Carbon emission permit prediction method and system based on empirical mode decomposition and neural network

Abstract

The invention discloses a carbon emission permit prediction method and system based on empirical mode decomposition and a neural network. The method comprises the steps of obtaining carbon emission permit data at different time and performing preprocessing; performing empirical mode decomposition on the preprocessed carbon emission permit data to extract sequence features; performing regression learning on the sequence features by using a convolutional neural network, learning a nonlinear relationship between the sequence features and the price tags, and constructing a carbon emission right prediction model according to the nonlinear relationship; and taking the carbon emission permit data of the continuous time section as the input of the carbon emission permit prediction model, outputting the carbon emission permit data of the future time section, and predicting the condition of the future time section according to the carbon emission permit data of the current time section. The method can better adapt to the nonlinear and non-stationary characteristics of the signal, the convolutional layer and the pooling layer in the CNN structure enable the model to capture the local features in the image, and the method has the advantages of being easy to implement and high in prediction precision.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emission rights prediction, and in particular to a carbon emission rights prediction method and system based on empirical mode decomposition and neural network. Background Art

[0002] As the global climate change problem becomes increasingly serious, carbon emission trading, as an effective market mechanism, has promoted the sustainable development of the economy while reducing greenhouse gas emissions. In China, the carbon emission trading pilot has been rolled out nationwide, aiming to control and reduce greenhouse gas emissions such as carbon dioxide through market means. However, due to the wide distribution of China's key emission enterprises and the lack of liquidity of the carbon trading market between pilot provinces and cities, the carbon trading prices in different regions vary significantly, which not only increases the operating costs of enterprises, but also brings challenges to the formulation of market strategies for enterprises.

[0003] In order to help key emission enterprises better understand and predict the changing trend of carbon emission rights prices, domestic and foreign scholars mainly start from three aspects: traditional econometric models, machine learning models, and decomposition and integration models to predict carbon emission rights prices. Although traditional econometric models such as ARIMA and VAR can capture some characteristics of time series, they perform poorly when dealing with nonlinear and non-stationary data; although machine learning models have incomparable advantages over traditional econometric models in predicting carbon emission rights prices, machine learning models have obvious disadvantages in terms of time consumption and parameter sensitivity, which will greatly limit the speed and accuracy of machine learning in predicting carbon emission rights prices. In view of the shortcomings of machine learning, many researchers have begun to study new prediction ideas. The study found that the decomposition and integration model reduces the complexity of prediction by decomposing complex time series into multiple modal components and using different models to make relevant predictions for each component. This new prediction model can present the internal data characteristics of the time series more completely, which reduces the difficulty of subsequent prediction work to the greatest extent and is conducive to improving the prediction effect. Therefore, it is necessary to propose a carbon emission rights prediction method based on empirical mode decomposition (EMD) and neural network (NN), which can more accurately capture the inherent laws of time series and thus achieve effective prediction of carbon emission rights prices. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a carbon emission rights prediction method and system based on empirical mode decomposition and neural network, which not only overcomes the limitations of traditional prediction models in processing nonlinear data, but also solves the overfitting or underfitting problems that a single model may encounter in complex time series prediction.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides a carbon emission rights prediction method based on empirical mode decomposition and neural networks, including: obtaining carbon emission rights data at different times and preprocessing them; performing empirical mode decomposition on the preprocessed carbon emission rights data to extract sequence features; using a convolutional neural network to perform regression learning on the sequence features, learning the nonlinear relationship between the sequence features and price tags, and constructing a carbon emission rights prediction model based on the nonlinear relationship; using the carbon emission rights data of continuous time sections as the input of the carbon emission rights prediction model, outputting the carbon emission rights data of future time sections, and realizing the prediction of future time sections by the carbon emission rights data of the current time section.

[0008] As a preferred solution of the carbon emission rights prediction method based on empirical mode decomposition and neural network described in the present invention, wherein: performing empirical mode decomposition on the preprocessed carbon emission rights data includes:

[0009] Creating an empirical mode decomposition object, performing empirical mode decomposition on the preprocessed carbon emission rights data by using an empirical mode decomposition algorithm, that is, extracting extreme points of the original signal, and obtaining the first round of approximate intrinsic mode function by connecting the mean values ​​of the extreme points;

[0010] Subtracting the first round of approximate intrinsic mode functions from the original signal to obtain a new signal, extracting extreme points and connecting the means of the new signal again to obtain the second round of approximate intrinsic mode functions, and so on until the obtained approximate intrinsic mode functions cover all data;

[0011] Among them, each inherent intrinsic mode function represents an oscillation component of different scales in the original signal, and each component of the inherent intrinsic mode function represents each frequency component in the original signal, and is arranged in sequence from high frequency to low frequency.

