CN115952888A - Multivariable grey model-based energy carbon emission prediction method and system - Google Patents
Multivariable grey model-based energy carbon emission prediction method and system Download PDFInfo
- Publication number
- CN115952888A CN115952888A CN202211551412.5A CN202211551412A CN115952888A CN 115952888 A CN115952888 A CN 115952888A CN 202211551412 A CN202211551412 A CN 202211551412A CN 115952888 A CN115952888 A CN 115952888A
- Authority
- CN
- China
- Prior art keywords
- carbon emission
- data
- model
- value
- influence factor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 title claims abstract description 116
- 229910052799 carbon Inorganic materials 0.000 title claims abstract description 116
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000012216 screening Methods 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000005265 energy consumption Methods 0.000 claims abstract description 11
- 239000002245 particle Substances 0.000 claims description 64
- 230000003044 adaptive effect Effects 0.000 claims description 28
- 238000010219 correlation analysis Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 13
- 238000005457 optimization Methods 0.000 claims description 12
- 238000002790 cross-validation Methods 0.000 claims description 6
- 230000006978 adaptation Effects 0.000 claims description 5
- 238000009825 accumulation Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 239000000654 additive Substances 0.000 claims description 3
- 230000000996 additive effect Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000000547 structure data Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 230000002087 whitening effect Effects 0.000 claims description 3
- 230000008685 targeting Effects 0.000 claims 1
- 238000013461 design Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/80—Management or planning
- Y02P90/84—Greenhouse gas [GHG] management systems
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an energy carbon emission prediction method based on a multivariable grey model, which comprises the steps of obtaining data information of a target area as a carbon emission influence factor set, and using energy consumption data as carbon emission source data; primarily screening factors of the carbon emission influence factor set; determining a final set of carbon emission influencing factors; inputting the carbon emission influence factor set into a multivariable grey model, and training to obtain an optimal carbon emission prediction model; and predicting the carbon emission of the target area by adopting a carbon emission prediction model. The invention also discloses a system for realizing the energy carbon emission prediction method based on the multivariable grey model. According to the energy carbon emission prediction method and system based on the multivariable gray model, the energy carbon emission is predicted through innovative algorithm design and prediction model design, and the method and system are high in reliability, good in practicability and high in precision.
Description
Technical Field
The invention belongs to the field of electrical automation, and particularly relates to a multivariable grey model-based energy carbon emission prediction method and system.
Background
With the development of economic technology and the improvement of living standard of people, electric energy becomes essential secondary energy in production and life of people, and brings endless convenience to production and life of people.
At present, with the increasing severity of environmental problems, carbon emission problems are receiving more and more attention. Accordingly, there is an increasing demand for electric energy, including the demand for electric vehicles, the demand for electric heaters and coolers, and so on. The carbon emission of one region is accurately predicted, and the subsequent work of power grid planning, power grid construction, power grid operation plan making and the like can be assisted to a power grid. Therefore, accurate carbon emission prediction becomes one of the important research points of the power system.
Currently, the existing carbon emission prediction methods include random factor regression, neural networks, time series and other methods. However, such methods take fewer factors into consideration, and the model requires a large amount of historical data to train. In addition, the accuracy of the existing prediction method is relatively poor.
Disclosure of Invention
The invention aims to provide an energy carbon emission prediction method based on a multivariable grey model, which has high reliability, good practicability and high precision.
The invention also aims to provide a system for realizing the multivariable grey model-based energy carbon emission prediction method.
The invention provides an energy carbon emission prediction method based on a multivariable grey model, which comprises the following steps:
s1, acquiring data information of a target area as a carbon emission influence factor set, and acquiring energy consumption data of the target area as carbon emission source data;
s2, primarily screening the factors in the carbon emission influence factor set based on a grey correlation analysis algorithm;
s3, based on the screening result of the step S2, determining a final carbon emission influence factor set by adopting a Lasso regression algorithm;
s4, optimizing a multivariable gray model by adopting a particle swarm algorithm, inputting the carbon emission influence factor set determined in the step S3 into the multivariable gray model, and training to obtain an optimal carbon emission prediction model by taking the average relative error absolute value as a target function;
and S5, adopting the carbon emission prediction model obtained in the step S4 to predict the carbon emission of the target area.
