A regional carbon emission prediction method and system based on deep learning model

Abstract

The present invention relates to the technical field of carbon emission prediction, and discloses a regional carbon emission prediction method and system based on a deep learning model. The real-time data and historical data of regional carbon emissions are acquired through a data acquisition module, and the historical data of regional carbon emissions are integrated and adapted to the transmission of the real-time data of regional carbon emissions; the model construction module constructs and trains a carbon emission prediction model through the integrated historical data of regional carbon emissions to obtain the loss value of the carbon emission prediction model so as to effectively predict the regional carbon emissions; the carbon emission prediction module inputs the transmitted real-time data of regional carbon emissions into the carbon emission prediction model to generate a prediction result, and compares the actual value of the regional carbon emissions with the prediction result to evaluate the carbon emission prediction model; the prediction optimization module optimizes the carbon emission prediction model according to the evaluation result; and the visual interaction module visually displays the real-time data and historical data of regional carbon emissions, the prediction result and the evaluation result in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emission prediction, and in particular to a regional carbon emission prediction method and system based on a deep learning model. Background Art

[0002] With the continuous development of deep learning, deep learning models can process large amounts of complex data. At the same time, the mixed application of different deep learning models can improve the accuracy and practicality of predictions and achieve more accurate carbon emissions predictions. However, most of them have not solved how to obtain regional carbon emission data to build a carbon emission prediction model for carbon emission prediction, evaluate the carbon emission prediction model, generate evaluation results to optimize the carbon emission prediction model, and visualize the carbon emission data and prediction results in real time.

[0003] For example, a Chinese patent application with publication number CN117495640A discloses a method and system for predicting regional carbon emissions, the method comprising obtaining a historical carbon emission data set and an initial influencing factor data set; performing a regional correlation analysis based on the carbon emission processing data set and the initial influencing factor processing data set, and performing a first screening of the initial influencing factor processing data set; performing a variable penalty based on the carbon emission processing data set and the first influencing factor data set, so as to perform a second screening of the first influencing factor data set; inputting the training data set into a preset prediction model for training, and inputting the second influencing factor data set into the trained preset prediction model for prediction, so as to obtain carbon emission prediction data. By screening the influencing factor data set twice, the preset prediction model can fully explore the nonlinear mapping relationship between the data, and at the same time can screen out factors that have no effect or little effect on carbon emissions, thereby improving the prediction speed and prediction accuracy of subsequent models.

[0004] For example, a Chinese patent with authorization announcement number CN116258277B discloses a method and system for predicting, planning and optimizing regional energy carbon emissions, which includes: performing a second-order decomposition on the current status of the regional energy carbon emission system, obtaining the trend baseline of the energy carbon emission system under the baseline scenario, and then setting the scenario parameters under scenario S to obtain the trend line of the energy carbon emission system under scenario S, and calculating the present value of the total investment cost of the path corresponding to the development path of scenario S; storing the path corresponding to scenario S in the feasible development path pool, and simultaneously storing its corresponding and rigid indicator scenario values, and when the rigid indicator constraints are met, sorting each scenario with the minimum as the single goal, and screening out the best regional development path. This method and system can be applied to the application paradigm of the departmental structure, realize the uniformity of carbon emission measurement in different regions, and can be used to guide targeted and effective regional planning work.

[0005] The above patent has the problem raised by this background technology: the above regional carbon emission prediction method performs regional correlation analysis on the acquired historical carbon emission data set and the initial influencing factor data set, performs a first screening on the initial influencing factor processing data set, performs variable penalty based on the carbon emission processing data set and the first influencing factor data set, performs a second screening on the first influencing factor data set, inputs the training data set into the preset prediction model for training, and inputs the second influencing factor data set into the trained preset prediction model for prediction to obtain carbon emission prediction data. By screening the influencing factor data set twice, the preset prediction model can fully mine the nonlinear relationship between the data. The above-mentioned regional energy carbon emission prediction, planning and optimization method performs a second-order decomposition on the current status of the regional energy carbon emission system to obtain the trend baseline of the energy carbon emission system under the baseline scenario, and then sets the scenario parameters under the scenario to obtain the trend line of the energy carbon emission system under the scenario, and calculates the present value of the total investment cost of the path corresponding to the development path of the scenario, and stores the path corresponding to the scenario in the feasible development path pool, and simultaneously stores the corresponding rigid indicator scenario value. When the rigid indicator constraint condition is met, the scenarios are sorted with the minimum as the single goal, so as to screen out the best regional development path. The above two patents do not solve the problem of how to obtain regional carbon emission data to construct a carbon emission prediction model for carbon emission prediction, and evaluate the carbon emission prediction model, generate evaluation results to optimize the carbon emission prediction model, and visualize the carbon emission data and prediction results in real time. To solve this problem, the present invention proposes a regional carbon emission prediction method and system based on a deep learning model. Summary of the invention

[0006] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0007] In view of the problems existing in the above-mentioned existing regional carbon emission prediction method and system based on deep learning model, the present invention is proposed.

