to obtain a first feature vector corresponding to the first time series data, a second feature vector corresponding to the second time series data, and the predicted feature vector. The dimension data obtaining module 520 is further configured to correspondingly perform medium and long term electricity price prediction by using the predicted feature vector, the historical feature vector, the bidding-related time series data, and the operation-related time series data to obtain the predicted electricity price, comprising: performing medium and long term electricity price prediction according to the first feature vector, the second feature vector, the predicted feature vector, the first time series data, the second time series data, and the bidding-related time series data to obtain the predicted electricity price.

[0135] In some embodiments, the dimension data obtaining module 520 is further configured to determine that the first type of feature dimension comprises any one of power load, wind power output, photovoltaic power output, and tie line.

[0136] The specific limitation of the medium and long term electricity price prediction device suitable for the electricity market can be referred to the limitation of the medium and long term electricity price prediction method suitable for the electricity market as described above, which will not be repeated here. Each module in the medium and long term electricity price prediction device suitable for the electricity market can be realized by software, hardware, and combinations thereof, in whole or in part. Each module described above can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to each module.

[0137] The medium and long term electricity price prediction device suitable for the electricity market in the embodiment is presented in the form of functional units, where the units refer to ASIC (Application Specific Integrated Circuit) circuits, processors and memories that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0138] The embodiment of the present application further provides a computer device, which can be a terminal, and an internal structure diagram of the computer device can be as shown in Figure 6 The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to perform wired or wireless communication with an external terminal, and the wireless communication can be achieved through WIFI, a mobile cellular network, NFC (Near Field Communication) or other technologies. The computer program is executed by the processor to implement a medium and long term electricity price prediction method suitable for an electricity market. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer overlaid on the display screen, or can be a key, trackball or touchpad arranged on the shell of the computer device, or can be an external keyboard, touchpad or mouse, etc.

[0139] Those skilled in the art can understand that Figure 6 The structure shown in the above figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. Specifically, the computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0140] The embodiments of the present application further provide a computer readable storage medium, and the method according to the embodiments of the present application can be implemented in hardware, firmware, or recorded in a storage medium, or stored in a remote storage medium or a non-transitory machine readable storage medium and downloaded to a local storage medium by network downloading of computer code, so that the method described herein can be processed by such software on a storage medium using a general purpose computer, a special purpose processor, or programmable or special hardware. The storage medium can be a disk, a compact disk, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid state disk, etc. Further, the storage medium can also include a combination of the above-mentioned memories. It can be understood that the computer, the processor, the microprocessor controller, or the programmable hardware includes a storage component that can store or receive software or computer code, when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the method shown in the above embodiments is implemented.

[0141] The embodiments of the present application provide a computer program product, which includes computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method of any of the embodiments of the present application.

[0142] The method, device and computer device for medium and long term electricity price prediction suitable for the electricity market described in the above embodiments can be implemented by a computer chip or an entity, or by a product having certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0143] For the convenience of description, the above device is described as various units in terms of functions. Of course, the functions of the units can be implemented in the same or more software and / or hardware in the implementation of the present application.

[0144] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0145] The computer program instructions can 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 block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps in the flowchart block or blocks.

[0146] These computer program instructions can 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 instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps in the flowchart block or blocks.

[0147] The computer program instructions can 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 block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps in the flowchart block or blocks.

[0148] In the description of the specification, the description of the terms "one embodiment", "some embodiments”, "example”, "specific example”, or "some examples” etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the description of the specification, the illustrative description of the above terms does not necessarily mean the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0149] In addition, the terms "first", "second" are used only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise explicitly specified.

[0150] It should also be noted that the terms "comprising", "comprises" or other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises a..." does not, without further constraints, exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0151] Various embodiments in the present specification are described in progressive manner, and the same or similar parts among the various embodiments can be mutually referred to, and each embodiment focuses on the difference from other embodiments. Since it is basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiments.

[0152] The above only describes the embodiments of the present application and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the scope of claims of the present application.