The step S1 of obtaining data information of the target area as a set of carbon emission influencing factors specifically includes the following steps:
obtaining socio-economic data, economic growth data, population data, urbanization level data, industry structure data, fixed asset investment data, outsourcer direct investment data, energy consumption data, vehicle holdup data and forest area data related to carbon emissions.
The grey correlation analysis algorithm-based preliminary screening of the factors in the carbon emission influence factor set in the step S2 specifically comprises the following steps:
setting variable sequence as X i Wherein the carbon emission sequence is X 0 ，X 1 ～X i Is a carbon emission influencing factor sequence; n is the number of samples within the sequence; i is the number of variables; total m carbon emission influencing factors;
X 0 ＝(x 0 (1),x 0 (2),...,x 0 (n)))
X i ＝(x i (1),x i (2),...,x i (n)))
X m ＝(x m (1),x m (2),...,x m (n)))
wherein i is 1,2,.. M;
mapping the data into an interval [0,1], thereby completing dimensionless processing;
then, the grey correlation analysis was performed using the following steps:
calculating the initial value X of the sequence i ' is X i '＝X i /X i (1)；
Calculating a difference sequence delta i (k) Is Δ i (k)＝|x' 0 (k)-x′ i (k)|；
Calculating a first range M of
Calculating a second pole difference m as>
Calculating a correlation coefficient gamma (x) 0 (k),x i (k) Is prepared from
Where xi is a resolution coefficient and xi is an element (0,1);
calculating the gray correlation degree gamma (X) 0 ,X i ) Is composed of
And finally, setting a correlation threshold, and judging the gray correlation of each influence factor:
if the grey correlation degree is larger than the correlation degree threshold value, retaining the corresponding influence factors;
if the grey correlation degree is smaller than or equal to the correlation degree threshold value, deleting the corresponding influence factors;
and finally, obtaining a carbon emission influence factor set after primary screening.
Step S3, determining a final set of carbon emission influencing factors by using a Lasso regression algorithm based on the screening result of step S2, specifically including the following steps:
the model for the Lasso regression is:
Y＝βX+ε
wherein Y is a vector of n X1, X is a vector of n X p, ε is a vector of n X1, β is a vector of regression coefficients of p X1 and β = (β) 1 ,β 2 ,...,β p )；
Lasso estimation
Is->
Wherein->
Lambda is a punishment item, and is a blending parameter and used for controlling the size of punishment degree;
the harmonic parameter λ is determined by a cross-validation algorithm: adopting K-fold cross validation, dividing the data set into K parts in equal proportion, training the model by taking K-1 parts as a training set, and fitting the model by taking the rest 1 part as a test set; and repeating the steps K times, and selecting a lambda value which finally enables the residual square sum to be minimum as a final harmonic parameter.
The step S4 of optimizing the multivariate gray model by adopting the particle swarm optimization, inputting the carbon emission influence factor set determined in the step S3 into the multivariate gray model, and training to obtain the optimal carbon emission prediction model by taking the absolute value of the average relative error as a target function specifically comprises the following steps:
A. performing smoothness verification:
carbon emission sequence X 0 Has a smoothness ratio ρ (k) of
k is k =2,3,.. N;
the quasi-smoothness ratio condition is constructed as
B. Generating a first order accumulation sequence
And &>
C. Generating an adjacent mean sequence:
constructing a first order additive sequence of carbon emissions
Is immediately adjacent to the generating sequence->
Is composed of
Wherein +>
D. Constructing a multivariate gray model as
Solving a parameter P by adopting a least square method, wherein the expression of the parameter P is P = [ a, b =] T (ii) a Finally, P = (B) is obtained T B) -1 B T Y is wherein
E. Constructing whitening equations
Resulting in a simulated value for the accumulation sequence>
Is composed of
F. Reducing the analog value of the accumulated sequence to the analog value of the original sequence to obtain
Is a reduced value;
for multivariate gray models, the background equation is constructed as:
optimizing a background equation parameter xi by adopting a self-adaptive particle swarm algorithm: using the minimization of the absolute value of the average relative error as a target function, performing parameter estimation by using a particle swarm optimization algorithm, and establishing an optimization model as
Wherein
The particle swarm optimization algorithm specifically comprises the following steps:
a. initialization: initializing a particle group, setting n particles in the particle group, and endowing each particle with a random initial position xi (i) and a speed v (i);
b. calculating an adaptive value: calculating the adaptive value of each particle according to the fitness function; the absolute value of the average relative error is used as an adaptive value;
