Interval prediction method for carbon dioxide emission in transportation industry

Abstract

The invention discloses an interval prediction method for carbon dioxide emission in the transportation industry, and belongs to the field of grey prediction, and the method comprises the following steps: S1, collecting original data, forming a data set, and carrying out the preprocessing; s2, constructing a grey multivariable NSGM (1, N) model; s3, predicting each variable of the system by using a single-variable prediction model; s4, constructing the boundary of each variable of the system; s5, outputting an interval prediction result of the carbon dioxide emission of the Chinese traffic transportation industry; the method provided by the invention has relatively high adaptability and relatively high prediction accuracy.

Description

Technical Field

[0001] The invention relates to the field of grey prediction models, and in particular to a method for interval prediction of carbon dioxide emissions in the transportation industry. Background Art

[0002] The prediction of carbon dioxide emissions in the transportation industry is an important part of carbon pollution prevention and control. By predicting carbon dioxide emissions, the development and construction activities of the transportation industry can be based on environmental coordination and sustainable development, which plays an important role in carbon emission management. Carbon dioxide emission prediction can help understand the changing trend of carbon emissions, especially in the event of sudden environmental events. Accurate carbon dioxide emission prediction can enable effective response measures to be taken in a timely manner. By predicting carbon dioxide emissions, we can effectively prevent environmental deterioration, promote the improvement of air quality, improve people's health level, and ensure social stability.

[0003] In order to predict the carbon dioxide emissions in the transportation industry, managers and researchers have established many prediction models, such as regression analysis, gray model, machine learning and artificial neural network. In these prediction modeling methods, when using a single variable model to predict the independent variable, the predicted data has errors, forming an interval that fluctuates around the precise data. In a complex and uncertain system, many data have dynamic volatility. Representing these data as interval numbers can better reflect their growth trend and range of variation. Therefore, in the case of less available data and limited data sample size, an interval prediction method is proposed to solve the interval prediction problem of carbon dioxide emissions in the transportation industry. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, this patent application proposes a method for interval prediction of carbon dioxide emissions in the transportation industry.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for interval prediction of carbon dioxide emissions in the transportation industry comprises the following steps:

[0007] S1: Collect raw data, form a data set, and perform preprocessing;

[0008] S2: construct grey multivariate NSGM (1, N) model;

[0009] S3: Use the univariate prediction model to predict the system's individual variables;

[0010] S4: construct the boundaries of the system’s respective variables;

[0011] S5: Output of interval forecast results of carbon dioxide emissions from China’s transportation industry.

[0012] Furthermore, the step of collecting raw data in step S1 includes: collecting raw data related to carbon dioxide and its emissions in the transportation industry, that is, dependent variable and independent variable data, dividing each variable sequence data into two groups, the first group is used as a training set to build an NSGM (1, N) model, and the second group is used as a test set to verify the prediction error of the model, and further preprocessing the collected raw data, the specific steps include:

[0013] A1: Calculate each variable The first-order cumulative generation sequence of

[0014]

[0015] in,

[0016] A2: Calculation The background value coefficient is the immediate generation sequence of the undetermined parameter λ

[0017]

[0018] in

[0019] Furthermore, the specific steps of step S2 to construct the grey multivariate NSGM (1, N) model are:

[0020] B1: construct parameter matrices B and Y;

[0021]

[0022] B2: Calculate the NSGM(1,N) model parameter list

[0023] (I) When m = N + 3 and |B| ≠ 0,

[0024] (II) When m>N+3 and |B T When B|≠0,

[0025] (III) When m<N+3 and |B T When B|≠0,

[0026] B3: The time response formula of NSGM(1,N) model is obtained by mathematical induction and iteration method;

[0027]

[0028] B4: Derive the cumulative generation formula of the NSGM(1,N) model.

[0029]

[0030] in,

[0031] B5: According to the cumulative reduction formula of the NSGM (1, N) model, calculate the simulation value / prediction value and the relative simulation / prediction comprehensive percentage error of the training set and the test set respectively; and determine whether the comprehensive error meets the system requirements.

[0032] Furthermore, in step S3, a corresponding univariate grey prediction model is used to perform simulation prediction according to the data characteristics of the independent variable, and a performance analysis is performed based on the prediction results obtained by each model to determine whether the model can be used in this experiment.

[0033] Furthermore, the CMRPE is used to determine whether each univariate prediction model meets the grey prediction standard and whether it can be applied in the experiment.

[0034] Furthermore, the interval of the independent variable is constructed, and the specific steps are as follows:

[0035] B1: Construct positive / negative independent variables

[0036] If b i >0, then and Positive correlation, defined as a positive independent variable If b i <0, then and Negative correlation, defined as a negative independent variable

[0037] B2: Constructing positive / negative independent variable sequences

[0038] when Then the positive / negative independent variable of the bounded interval is defined as a positive / negative sequence, is a forward sequence, is a negative sequence,

[0039] B3: Constructing the matrix and

[0040] Since the positive independent variable sequence is positively correlated with the dependent variable, the upper bound f of the NSGM (1, N) model is calculated. s (k), take Sequence upper bound Negative independent variable sequence Negatively correlated with the dependent variable, so when calculating the upper bound f of the NSGM(1,N) model s (k), take Sequence lower bound and The sequence constitutes a new sequence Matrix, similarly, and The sequence constitutes a new sequence matrix,

[0041] B4: Application Matrix The data inside predicts the future value of the system dependent variable of the NSGM (1, N) model to obtain the upper boundary, and the matrix is ​​applied The data inside predicts the future value of the system dependent variable using the NSGM(1,N) model to obtain the lower bound.