[0008] Therefore, the object of the present invention is to provide a regional carbon emission prediction method and system based on a deep learning model.

[0009] In order to solve the above technical problems, the present invention provides a regional carbon emission prediction system based on a deep learning model: a data acquisition module, a model building module, a carbon emission prediction module, a prediction optimization module and a visual interaction module;

[0010] The data acquisition module is used to acquire real-time data and historical data of regional carbon emissions, integrate the historical data of regional carbon emissions and adapt to the transmission of real-time data of regional carbon emissions, and integrate the historical data of regional carbon emissions including dynamically integrating the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions;

[0011] The model building module is used to build and train a carbon emission prediction model through the fused historical data of regional carbon emissions to obtain a loss value of the carbon emission prediction model;

[0012] The functional expression of the constructed carbon emission prediction model is as follows:

[0013] ;

[0014] In the formula, represents the predicted value of regional carbon emissions, represents the prediction weight, represents the historical data of regional carbon emissions after integration, represents the prediction bias;

[0015] The carbon emission prediction module is used to input the transmitted real-time data of regional carbon emissions into the carbon emission prediction model, generate a prediction result, compare the actual value of regional carbon emissions with the prediction result, evaluate the carbon emission prediction model, and generate an evaluation result. The evaluation of the carbon emission prediction model includes subtracting the actual value of regional carbon emissions from the predicted value of regional carbon emissions and summing them up;

[0016] The prediction optimization module is used to optimize the carbon emission prediction model according to the evaluation results;

[0017] The visual interaction module is used to visualize and display the real-time data and historical data, prediction results and evaluation results of regional carbon emissions in real time.

[0018] As a preferred solution of the regional carbon emission prediction system based on the deep learning model described in the present invention, the real-time data and historical data of the regional carbon emission are obtained, and the historical data of the regional carbon emission is dynamically integrated according to the quality of the historical data of the regional carbon emission;

[0019] The judgment rules for the quality of the historical data of regional carbon emissions include:

[0020] The quality of the historical data of regional carbon emissions is comprehensively judged based on its accuracy, completeness, consistency and timeliness. The calculation formula for comprehensively judging the quality of the historical data of regional carbon emissions is as follows:

[0021] ;

[0022] In the formula, Indicates the quality value of the historical data of regional carbon emissions, Indicates the accuracy value of historical data of regional carbon emissions, Indicates the integrity value of the historical data of regional carbon emissions, Represents the consistency value of historical data of regional carbon emissions, Indicates the timeliness of historical data of regional carbon emissions.

[0023] As a preferred solution of the regional carbon emission prediction system based on the deep learning model described in the present invention, the historical data of the regional carbon emissions are dynamically integrated according to the quality of the historical data of the regional carbon emissions, and the calculation formula for dynamically integrating the historical data of regional carbon emissions is as follows:

[0024] ;

[0025] In the formula, represents the historical data of regional carbon emissions after integration, Indicates The weight of each monitoring point, Indicates Historical data of regional carbon emissions at monitoring points, Indicates the total number of monitoring points;

[0026] Among them, the weight of the monitoring point changes dynamically according to the quality of the historical data of regional carbon emissions. The weight change formula of the monitoring point is as follows:

[0027] ;

[0028] In the formula, Represents the weight coefficient set according to the quality of historical data on regional carbon emissions, Indicates the quality value of the historical data of regional carbon emissions, The historical data of regional carbon emissions A quality value.

[0029] As a preferred solution of the regional carbon emission prediction system based on the deep learning model described in the present invention, the carbon emission prediction model is constructed and trained by the fused historical data of regional carbon emissions to obtain the loss value of the carbon emission prediction model. The function expression of the loss value of the carbon emission prediction model is as follows:

[0030] ;

[0031] In the formula, represents the loss value of the carbon emission prediction model, represents the quantile loss function, represents the actual value of regional carbon emissions, represents the predicted value of regional carbon emissions, Represents the quantile level.