[0153] Although the embodiments of the present application are described in conjunction with the drawings, various modifications and changes can be made by those skilled in the art without departing from the spirit and scope of the present application, and such modifications and changes fall within the scope defined by the appended claims.

Claims

1. A method for mid-long term electricity price forecasting suitable for electricity market, characterized in that, The method comprises: obtaining operation-related time series data of a power market; obtaining first time series data on a first feature dimension from the operation-related time series data according to the first feature dimension; wherein the first feature dimension is a feature dimension related to a bidding space; the first feature dimension comprises power load reflecting total power demand of all users in a power system within a specific time period, power output of various types of generator sets such as wind power output and photovoltaic output affecting supply capacity of the power market, and tie-line transmission data reflecting power flow between different regions; performing time series prediction on feature performance on the first feature dimension according to the first time series data to obtain bidding-related time series data; wherein the bidding-related time series data is time series data for reflecting predicted performance of the bidding space-related features within a future period of time; obtaining second time series data on a second feature dimension from the operation-related time series data according to the second feature dimension; wherein the second feature dimension is a feature dimension irrelevant to the bidding space, and the second feature dimension does not directly participate in the bidding process but has an indirect impact on the electricity price; the second time series data comprises historical data of time-varying GDP growth rate or industrial output value or coal price or liquefied natural gas price; performing feature dynamic enhancement processing on the bidding-related time series data, the first time series data and the second time series data to obtain a first feature vector corresponding to the first time series data, a second feature vector corresponding to the second time series data, and a predicted feature vector corresponding to the bidding-related time series data; performing medium and long term electricity price prediction according to the first feature vector, the second feature vector, the predicted feature vector, the first time series data, the second time series data and the bidding-related time series data to obtain a predicted electricity price.

2. The method of claim 1, wherein, The method comprises: inputting the first time series data into an LSTM model; performing time series prediction on feature performance on the first feature dimension by the LSTM model to obtain the bidding-related time series data.

3. The method of claim 1, wherein, The second feature dimension comprises any one of economic development, coal price, liquefied natural gas price, and climate change.

4. An apparatus for mid-long term electricity price forecasting suitable for electricity markets, characterized by, The device comprises: a time series data acquisition module configured to acquire operation-related time series data of a power market; a dimension data acquisition module configured to obtain first time series data on a first feature dimension from the operation-related time series data according to the first feature dimension; wherein the first feature dimension is a feature dimension related to a bidding space; the first feature dimension comprises power load reflecting total power demand of all users in a power system within a specific time period, power output of various types of generator sets such as wind power output and photovoltaic output affecting supply capacity of the power market, and tie-line transmission data reflecting power flow between different regions; a time series prediction module configured to perform time series prediction on the performance of features in the first feature dimension according to the first time series data, to obtain bidding-related time series data; wherein the bidding-related time series data is time series data reflecting the predicted performance of features in a bidding space over a future period of time; a price prediction module configured to obtain second time series data in the second feature dimension from the operation-related time series data according to the second feature dimension; wherein the second feature dimension is a feature dimension irrelevant to the bidding space, and the second feature dimension does not directly participate in the bidding process but has an indirect impact on the electricity price; and the second time series data includes historical data of time-varying GDP growth rate, industrial output value, coal price or liquefied natural gas price; performing feature dynamic enhancement processing on the bidding-related time series data, the first time series data and the second time series data, to obtain a first feature vector corresponding to the first time series data, a second feature vector corresponding to the second time series data, and a predicted feature vector corresponding to the bidding-related time series data; performing medium and long term electricity price prediction according to the first feature vector, the second feature vector, the predicted feature vector, the first time series data, the second time series data and the bidding-related time series data, to obtain a predicted electricity price.

5. A computer device, comprising: comprising: a memory and a processor in communication connection with each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method of any one of claims 1 to 3.

6. A computer readable storage medium characterized by, the computer readable storage medium stores computer instructions for causing a computer to perform the method of any one of claims 1 to 3.

Images