[0012] As a preferred solution of the carbon emission rights prediction method based on empirical mode decomposition and neural network described in the present invention, the extraction of sequence features includes:

[0013] Arranging the decomposed intrinsic mode functions into a matrix by rows, and transposing the matrix to construct a characteristic matrix;

[0014] A window size window_size is defined, and the feature matrix is ​​traversed. Each traversal takes a row continuous segment of the length of the window size window_size as a feature, and a data point after the corresponding time window is taken as a label, and a training database based on carbon emission rights is constructed to extract sequence features.

[0015] As a preferred solution of the carbon emission rights prediction method based on empirical mode decomposition and neural network described in the present invention, wherein: the regression learning of the sequence features using a convolutional neural network includes:

[0016] The convolutional neural network includes two convolutional layers, a maximum pooling layer, a Flatten layer and two fully connected layers, wherein the first convolutional layer includes 128 filters and the second convolutional layer includes 64 filters. Both convolutional layers use a convolution kernel of size 3 and adopt ReLU as an activation function to capture local features of input data. The step size of the convolution operation is 1.

[0017] The features passed through the convolution layer are input into the maximum pooling layer for dimensionality reduction. The size of the pooling window is 2. The pooled data is flattened into a one-dimensional vector and input into the fully connected layer.

[0018] As a preferred solution of the carbon emission rights prediction method based on empirical mode decomposition and neural network described in the present invention, wherein: the regression learning of the sequence features using a convolutional neural network also includes:

[0019] The first fully connected layer has 64 neurons and uses ReLU as the activation function. It receives the pooled data of the maximum pooling layer and performs linear transformation and nonlinear activation on it to capture the complex pattern of the data. The second fully connected layer has 1 neuron, which is used to output the nonlinear relationship between the sequence features and the price tags.

[0020] As a preferred solution of the carbon emission rights prediction method based on empirical mode decomposition and neural network described in the present invention, the prediction of the future time section by the carbon emission rights data of the current time section includes:

[0021] Generate training data based on carbon emission rights for the training database based on carbon emission rights, and train the carbon emission rights prediction model based on the training data. After the training is completed, use the carbon emission rights data of four consecutive time sections as the input of the carbon emission rights prediction model, and output the carbon emission rights data of a future time section to predict the situation of the future section based on the carbon emission rights data of the current section.

[0022] As a preferred solution of the carbon emission rights prediction method based on empirical mode decomposition and neural network described in the present invention, wherein: the carbon emission rights data at different times are obtained and preprocessed, wherein the preprocessing process includes:

[0023] Cut the long time series to separate the daily carbon emission rights time series;

[0024] Analyze, clean and supplement the separated daily carbon emission rights time series;

[0025] Determine the preprocessing time range. Within the time range, if the carbon emission rights time series is always zero, it is judged as abnormal and the data is removed. If the carbon emission rights time series is missing only at a certain sampling point, it is supplemented by the average value of the sample data.

[0026] In a second aspect, the present invention provides a carbon emission rights prediction system based on empirical mode decomposition and neural network, comprising:

[0027] Data collection and processing module, used to obtain carbon emission rights data at different times and perform preprocessing;

[0028] A feature extraction module, used for performing empirical mode decomposition on the preprocessed carbon emission rights data to extract sequence features;

[0029] A regression learning module, used for performing regression learning on the sequence features using a convolutional neural network, obtaining a nonlinear relationship between the sequence features and the price tag, and constructing a carbon emission rights prediction model based on the nonlinear relationship;

[0030] The identification prediction module is used to use the carbon emission rights data of continuous time sections as the input of the carbon emission rights prediction model, output the carbon emission rights data of future time sections, and realize the prediction of future time sections by the carbon emission rights data of the current time section.

[0031] In a third aspect, the present invention provides a computing device, comprising:

[0032] Memory and processor;

[0033] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the carbon emission rights prediction method based on empirical mode decomposition and neural network are implemented.