[0032] As a preferred solution of the regional carbon emission prediction system based on the deep learning model described in the present invention, the actual value of the regional carbon emission is compared with the prediction result, the carbon emission prediction model is evaluated, and the evaluation result is generated. The specific steps of generating the evaluation result include:

[0033] The prediction result generated by inputting the real-time data of regional carbon emissions into the carbon emission prediction model is the predicted value of regional carbon emissions. The actual value of regional carbon emissions is subtracted from the predicted value of regional carbon emissions and the sum is processed to obtain the predicted difference. The difference threshold is configured and the predicted difference is compared with the difference threshold. If the predicted difference is less than or equal to the difference threshold, it means that the prediction result of the carbon emission prediction model is correct.

[0034] If the prediction difference is greater than the difference threshold, it means that the prediction result of the carbon emission prediction model is incorrect;

[0035] When the prediction result of the carbon emission prediction model is incorrect, the monitoring point showing the incorrect prediction result of the carbon emission prediction model is an error point, the number of error points is counted, the error number threshold is configured, and the number of error points is compared with the error number threshold. If the number of error points is less than the error number threshold, the evaluation result is that the carbon emission prediction model is not optimized;

[0036] If the number of error points is greater than or equal to the error number threshold, the evaluation result is an optimized carbon emission prediction model.

[0037] As a preferred solution of the regional carbon emission prediction system based on the deep learning model described in the present invention, when the evaluation result is to optimize the carbon emission prediction model, the difference threshold is adjusted. If the evaluation result after adjusting the difference threshold is displayed as not optimizing the carbon emission prediction model, it means that the configuration of the difference threshold is incorrect;

[0038] If the evaluation result still shows that the carbon emission prediction model is optimized after adjusting the difference threshold, the prediction weight of the carbon emission prediction model is adjusted. The function expression for adjusting the prediction weight of the carbon emission prediction model is as follows:

[0039] ;

[0040] In the formula, represents the adjusted prediction weight of the carbon emission prediction model, represents the prediction weight, represents the learning rate of the carbon emission prediction model, represents the gradient of the prediction weight of the carbon emission prediction model, Represents the loss value of the carbon emission prediction model.

[0041] As a preferred solution of the regional carbon emission prediction system based on deep learning model described in the present invention, the function expression for adjusting the learning rate of the carbon emission prediction model is as follows:

[0042] ;

[0043] In the formula, represents the initial learning rate of the carbon emission prediction model, represents the decay rate of the carbon emission prediction model, represents the number of training times of the carbon emission prediction model, Represents the attenuation times of the carbon emission prediction model.

[0044] A regional carbon emission prediction method based on a deep learning model, comprising: S1, acquiring real-time data and historical data of regional carbon emissions, fusing the historical data of regional carbon emissions and adapting to the transmission of real-time data of regional carbon emissions, fusing the historical data of regional carbon emissions including dynamically fusing the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions;

[0045] S2. Build and train a carbon emission prediction model through the integrated historical data of regional carbon emissions to obtain the loss value of the carbon emission prediction model;

[0046] S3, inputting the transmitted real-time data of regional carbon emissions into the carbon emission prediction model, generating a prediction result, and comparing the actual value of regional carbon emissions with the prediction result, evaluating the carbon emission prediction model, and generating an evaluation result, wherein the evaluation of the carbon emission prediction model includes making a difference comparison between the actual value of regional carbon emissions and the predicted value of regional carbon emissions;

[0047] S4. Optimize the carbon emission prediction model based on the evaluation results;

[0048] S5. Real-time visualization of real-time and historical data, forecast results, and evaluation results of regional carbon emissions.

[0049] A computer device comprises: a memory for storing instructions; and a processor for executing the instructions, so that the computer device implements a regional carbon emission prediction method based on a deep learning model.

[0050] A computer-readable storage medium stores a computer program, which, when executed, implements a regional carbon emission prediction method based on a deep learning model.

[0051] Beneficial effects of the present invention: The present invention acquires the real-time data and historical data of regional carbon emissions through the data acquisition module, integrates the historical data of regional carbon emissions and adapts to the transmission of real-time data of regional carbon emissions, and integrates the historical data of regional carbon emissions, including dynamically integrating the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions, ensuring that the information used in training and building the carbon emission prediction model is comprehensive and multi-dimensional, which not only improves the richness of the data, but also enhances the ability of the carbon emission prediction model to recognize complex patterns; the model construction module constructs and trains the carbon emission prediction model through the integrated historical data of regional carbon emissions, and obtains the loss value of the carbon emission prediction model, so as to effectively predict the regional carbon emissions; the carbon emission prediction module transmits the real-time data of regional carbon emissions The data is input into the carbon emission prediction model to generate prediction results, and the actual value of regional carbon emissions is compared with the prediction results to evaluate the carbon emission prediction model and generate evaluation results. The evaluation of the carbon emission prediction model includes making a difference comparison between the actual value of regional carbon emissions and the predicted value of regional carbon emissions, so as to effectively evaluate the carbon emission prediction model; the prediction optimization module optimizes the carbon emission prediction model according to the evaluation results, and continuously adjusts and optimizes the model parameters according to the evaluation results to form a closed-loop feedback system to ensure that the carbon emission prediction model evolves over time and improves long-term accuracy and stability; the visual interaction module can visualize the real-time and historical data of regional carbon emissions, prediction results and evaluation results in real time, making complex data and the prediction results of the carbon emission prediction model easy to understand. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0053] Figure 1 This is a system structure diagram of a regional carbon emission prediction system based on a deep learning model according to the present invention;