[0042] Furthermore, in step S5, the prediction result is output, and the interval prediction value is obtained according to the upper and lower boundary values ​​obtained in step S4. The specific steps of the interval expansion prediction output based on the grey multidimensional model are: according to the training input-output mode, the data is input into the new model, and the model output is the result of the interval prediction of China's carbon dioxide emissions in the transportation industry.

[0043] Beneficial Effects

[0044] (1) Through the construction of independent variable boundaries and the expansion of dependent variable intervals, the multivariate grey prediction model method system is enriched and improved;

[0045] (2) The multidimensional grey interval prediction model was used to predict carbon emissions in my country's transportation industry. Compared with an exact numerical value, the interval number is more flexible in policy reference. The research results can provide a theoretical basis for relevant government departments to formulate relevant policies;

[0046] (3) It provides an effective way to solve other small sample interval prediction problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 The present invention provides a method for predicting carbon dioxide emissions in the transportation industry;

[0048] Figure 2 This is an interval construction flow chart of an interval prediction method for carbon dioxide emissions in the transportation industry described in the present invention. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0050] Reference Figure 1-Figure 2 , a method for interval prediction of carbon dioxide emissions in the transportation industry, comprising the following steps:

[0051] S1: Collect raw data, form a data set, and perform preprocessing;

[0052] S2: construct grey multivariate NSGM (1, N) model;

[0053] S3: construct the boundaries of the system’s respective variables;

[0054] S4: Output of interval forecast results of carbon dioxide emissions from China’s transportation industry.

[0055] Optimized, in step S1, the original data, i.e., the dependent variable, is collected, the independent variable data is screened according to the grey correlation degree, and then the determined data is preprocessed; in step S3, according to the data characteristics of each variable, a suitable single variable prediction model is selected to predict the independent variable, and the boundary is constructed to expand the prediction interval; in step S4, the boundary of the independent variable is constructed, and the NSGM (1, N) model is used to perform interval prediction on the future value of the system dependent variable.

[0056] Optimized, the step of collecting historical data in step S1 also includes: step S1 collects original data, i.e., dependent variable and independent variable data, divides each variable sequence data into two groups, the first group is used as a training set to build an NSGM (1, N) model, and the second group is used as a test set to test the prediction error of the model and perform data preprocessing, and the specific steps include:

[0057] A1: Calculate each variable The first-order cumulative generation sequence of

[0058]

[0059] in,

[0060] A2: Calculation The background value coefficient is the immediate generation sequence of the undetermined parameter λ

[0061]

[0062] in

[0063] Data preprocessing is performed on the original data of different dimensions, and the initialization method is selected to reduce the impact on the system.

[0064] Optimized, step S2 constructs a multivariate NSGM (1, N) model and tests the model error. The specific steps are:

[0065] B1: construct parameter matrices B and Y;

[0066]

[0067] B2: Calculate the NSGM(1,N) model parameter list

[0068] (I) When m = N + 3 and |B| ≠ 0,

[0069] (II) When m>N+3 and |B T When B|≠0,

[0070] (III) When m<N+3 and |B T When B|≠0,

[0071] B3: The time response formula of NSGM(1,N) model is obtained by mathematical induction and iteration method;

[0072]

[0073] B4: Derive the cumulative generation formula of the NSGM(1,N) model.

[0074]

[0075] in,

[0076] B5: According to the cumulative reduction formula of the NSGM (1, N) model, calculate the simulated value / prediction value and the relative simulation / prediction comprehensive percentage error of the training set and the test set respectively; determine whether the comprehensive error meets the system requirements.

[0077] In step S3, the gray single variable prediction model is used to simulate and predict each independent variable, and the performance analysis is performed based on the prediction results obtained by each model.

[0078] Specifically: According to the simulation data, the mean relative simulation percentage error MRSPE is obtained.

[0079] According to the prediction data, the mean relative prediction percentage error MRPPE is obtained.

[0080] The comprehensive mean relative error CMRPE is obtained based on the simulation error and prediction error.

[0081] According to CMRPE, it is judged whether each single variable prediction model meets the grey prediction standard, and further whether it can be applied in the experiment.

[0082] The specific steps of optimizing the variable boundary in step S3 are:

[0083] B1: Construct positive / negative independent variables

[0084] If b i >0, then and Positive correlation, defined as a positive independent variable If b i <0, then and Negative correlation, defined as a negative independent variable

[0085] B2: Constructing positive / negative independent variable sequences

[0086] when Then the positive / negative independent variable of the bounded interval is defined as a positive / negative sequence, is a forward sequence, Negative sequence.