[0034] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the carbon emission rights prediction method based on empirical mode decomposition and neural network.

[0035] Compared with the prior art, the present invention has the following beneficial effects: the present invention provides a carbon emission rights prediction method and system based on empirical mode decomposition and neural network, wherein the empirical mode decomposition method is used for signal processing and data analysis, can be decomposed according to the local characteristics of the signal, does not require pre-setting of basis functions or filters, and is applicable to various signal types. Unlike traditional frequency domain analysis methods, the EMD method does not require prior knowledge of the spectral characteristics of the signal, and can better adapt to the nonlinear and non-stationary characteristics of the signal. EMD decomposes the signal into a series of intrinsic mode functions (IMFs), each IMF represents an intrinsic vibration mode in the signal, has a clear physical meaning, and is helpful for understanding and interpreting the signal characteristics; the convolutional layers and pooling layers in the CNN structure enable the model to capture local features in the image, and have a certain degree of invariance to changes in the position of the target in the image, which helps to improve the robustness of the model to the input data. The parameter sharing mechanism in CNN reduces the number of parameters that need to be trained, making the model more concise and more generalizable, thereby improving the robustness of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0037] Figure 1 This is a schematic diagram of the overall process logic of a carbon emission rights prediction method based on empirical mode decomposition and neural network according to an embodiment of the present invention;

[0038] Figure 2 This is an overall flow chart of a carbon emission rights prediction method based on empirical mode decomposition and neural network according to an embodiment of the present invention;

[0039] Figure 3 A schematic diagram of an empirical mode decomposition process of a carbon emission rights prediction method based on empirical mode decomposition and neural network according to an embodiment of the present invention;

[0040] Figure 4 A schematic diagram of an empirical mode decomposition result of a carbon emission rights prediction method based on empirical mode decomposition and neural network according to an embodiment of the present invention;

[0041] Figure 5 A schematic diagram of the structure of a convolutional neural network of a carbon emission rights prediction method based on empirical mode decomposition and neural network according to an embodiment of the present invention;

[0042] Figure 6This is a graph of prediction results of a carbon emission rights prediction method based on empirical mode decomposition and neural network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0044] Example 1

[0045] Reference Figure 1-Figure 5 As an embodiment of the present invention, a carbon emission rights prediction method based on empirical mode decomposition and neural network is provided, such as Figure 1 The specific steps shown include:

[0046] S100: Obtain carbon emission rights data at different times and perform preprocessing;

[0047] S200: Performing empirical mode decomposition on the preprocessed carbon emission rights data to extract sequence features;

[0048] S300: Use convolutional neural networks to perform regression learning on sequence features, learn the nonlinear relationship between sequence features and price tags, and build a carbon emission rights prediction model based on the nonlinear relationship;

[0049] S400: Using the carbon emission rights data of continuous time sections as the input of the carbon emission rights prediction model, outputting the carbon emission rights data of future time sections, and realizing the prediction of the future time sections by the carbon emission rights data of the current time section.

[0050] It should be noted that the present invention provides a carbon emission rights prediction method and system based on empirical mode decomposition and neural network, wherein the empirical mode decomposition method is used for signal processing and data analysis, can be decomposed according to the local characteristics of the signal, does not require pre-setting basis functions or filters, and is applicable to various signal types. Unlike traditional frequency domain analysis methods, the EMD method does not require a priori understanding of the spectral characteristics of the signal, and can better adapt to the nonlinear and non-stationary characteristics of the signal. EMD decomposes the signal into a series of intrinsic mode functions (IMFs), each IMF represents an inherent vibration mode in the signal, has a clear physical meaning, and is helpful for understanding and interpreting the signal characteristics; the convolution layer and pooling layer in the CNN structure enable the model to capture local features in the image, and have a certain degree of invariance to the position change of the target in the image, which helps to improve the robustness of the model to the input data. The parameter sharing mechanism in CNN reduces the number of parameters that need to be trained, making the model more concise and more generalizable, thereby improving the robustness of the model.