[0054] Figure 2 A diagram of steps for generating evaluation results of a regional carbon emission prediction system based on a deep learning model according to the present invention;

[0055] Figure 3 This is a flow chart of an optimization model strategy for a regional carbon emission prediction system based on a deep learning model according to the present invention;

[0056] Figure 4 This is a method flow chart of a regional carbon emissions prediction method based on a deep learning model of the present invention. DETAILED DESCRIPTION

[0057] In order to make the above-mentioned objects, 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 accompanying drawings.

[0058] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0059] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0060] Example 1

[0061] In this embodiment, a system structure diagram of a regional carbon emission prediction system based on a deep learning model is provided, such as Figure 1 As shown, a regional carbon emission prediction system based on a deep learning model includes a data acquisition module, a model building module, a carbon emission prediction module, a prediction optimization module and a visual interaction module.

[0062] The data acquisition module is used to acquire real-time data and historical data of regional carbon emissions, integrate the historical data of regional carbon emissions and adapt to the transmission of real-time data of regional carbon emissions. The integration of historical data of regional carbon emissions includes dynamically integrating the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions.

[0063] Acquire real-time and historical data of regional carbon emissions, and dynamically integrate the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions to improve the accuracy and efficiency of data fusion.

[0064] The rules for judging the quality of historical data on regional carbon emissions include:

[0065] The quality of the historical data of regional carbon emissions is comprehensively judged based on its accuracy, completeness, consistency and timeliness. The calculation formula for comprehensively judging the quality of the historical data of regional carbon emissions is as follows:

[0066] ;

[0067] In the formula, Indicates the quality value of the historical data of regional carbon emissions, Indicates the accuracy value of historical data of regional carbon emissions, Indicates the integrity value of the historical data of regional carbon emissions, Represents the consistency value of historical data of regional carbon emissions, Indicates the timeliness of historical data of regional carbon emissions.

[0068] Real-time data and historical data of regional carbon emissions are obtained from different monitoring points, and the accuracy of historical data is obtained by comparing with known correct data. The absolute error between historical data and known correct data is calculated to obtain the overall root mean square error, and the accuracy value of historical data is obtained. The completeness of historical data is obtained by calculating the proportion of missing values, and the consistency of historical data is obtained by comparing the differences of similar data points of historical data from different sources. The variance of similar data points of historical data from different sources is calculated to obtain the differences of similar data points of historical data from different sources, thereby obtaining the consistency value of historical data. The timeliness of historical data is determined by comparing the update time of historical data with the time interval of historical data.

[0069] Configure the quality threshold and compare the quality value of the historical data with the quality threshold. If the quality value of the historical data is less than the quality threshold, it means that the historical data is of low quality.

[0070] If the quality value of the historical data is greater than or equal to the quality threshold, it means that the historical data is of high quality.

[0071] The historical data of regional carbon emissions are dynamically integrated according to the quality of the historical data of regional carbon emissions. The calculation formula for dynamically integrating the historical data of regional carbon emissions is as follows:

[0072] ;

[0073] In the formula, represents the historical data of regional carbon emissions after integration, Indicates The weight of each monitoring point, Indicates Historical data of regional carbon emissions at monitoring points, Indicates the total number of monitoring points.

[0074] Among them, the weight of the monitoring point changes dynamically according to the quality of the historical data of regional carbon emissions. The weight change formula of the monitoring point is as follows:

[0075] ;

[0076] In the formula, Indicates the weight coefficient set according to the quality of the historical data of regional carbon emissions. For example, if the historical data is of low quality , when historical data is high quality , Indicates the quality value of the historical data of regional carbon emissions, The historical data of regional carbon emissions A quality value.