[0087] B3: Constructing the matrix

[0088] Since the positive independent variable sequence is positively correlated with the dependent variable, the upper bound f of the NSGM (1, N) model is calculated. s (k), take Sequence upper bound Negative independent variable sequence Negatively correlated with the dependent variable, so when calculating the upper bound f of the NSGM(1,N) model s (k), take Sequence lower bound and The sequence constitutes a new sequence Matrix, similarly, and The sequence constitutes a new sequence matrix.

[0089] B4: Application Matrix The data inside predicts the future value of the system dependent variable of the NSGM (1, N) model to obtain the upper boundary, and the matrix is ​​applied The data inside predicts the future value of the system dependent variable using the NSGM(1,N) model to obtain the lower bound.

[0090] Optimized, the specific steps of outputting the interval prediction results of China's carbon dioxide emissions from the transportation industry are: according to the training input-output mode, the model outputs the interval prediction results of carbon dioxide emissions.

[0091] Beneficial Effects

[0092] (1) Through the construction of independent variable boundaries and the expansion of dependent variable intervals, the multivariate grey prediction model method system is enriched and improved;

[0093] (2) The multidimensional grey interval prediction model was used to predict carbon emissions in my country's transportation industry. Compared with an exact numerical value, the interval number is more flexible in policy reference. The research results can provide a theoretical basis for relevant government departments to formulate relevant policies;

[0094] (3) It provides an effective way to solve other small sample interval prediction problems.

[0095] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for interval prediction of carbon dioxide emissions in the transportation industry, characterized in that: The following steps are involved: S1: Collect raw data, form a data set, and perform preprocessing; S2: construct grey multivariate NSGM (1, N) model; S3: Use the univariate prediction model to predict the system's individual variables; S4: construct the boundaries of the system’s respective variables; S5: Output of interval forecast results of carbon dioxide emissions from China’s transportation industry.

2. The interval prediction method for carbon dioxide emissions in the transportation industry according to claim 1 is characterized in that: In step S1, the original data, i.e., the dependent variable, is collected, the independent variable data is screened according to the gray correlation degree, and then the determined data is preprocessed; in step S3, according to the data characteristics of each variable, a suitable single variable prediction model is selected to predict the independent variable, and the boundary is constructed to expand the prediction interval; in step S4, the boundary of the independent variable is constructed, and the NSGM (1, N) model is used to perform interval prediction on the future value of the system dependent variable.

3. The interval prediction method for carbon dioxide emissions in the transportation industry according to claim 2 is characterized in that: The specific steps of step S1 data preprocessing are: A1: Calculate each variable The first-order cumulative generation sequence of in, A2: Calculation The background value coefficient is the immediate generation sequence of the undetermined parameter λ in 4. The interval prediction method for carbon dioxide emissions in the transportation industry according to claim 1 is characterized in that: On the basis of the multivariate grey NSGM (1, N) model, the boundary of the independent variable is constructed so as to expand the prediction interval of the dependent variable. The specific steps of step S2 to construct the NSGM (1, N) model are as follows: B1: construct parameter matrices B and Y; B2: Calculate the parameters of the NSGM(1,N) model (I) When m = N + 3 and |B| ≠ 0, (II) When m>N+3 and |B T When B|≠0, (III) When m<N+3 and |B T When B|≠0, B3: The time response formula of NSGM(1,N) model is obtained by mathematical induction and iteration method; B4: Derive the cumulative generation formula of the NSGM(1,N) model; in, 5. The method for predicting carbon dioxide intervals in the transportation industry according to claim 4, characterized in that: In step S3, a suitable single variable prediction model is selected to predict each variable, and a performance analysis is performed based on the prediction results obtained by each model to determine whether the model can be used in this experiment.

6. The method for predicting carbon dioxide intervals in the transportation industry according to claim 5, characterized in that: The specific steps to construct variable boundaries are: B1: Construct positive / negative independent variables If b i >0, then and Positive correlation, defined as a positive independent variable If b i <0, then and Negative correlation, defined as a negative independent variable B2: Constructing positive / negative independent variable sequences when Then the positive / negative independent variable of the bounded interval is defined as a positive / negative sequence, is a forward sequence, is a negative sequence, B3: Constructing the matrix and Since the positive independent variable sequence is positively correlated with the dependent variable, the upper bound f of the NSGM (1, N) model is calculated. s (k), take Sequence upper bound Negative independent variable sequence Negatively correlated with the dependent variable, so when calculating the upper bound f of the NSGM (1, N) model s (k), take Sequence lower bound and The sequence constitutes a new sequence Matrix, similarly, and The sequence constitutes a new sequence matrix, B4: Application Matrix The data in the NSGM (1, N) model predicts the future value of the system dependent variable to obtain the upper boundary, and the matrix is ​​applied The data inside predicts the future value of the system dependent variable using the NSGM(1,N) model to obtain the lower bound.

7. The method for predicting carbon dioxide intervals in the transportation industry according to claim 6, characterized in that: The specific steps of outputting the prediction interval results of China's carbon dioxide emissions from the transportation industry are: according to the training input-output mode, the model output is the prediction interval results of China's carbon dioxide emissions from the transportation industry.

Images