[0051] In the embodiment of the present application, the above step S100 includes the following sub-steps A1-A2;

[0052] In A1: obtain carbon emission rights data at different times;

[0053] In A2: preprocessing the carbon emission rights data obtained at different times;

[0054] Specifically, the preprocessing process includes:

[0055] Cut the long time series to separate the daily carbon emission rights time series;

[0056] Analyze, clean and supplement the separated daily carbon emission rights time series;

[0057] Determine the preprocessing time range. Within the time range, if the carbon emission rights time series is always zero, it is judged as abnormal and the data is removed. If the carbon emission rights time series is missing only at a certain sampling point, it is supplemented by the average value of the sample data.

[0058] It should be noted that the above step S100 not only ensures the quality and integrity of the data, but also effectively removes abnormal data, maintains the temporal continuity of the data, and avoids the situation where the overall analysis results are affected by the missing of individual data points; this preprocessing process can provide a more accurate and reliable data basis for subsequent carbon emission rights price forecasts, thereby improving the stability of the forecasting model and the forecasting accuracy.

[0059] In the embodiment of the present application, the above step S200 performs empirical mode decomposition on the pre-processed carbon emission rights data to extract sequence features including:

[0060] Specifically, Figure 2 The empirical mode decomposition of the pre-processed carbon emission rights data is shown as follows:

[0061] Create an empirical mode decomposition object, and perform empirical mode decomposition on the preprocessed carbon emission rights data using the empirical mode decomposition algorithm, that is, extract the extreme points of the original signal, and obtain the first round of approximate intrinsic mode function by connecting the mean values ​​of the extreme points;

[0062] The first round of approximate intrinsic mode function is subtracted from the original signal to obtain a new signal, and the extreme point extraction and mean connection are performed again on the new signal to obtain the second round of approximate intrinsic mode function, such as Figure 3 As shown in the figure, the approximate intrinsic mode function is obtained until it covers all data. The EMD decomposition result of the carbon emission volume time series is as follows: Figure 4 As shown;

[0063] Among them, each inherent intrinsic mode function represents the oscillation components of different scales in the original signal, and each component of the inherent intrinsic mode function represents the frequency components in the original signal, and is arranged in order from high frequency to low frequency.

[0064] It should be noted that the EMD is used to construct the training data fingerprint information, in which the empirical mode decomposition method is an adaptive signal time-frequency processing method, which is particularly suitable for the analysis and processing of nonlinear and non-stationary signals. Its essence is a method of decomposing the signal, decomposing the signal into the superposition of each independent component, and performing signal decomposition based on the time scale characteristics of the data itself, which is adaptive. EMD believes that any complex sequence is composed of multiple single-frequency signals, so it can be decomposed into several intrinsic mode functions. Each component of IMF represents the frequency components in the original signal, and is arranged in order from high frequency to low frequency, which is also the physical meaning of IMF. Intrinsic mode functions (IMF) are the signal components of each layer obtained after the original signal is decomposed by EMD. Any signal can be decomposed into the sum of several IMFs. IMF has two assumptions: (1) in the entire data segment, the number of extreme points and the number of zero-crossing points must be equal or differ by no more than one; (2) at any time, the average value of the upper envelope formed by the local maximum points and the average value of the lower envelope formed by the local minimum points is zero, that is, the upper and lower envelopes are locally symmetric with respect to the time axis.

[0065] Furthermore, the decomposed intrinsic mode functions are arranged into a matrix by rows, and the matrix is ​​transposed to construct a feature matrix; a window_size of 50 is defined, and the feature matrix is ​​traversed, and each traversal takes a row continuous segment of the length of the window size window_size as a feature, and a data point after the corresponding time window is taken as a label to construct a training database based on carbon emission rights to extract sequence features.

[0066] It should be noted that the above step S200 can effectively decompose the original data into multiple intrinsic mode functions with different frequency characteristics by performing empirical mode decomposition on the pre-processed carbon emission rights data, which can not only reveal the multi-scale structure in the data, but also filter out noise, making the intrinsic characteristics of the data clearer; each IMF and trend item obtained after EMD decomposition represents the dynamic behavior of the data at different time scales, providing a richer and more accurate source of information for subsequent feature extraction and model training, and helping to improve the model's understanding and prediction of the trend of carbon emission rights price changes. In addition, this multi-scale analysis method can better capture the periodicity and nonlinear characteristics of carbon emission rights price fluctuations, laying a solid foundation for building an efficient prediction model.