[0077] At the same time, when real-time data is obtained, the transmission time of the real-time data is adapted to the transmission of the real-time data of regional carbon emissions. The formula for the transmission time of the real-time data is as follows:

[0078] ;

[0079] In the formula, Indicates the transmission time of real-time data of regional carbon emissions, The compressed data size of the real-time data of regional carbon emissions. Indicates the compression rate of real-time data of regional carbon emissions, The rate at which real-time data representing regional carbon emissions is transmitted.

[0080] What needs to be explained is: the compression rate of real-time data on regional carbon emissions It refers to the ratio of the compressed data size of the real-time data of regional carbon emissions to the original data size. The transmission rate of the real-time data of regional carbon emissions It is determined based on the actual network bandwidth and network conditions.

[0081] The transmission time of the real-time data of regional carbon emissions is determined by the above formula to ensure that the real-time data of regional carbon emissions meets the requirements of fast transmission and efficient delivery.

[0082] The model building module is used to build and train a carbon emission prediction model through the integrated historical data of regional carbon emissions to obtain the loss value of the carbon emission prediction model.

[0083] The carbon emission prediction model is constructed and trained through the fused historical data of regional carbon emissions. The function expression of the constructed carbon emission prediction model is as follows:

[0084] ;

[0085] In the formula, represents the predicted value of regional carbon emissions, represents the prediction weight, Represents the prediction bias.

[0086] The fused historical data of regional carbon emissions are mapped to the predicted values ​​of regional carbon emissions through the multi-layer perceptron in deep learning technology, where the prediction weight and prediction bias This needs to be determined through continuous training and iteration.

[0087] The carbon emission prediction model is continuously trained to obtain the loss value of the carbon emission prediction model. The function expression of the loss value of the carbon emission prediction model is as follows:

[0088] ;

[0089] In the formula, represents the loss value of the carbon emission prediction model, represents the quantile loss function, represents the actual value of regional carbon emissions, Represents the quantile level.

[0090] The quantile level The determination is based on whether the carbon emission prediction model needs to be highly sensitive to the predicted losses and the value is taken according to the actual situation.

[0091] The function expression of the quantile loss function is as follows:

[0092] ;

[0093] The loss value of the above carbon emission prediction model can reflect the gap between the predicted value of regional carbon emissions and the actual value of regional carbon emissions, so as to serve as a data reference when subsequently determining whether the carbon emission prediction model needs to be optimized.

[0094] The carbon emission prediction module is used to input the real-time data of regional carbon emissions into the carbon emission prediction model, generate prediction results, compare the actual value of regional carbon emissions with the prediction results, evaluate the carbon emission prediction model, and generate evaluation results. The evaluation of the carbon emission prediction model includes subtracting the actual value of regional carbon emissions from the predicted value of regional carbon emissions and summing them up.

[0095] Compare the actual value of regional carbon emissions with the predicted results, evaluate the carbon emission prediction model, and generate the evaluation results. The steps for generating the evaluation results are as follows: Figure 2 As shown, specifically including:

[0096] The prediction result generated by inputting the real-time data of regional carbon emissions into the carbon emission prediction model is the predicted value of regional carbon emissions. The actual value of regional carbon emissions is subtracted from the predicted value of regional carbon emissions and the sum is processed, that is, the loss value obtained by the above-mentioned training carbon emission prediction model, to obtain the predicted difference, configure the difference threshold, and compare the predicted difference with the difference threshold. If the predicted difference is less than or equal to the difference threshold, it means that the prediction result of the carbon emission prediction model is correct.

[0097] If the prediction difference is greater than the difference threshold, it means that the prediction result of the carbon emission prediction model is incorrect;

[0098] When the prediction result of the carbon emission prediction model is incorrect, the monitoring point showing the incorrect prediction result of the carbon emission prediction model is an error point, the number of error points is counted, the error number threshold is configured, and the number of error points is compared with the error number threshold. If the number of error points is less than the error number threshold, the evaluation result is that the carbon emission prediction model is not optimized;

[0099] If the number of error points is greater than or equal to the error number threshold, the evaluation result is an optimized carbon emission prediction model.

[0100] The monitoring points with errors are determined by comparing the difference threshold with the predicted difference. Then, the number of monitoring points with errors is counted to determine whether the carbon emission prediction model needs to be optimized. If optimization is required, the prediction optimization module is entered.

[0101] The prediction optimization module is used to optimize the carbon emission prediction model based on the evaluation results.

[0102] like Figure 3 As shown, when the evaluation result is to optimize the carbon emission prediction model, adjust the difference threshold. If the evaluation result shows that the carbon emission prediction model is not optimized after adjusting the difference threshold, it means that the configuration of the difference threshold is incorrect.