[0067] In the embodiment of the present application, the convolutional neural network in the above step S300 includes two convolutional layers, a maximum pooling layer, a Flatten layer and two fully connected layers, such as Figure 5 Specifically shown are:

[0068] Input layer: The shape of the input data is (96,2), which means that each sample has 96 time steps and each time step contains 2 features;

[0069] The first convolution layer includes 128 filters, and the second convolution layer includes 64 filters. Both convolution layers use a convolution kernel of size 3 and ReLU as the activation function to capture the local features of the input data. The stride of the convolution operation is 1.

[0070] The features after the convolution layer are input into the maximum pooling layer for dimensionality reduction. The size of the pooling window is 2. The pooled data is flattened into a one-dimensional vector and input into the fully connected layer.

[0071] The first fully connected layer has 64 neurons and uses ReLU as the activation function. It receives the data after pooling in the maximum pooling layer and performs linear transformation and nonlinear activation on it to capture the complex patterns of the data. The second fully connected layer has 1 neuron, which is used to output the nonlinear relationship between the sequence features and the price tags.

[0072] The mean square error loss function and Adam optimizer are used, and the learning rate is set to 0.001.

[0073] It should be noted that the convolutional layer (Convolutional Layer): The convolution operation is the core of CNN. The convolutional layer extracts features on the input data through a sliding window (convolution kernel). The convolution operation can capture local features and retain the spatial structure, allowing the network to learn the patterns in the image. Pooling layer (Pooling Layer): The pooling layer is used to reduce the spatial dimension of the feature map, reduce the number of parameters, and retain important information. Common pooling operations include max pooling and average pooling. Activation function (Activation Function): A nonlinear activation function, such as ReLU (Rectified Linear Unit), is usually added between the convolutional layer and the fully connected layer to introduce nonlinearity and increase the expression ability of the network. Fully connected layer (Fully Connected Layer): The fully connected layer converts the feature maps extracted by the convolutional layer and the pooling layer into the final output. The fully connected layer is usually used for classification tasks, and the softmax function is used in the output layer for classification prediction. Loss Function: CNN usually uses the cross entropy loss function to measure the difference between the model output and the true label, and updates the network parameters through the back propagation algorithm to minimize the loss function. Back propagation algorithm: The derivative of the loss function with respect to the network parameters is calculated through the back propagation algorithm, and then the parameters are updated using optimization methods such as gradient descent to continuously optimize the performance of the network. In general, CNN gradually extracts and abstracts the features of the input data through the stacking of multiple layers of convolution and pooling layers, and finally implements tasks such as classification or regression through the fully connected layer. This hierarchical structure and parameter sharing design enable CNN to perform well in processing large-scale data sets and complex tasks, and become one of the important technologies in the field of deep learning.

[0074] It should be noted that the above step S300 uses a convolutional neural network to perform regression learning on the sequence features extracted by empirical mode decomposition, which can deeply explore the complex nonlinear relationship between the sequence features and the price tags. The powerful representation learning ability of the convolutional neural network enables it to automatically identify and learn the local dependencies and long-term dependencies in the data, thereby building a more accurate and robust carbon emission rights prediction model. This method not only overcomes the problem that traditional linear models cannot effectively handle nonlinear data, but also improves the generalization ability of the model through multi-level abstract representation, so that the model can make reliable predictions when facing new and unseen data, which is of great significance to improving the accuracy and practicality of carbon emission rights price prediction.

[0075] In the embodiment of the present application, the above step S400 predicts the situation of the future time section based on the carbon emission rights data of the current time section, including:

[0076] For the training database based on carbon emission rights, training data based on carbon emission rights is generated, and the carbon emission rights prediction model is trained based on the training data. After the training is completed, the carbon emission rights data of four consecutive time sections are used as the input of the carbon emission rights prediction model, and the carbon emission rights data of a future time section is output to realize the prediction of the future section based on the carbon emission rights data of the current section.

[0077] It should be noted that the above step S400 can not only make full use of the time dependency and pattern in the historical data, but also effectively apply the nonlinear relationship learned by the model in the training phase to the actual prediction, thereby improving the accuracy and reliability of the prediction. In this way, the model can dynamically respond to the changing trend of the carbon emission rights market price, provide timely and accurate decision support for enterprises and policy makers, and also help optimize resource allocation and promote the healthy development of the carbon market.