[0103] If the evaluation result still shows that the carbon emission prediction model is optimized after adjusting the difference threshold, the prediction weight of the carbon emission prediction model is adjusted. The function expression for adjusting the prediction weight of the carbon emission prediction model is as follows:

[0104] ;

[0105] In the formula, represents the adjusted prediction weight of the carbon emission prediction model, represents the learning rate of the carbon emission prediction model, represents the gradient of the prediction weight of the carbon emission prediction model, Represents the loss value of the carbon emission prediction model.

[0106] The function expression for adjusting the learning rate of the carbon emission prediction model is as follows:

[0107] ;

[0108] In the formula, represents the initial learning rate of the carbon emission prediction model, represents the decay rate of the carbon emission prediction model, represents the number of training times of the carbon emission prediction model, Represents the attenuation times of the carbon emission prediction model.

[0109] What needs to be explained is: the decay rate of the carbon emission prediction model A series of possible decay rates are pre-set, such as 0.9, 0.95 and 0.99, and these decay rates are used to train the carbon emission prediction model. At the same time, the loss value of the carbon emission prediction model under each decay rate is recorded to select the decay rate that makes the carbon emission prediction model perform best.

[0110] In the above, the difference threshold is first adjusted to determine whether the error point is misjudged. However, the adjustment of the difference threshold is not without limit, but is within an adjustment range. If the carbon emission prediction model still needs to be optimized after adjusting the difference threshold, then the training effect and prediction accuracy of the carbon emission prediction model can be improved by adjusting the learning rate and prediction weight of the carbon emission prediction model.

[0111] The visual interaction module is used to visualize the real-time and historical data, prediction results and evaluation results of regional carbon emissions.

[0112] The real-time and historical data of carbon emissions in each region are comprehensively displayed by combining time series graphs and geographic heat maps. The prediction results obtained by the carbon emission prediction model are also displayed, and the evaluation results of the carbon emission prediction model are displayed. The prediction difference is compared with the difference threshold to determine whether the monitoring point is an error point. When it is displayed as an error point, it is marked with different colors in the visualization, and areas with relatively high carbon emissions are highlighted.

[0113] Example 2

[0114] In this embodiment, a method flow chart of a method for predicting regional carbon emissions based on a deep learning model is provided, such as Figure 4 As shown, a regional carbon emission prediction method based on a deep learning model includes:

[0115] S1. Acquire real-time data and historical data of regional carbon emissions, integrate the historical data of regional carbon emissions and adapt to the transmission of real-time data of regional carbon emissions. The integration of historical data of regional carbon emissions includes dynamically integrating the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions.

[0116] Real-time data and historical data of regional carbon emissions are obtained from different monitoring points, the accuracy of historical data is obtained by comparing with known correct data, the completeness of historical data is obtained by calculating the proportion of missing values, the consistency of historical data is obtained by comparing the differences of similar data points of historical data from different sources, the timeliness of historical data is obtained by comparing the update time of historical data with the time interval of historical data, and the quality of regional carbon emissions historical data is comprehensively judged according to the accuracy, completeness, consistency and timeliness of regional carbon emissions historical data, and the historical data of regional carbon emissions is dynamically integrated according to the quality of regional carbon emissions historical data.

[0117] S2. Build and train a carbon emission prediction model through the integrated historical data of regional carbon emissions to obtain the loss value of the carbon emission prediction model.

[0118] The fused historical data of regional carbon emissions are mapped to the predicted value of regional carbon emissions through the multi-layer perceptron in deep learning technology. The carbon emission prediction model is continuously trained to obtain the loss value of the carbon emission prediction model. The loss value of the carbon emission prediction model can reflect the gap between the predicted value of regional carbon emissions and the actual value of regional carbon emissions, so as to provide data reference for subsequent judgment on whether the carbon emission prediction model needs to be optimized.

[0119] S3. Input the transmitted real-time data of regional carbon emissions into the carbon emission prediction model to generate prediction results, compare the actual value of regional carbon emissions with the prediction results, evaluate the carbon emission prediction model, and generate evaluation results. The evaluation of the carbon emission prediction model includes making a difference comparison between the actual value of regional carbon emissions and the predicted value of regional carbon emissions.

[0120] The monitoring points with errors are determined by comparing the difference threshold with the predicted difference, and then the number of monitoring points with errors is counted to determine whether the carbon emission prediction model needs to be optimized. If optimization is required, enter S4.

[0121] S4. Optimize the carbon emission prediction model based on the evaluation results.

[0122] First, the difference threshold is adjusted to determine whether the error point is misjudged. However, the adjustment of the difference threshold is not unlimited, but is within an adjustment range. If the carbon emission prediction model still needs to be optimized after adjusting the difference threshold, then the training effect and prediction accuracy of the carbon emission prediction model can be improved by adjusting the learning rate and prediction weight of the carbon emission prediction model.

[0123] S5. Real-time visualization of real-time and historical data, forecast results, and evaluation results of regional carbon emissions.

[0124] The real-time and historical data of carbon emissions in each region are comprehensively displayed by combining time series graphs and geographic heat maps. The prediction results obtained by the carbon emission prediction model are also displayed, and the evaluation results of the carbon emission prediction model are displayed. The prediction difference is compared with the difference threshold to determine whether the monitoring point is an error point. When it is displayed as an error point, it is marked with different colors in the visualization, and areas with relatively high carbon emissions are highlighted.

[0125] For the specific content of a regional carbon emission prediction method based on a deep learning model, please refer to a regional carbon emission prediction system based on a deep learning model, which will not be repeated here.

[0126] Example 3

[0127] In this embodiment, a computer device is provided, including a memory and a processor, the memory is used to store instructions, and the processor is used to execute the instructions, so that the computer device executes the steps of implementing the above-mentioned regional carbon emission prediction method based on a deep learning model.

[0128] Example 4

[0129] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed, the steps of the above-mentioned regional carbon emission prediction method based on the deep learning model are implemented.

[0130] The computer-readable storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and other media for storing program codes.

[0131] 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 can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A regional carbon emission prediction system based on a deep learning model, characterized by: include: Data acquisition module, model building module, carbon emission prediction module, prediction optimization module and visual interaction module; The data acquisition module is used to acquire real-time data and historical data of regional carbon emissions, integrate the historical data of regional carbon emissions and adapt to the transmission of real-time data of regional carbon emissions, and integrate the historical data of regional carbon emissions including dynamically integrating the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions; Acquire real-time data and historical data of carbon emissions in the region, and dynamically integrate the historical data of carbon emissions in the region according to the quality of the historical data of carbon emissions in the region; The judgment rules for the quality of the historical data of regional carbon emissions include: The quality of the historical data of regional carbon emissions is comprehensively judged based on its accuracy, completeness, consistency and timeliness. The calculation formula for comprehensively judging the quality of the historical data of regional carbon emissions is as follows: ； In the formula, Indicates the quality value of the historical data of regional carbon emissions, Indicates the accuracy value of historical data of regional carbon emissions, Indicates the integrity value of the historical data of regional carbon emissions, Represents the consistency value of historical data of regional carbon emissions, A timeliness value representing the historical data of regional carbon emissions; Real-time data and historical data of regional carbon emissions are obtained from different monitoring points. The accuracy of historical data is obtained by comparing with known correct data. The absolute error between historical data and known correct data is calculated to obtain the overall root mean square error, and the accuracy value of historical data is obtained. The integrity of historical data is obtained by calculating the proportion of missing values. The consistency of historical data is obtained by comparing the differences of similar data points of historical data from different sources. The variance of similar data points of historical data from different sources is calculated to obtain the differences of similar data points of historical data from different sources, thereby obtaining the consistency value of historical data. The timeliness of historical data is obtained by comparing the update time of historical data with the time interval of historical data. The historical data of carbon emissions in the region are dynamically integrated according to the quality of the historical data of carbon emissions in the region. The calculation formula for dynamically integrating the historical data of carbon emissions in the region is as follows: ； In the formula, represents the historical data of regional carbon emissions after integration, Indicates The weight of each monitoring point, Indicates Historical data of regional carbon emissions at monitoring points, Indicates the total number of monitoring points; Among them, the weight of the monitoring point changes dynamically according to the quality of the historical data of regional carbon emissions. The weight change formula of the monitoring point is as follows: ； In the formula, Represents the weight coefficient set according to the quality of historical data on regional carbon emissions, Indicates the quality value of the historical data of regional carbon emissions, The historical data of regional carbon emissions Quality value; The model building module is used to build and train a carbon emission prediction model through the fused historical data of regional carbon emissions to obtain a loss value of the carbon emission prediction model; The functional expression of the constructed carbon emission prediction model is as follows: ； In the formula, represents the predicted value of regional carbon emissions, represents the prediction weight, represents the historical data of regional carbon emissions after integration, represents the prediction bias; The carbon emission prediction module is used to input the transmitted real-time data of regional carbon emissions into the carbon emission prediction model, generate a prediction result, compare the actual value of regional carbon emissions with the prediction result, evaluate the carbon emission prediction model, and generate an evaluation result. The evaluation of the carbon emission prediction model includes subtracting the actual value of regional carbon emissions from the predicted value of regional carbon emissions and summing them up; The actual value of the regional carbon emissions is compared with the prediction result, the carbon emissions prediction model is evaluated, and an evaluation result is generated. The specific steps of generating the evaluation result include: The prediction result generated by inputting the real-time data of regional carbon emissions into the carbon emission prediction model is the predicted value of regional carbon emissions. The actual value of regional carbon emissions is subtracted from the predicted value of regional carbon emissions and the sum is processed to obtain the predicted difference. The difference threshold is configured and the predicted difference is compared with the difference threshold. If the predicted difference is less than or equal to the difference threshold, it means that the prediction result of the carbon emission prediction model is correct. If the prediction difference is greater than the difference threshold, it means that the prediction result of the carbon emission prediction model is incorrect; When the prediction result of the carbon emission prediction model is incorrect, the monitoring point showing the incorrect prediction result of the carbon emission prediction model is an error point, the number of error points is counted, the error number threshold is configured, and the number of error points is compared with the error number threshold. If the number of error points is less than the error number threshold, the evaluation result is that the carbon emission prediction model is not optimized; If the number of error points is greater than or equal to the error number threshold, the evaluation result is an optimized carbon emission prediction model; The prediction optimization module is used to optimize the carbon emission prediction model according to the evaluation results; When the evaluation result is to optimize the carbon emission prediction model, the difference threshold is adjusted. If the evaluation result after adjusting the difference threshold is displayed as not optimizing the carbon emission prediction model, it means that the configuration of the difference threshold is incorrect; If the evaluation result still shows that the carbon emission prediction model is optimized after adjusting the difference threshold, the prediction weight of the carbon emission prediction model is adjusted. The function expression for adjusting the prediction weight of the carbon emission prediction model is as follows: ； In the formula, represents the adjusted prediction weight of the carbon emission prediction model, represents the prediction weight, represents the learning rate of the carbon emission prediction model, represents the gradient of the prediction weight of the carbon emission prediction model, Represents the loss value of the carbon emission prediction model; The visual interaction module is used to visualize and display the real-time data and historical data, prediction results and evaluation results of regional carbon emissions in real time.