[0078] The above is a schematic scheme of the carbon emission rights prediction method based on empirical mode decomposition and neural network in this embodiment. It should be noted that the technical scheme of the system for predicting carbon emission rights based on empirical mode decomposition and neural network and the technical scheme of the carbon emission rights prediction method based on empirical mode decomposition and neural network belong to the same concept. The details of the technical scheme of the carbon emission rights prediction system based on empirical mode decomposition and neural network in this embodiment that are not described in detail can be found in the description of the technical scheme of the carbon emission rights prediction method based on empirical mode decomposition and neural network.

[0079] This embodiment provides a carbon emission rights prediction system based on empirical mode decomposition and neural network, including:

[0080] Data collection and processing module, used to obtain carbon emission rights data at different times and perform preprocessing;

[0081] A feature extraction module is used to perform empirical mode decomposition on the preprocessed carbon emission rights data to extract sequence features;

[0082] The regression learning module is used to perform regression learning on sequence features using convolutional neural networks, obtain the nonlinear relationship between sequence features and price tags, and build a carbon emission rights prediction model based on the nonlinear relationship;

[0083] The identification and prediction module is used to use the carbon emission rights data of continuous time sections as the input of the carbon emission rights prediction model, output the carbon emission rights data of future time sections, and realize the prediction of future time sections by the carbon emission rights data of the current time section.

[0084] The above-mentioned unit modules may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to the above-mentioned modules.

[0085] This embodiment also provides a computing device, which includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, 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 the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a carbon emission rights prediction method based on empirical mode decomposition and neural network is implemented. 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 covered on the display screen, or a button, trackball or touchpad set on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0086] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method proposed in the above embodiment is implemented.

[0087] The storage medium proposed in this embodiment and the method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0088] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0089] Example 2

[0090] Reference Figure 6 Based on the previous embodiment, this embodiment provides an application example of a carbon emission rights prediction method and system based on empirical mode decomposition and neural network, in order to verify and illustrate the technical effects adopted in this method.

[0091] The data set of this embodiment includes the transaction price data of carbon emission rights in the Hubei carbon emission rights trading market from April 2, 2014 to March 1, 2024.

[0092] First, the time series of carbon emission rights transaction prices at different times are decomposed by EMD to obtain several intrinsic mode functions (IMFs). After the IMFs are organized into feature matrices, the training set and test set are constructed using a sliding window method. Finally, a deep neural network is used for regression learning to learn the nonlinear relationship between the main features and price tags, and the prediction results of the carbon emission rights price prediction model are obtained, such as Figure 6 shown.

[0093] It can be seen from the results that the present invention first ensures the data quality and integrity of the input model through the fine preprocessing of the original data, laying a solid foundation for subsequent analysis; then uses the empirical mode decomposition technology to extract multi-scale features of the data, effectively reveals the inherent laws of carbon emission rights price changes, and enhances the model's ability to capture the nonlinear characteristics of the data; then uses a convolutional neural network to perform deep learning on the extracted features, which can not only automatically discover the complex nonlinear relationship between sequence features and price tags, but also improve the learning efficiency and generalization ability of the model, ensuring the accuracy and stability of the prediction model; finally, by using the data of continuous time sections as the model input, a dynamic prediction of future carbon emission rights prices is achieved, which provides important decision-making basis for market participants, helps to optimize resource allocation, and promotes the healthy and orderly development of the carbon trading market.

[0094] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0095] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt 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.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0096] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0097] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0098] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0099] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0100] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A carbon emission rights prediction method based on empirical mode decomposition and neural network, characterized in that: include: Obtain carbon emission rights data at different times and perform preprocessing; Performing empirical mode decomposition on the preprocessed carbon emission rights data to extract sequence features; Using a convolutional neural network to perform regression learning on the sequence features, learn the nonlinear relationship between the sequence features and the price tags, and build a carbon emission rights prediction model based on the nonlinear relationship; The carbon emission rights data of continuous time sections are used as the input of the carbon emission rights prediction model, and the carbon emission rights data of future time sections are output, so as to predict the situation of future time sections by the carbon emission rights data of the current time section.

2. The carbon emission rights prediction method based on empirical mode decomposition and neural network according to claim 1, characterized in that: Performing empirical mode decomposition on the preprocessed carbon emission rights data includes: Creating an empirical mode decomposition object, performing empirical mode decomposition on the preprocessed carbon emission rights data by using an empirical mode decomposition algorithm, that is, extracting extreme points of the original signal, and obtaining the first round of approximate intrinsic mode function by connecting the mean values ​​of the extreme points; Subtracting the first round of approximate intrinsic mode functions from the original signal to obtain a new signal, extracting extreme points and connecting the means of the new signal again to obtain the second round of approximate intrinsic mode functions, and so on until the obtained approximate intrinsic mode functions cover all data; Among them, each inherent intrinsic mode function represents an oscillation component of different scales in the original signal, and each component of the inherent intrinsic mode function represents each frequency component in the original signal, and is arranged in sequence from high frequency to low frequency.

3. The carbon emission rights prediction method based on empirical mode decomposition and neural network according to claim 2, characterized in that: The extracting sequence features comprises: Arranging the decomposed intrinsic mode functions into a matrix by rows, and transposing the matrix to construct a characteristic matrix; A window size window_size is defined, and the feature matrix is ​​traversed. Each traversal takes a row continuous segment of the length of the window size window_size as a feature, and a data point after the corresponding time window is taken as a label, and a training database based on carbon emission rights is constructed to extract sequence features.

4. The carbon emission rights prediction method based on empirical mode decomposition and neural network according to claim 3, characterized in that: The method of performing regression learning on the sequence features by using a convolutional neural network includes: The convolutional neural network includes two convolutional layers, a maximum pooling layer, a Flatten layer and two fully connected layers, wherein the first convolutional layer includes 128 filters and the second convolutional layer includes 64 filters. Both convolutional layers use a convolution kernel of size 3 and adopt ReLU as an activation function to capture local features of input data. The step size of the convolution operation is 1. The features passed through the convolution layer are input into the maximum pooling layer for dimensionality reduction. The size of the pooling window is 2. The pooled data is flattened into a one-dimensional vector and input into the fully connected layer.

5. The carbon emission rights prediction method based on empirical mode decomposition and neural network according to claim 4, characterized in that: The step of performing regression learning on the sequence features using a convolutional neural network also includes: The first fully connected layer has 64 neurons and uses ReLU as the activation function. It receives the pooled data of the maximum pooling layer and performs linear transformation and nonlinear activation on it to capture the complex pattern of the data. The second fully connected layer has 1 neuron, which is used to output the nonlinear relationship between the sequence features and the price tags.

6. The carbon emission rights prediction method based on empirical mode decomposition and neural network according to claim 5, characterized in that: The situation of predicting the future time section by the carbon emission rights data of the current time section includes: Generate training data based on carbon emission rights for the training database based on carbon emission rights, and train the carbon emission rights prediction model based on the training data. After the training is completed, use the carbon emission rights data of four consecutive time sections as the input of the carbon emission rights prediction model, and output the carbon emission rights data of a future time section to predict the situation of the future section based on the carbon emission rights data of the current section.

7. The carbon emission rights prediction method based on empirical mode decomposition and neural network according to claim 1, characterized in that: The carbon emission rights data at different times are obtained and preprocessed, wherein the preprocessing process includes: Cut the long time series to separate the daily carbon emission rights time series; Analyze, clean and supplement the separated daily carbon emission rights time series; Determine the preprocessing time range. Within the time range, if the carbon emission rights time series is always zero, it is judged as abnormal and the data is removed. If the carbon emission rights time series is missing only at a certain sampling point, it is supplemented by the average value of the sample data.

8. A system using the carbon emission rights prediction method based on empirical mode decomposition and neural network as described in any one of claims 1 to 7, characterized in that: include: Data collection and processing module, used to obtain carbon emission rights data at different times and perform preprocessing; A feature extraction module, used for performing empirical mode decomposition on the preprocessed carbon emission rights data to extract sequence features; A regression learning module, used for performing regression learning on the sequence features using a convolutional neural network, obtaining a nonlinear relationship between the sequence features and the price tag, and constructing a carbon emission rights prediction model based on the nonlinear relationship; The identification prediction module is used to use the carbon emission rights data of continuous time sections as the input of the carbon emission rights prediction model, output the carbon emission rights data of future time sections, and realize the prediction of future time sections by the carbon emission rights data of the current time section.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the carbon emission rights prediction method based on empirical mode decomposition and neural network as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the carbon emission rights prediction method based on empirical mode decomposition and neural network as described in any one of claims 1 to 7.

Images