2. A regional carbon emission prediction system based on a deep learning model as claimed in claim 1, characterized in that: The carbon emission prediction model is constructed and trained through the fused historical data of regional carbon emissions to obtain the loss value of the carbon emission prediction model. The function expression of the loss value of the carbon emission prediction model is as follows: ； In the formula, represents the loss value of the carbon emission prediction model, represents the quantile loss function, represents the actual value of regional carbon emissions, represents the predicted value of regional carbon emissions, Represents the quantile level.

3. A regional carbon emission prediction system based on a deep learning model as claimed in claim 2, characterized in that: The function expression for adjusting the learning rate of the carbon emission prediction model is as follows: ； In the formula, represents the initial learning rate of the carbon emission prediction model, represents the decay rate of the carbon emission prediction model, represents the number of training times of the carbon emission prediction model, Represents the attenuation times of the carbon emission prediction model.

4. A regional carbon emission prediction method based on a deep learning model, implemented based on a regional carbon emission prediction system based on a deep learning model according to any one of claims 1 to 3, characterized in that: include: S1. Acquire real-time data and historical data of regional carbon emissions, integrate the historical data of regional carbon emissions and adapt to the transmission of real-time data of regional carbon emissions, and integrate the historical data of regional carbon emissions including dynamically integrating the historical data of regional carbon emissions according to the quality of the historical data of regional carbon emissions; S2. Build and train a carbon emission prediction model through the integrated historical data of regional carbon emissions to obtain the loss value of the carbon emission prediction model; S3, inputting the transmitted real-time data of regional carbon emissions into the carbon emission prediction model, generating a prediction result, and comparing the actual value of regional carbon emissions with the prediction result, evaluating the carbon emission prediction model, and generating an evaluation result, wherein the evaluation of the carbon emission prediction model includes making a difference comparison between the actual value of regional carbon emissions and the predicted value of regional carbon emissions; S4. Optimize the carbon emission prediction model based on the evaluation results; S5. Real-time visualization of real-time and historical data, forecast results, and evaluation results of regional carbon emissions.

5. A computer device, characterized in that: include: A memory for storing instructions; A processor is used to execute the instructions so that the computer device implements the regional carbon emission prediction method based on a deep learning model as described in claim 4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, it implements the regional carbon emission prediction method based on deep learning model as described in claim 4.

Images