c. solving the individual optimal adaptive value: for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical best position: if the adaptive value of the current position is higher, the current position is used as a new historical optimal position;
d. solving the group optimal adaptive value: for each particle, comparing the fitness value of its current position with the fitness value corresponding to its global optimal position: if the adaptive value of the current position is higher, taking the current position as a new global optimal position;
e. update particle position and velocity: the velocity and position of each particle is updated using the following formula:
v(i+1)＝ω×v(i)+c 1 ×rand()×(Pbest(i)-ξ(i))+c 2 ×rand()×(Gbest(i)-ξ(i))
ξ(i+1)＝ξ(i)+v(i)
wherein v (i + 1) is the updated particle velocity; v (i) is the particle velocity before update; omega is the inertial weight; c. C 1 And c 2 Is a learning factor; rand () is a generated random number between 0 and 1; pbest (i) is a local optimum adaptation value; gbest (i) is a global optimal adaptation value; ξ (i + 1) is the updated particle position; ξ (i) is the particle position before update;
f. judging whether the set finishing condition is met:
if the set ending condition is met, ending the algorithm, and finally obtaining the global optimal position as a global optimal solution;
and if the set ending condition is not met, repeating the steps b to f until the set ending condition is met.
The invention also provides a system for realizing the energy carbon emission prediction method based on the multivariable grey model, which specifically comprises a data acquisition module, a primary screening module, an influence factor determination module, a prediction model training module and a prediction module; the data acquisition module, the preliminary screening module, the influence factor determination module, the prediction model training module and the prediction module are sequentially connected in series; the data acquisition module is used for acquiring data information of a target area as a carbon emission influence factor set, acquiring energy consumption data of the target area as carbon emission source data, and uploading the data to the preliminary screening module; the preliminary screening module is used for preliminarily screening the factors in the carbon emission influence factor set based on a grey correlation analysis algorithm according to the acquired data and uploading the data to the influence factor determining module; the influence factor determining module is used for determining a final carbon emission influence factor set by adopting a Lasso regression algorithm according to the acquired data and based on a screening result, and uploading the data to the prediction model training module; the prediction model training module is used for optimizing the multivariable grey model by adopting a particle swarm algorithm according to the acquired data, inputting the determined carbon emission influence factor set into the multivariable grey model, training by taking the average relative error absolute value as a target function to obtain an optimal carbon emission prediction model, and uploading the data to the prediction module; and the prediction module is used for predicting the carbon emission of the target area by adopting the obtained carbon emission prediction model according to the acquired data.
According to the energy carbon emission prediction method and system based on the multivariable gray model, the energy carbon emission is predicted through innovative algorithm design and prediction model design, and the method and system are high in reliability, good in practicability and high in precision.
Drawings
FIG. 1 is a schematic process flow diagram of the process of the present invention.
FIG. 2 is a functional block diagram of the system of the present invention.
Detailed Description
FIG. 1 is a schematic flow chart of the method of the present invention: the invention provides an energy carbon emission prediction method based on a multivariable grey model, which comprises the following steps:
s1, acquiring data information of a target area as a carbon emission influence factor set, and acquiring energy consumption data of the target area as carbon emission source data; the method specifically comprises the following steps:
acquiring socioeconomic data, economic growth data, population data, urbanization level data, industrial structure data, fixed asset investment data, outsourcer direct investment data, energy consumption data, vehicle holdup data and forest area data related to carbon emission; the total carbon emission is obtained by adding the product of the total energy consumption and the corresponding carbon emission coefficient;
s2, primarily screening the factors in the carbon emission influence factor set based on a grey correlation analysis algorithm; the method specifically comprises the following steps:
setting variable sequence as X i Wherein the carbon emission sequence is X 0 ，X 1 ～X i Is a carbon emission influencing factor sequence; n is the number of samples within the sequence; i is the number of variables; total m carbon emission influencing factors;
X 0 ＝(x 0 (1),x 0 (2),...,x 0 (n)))
X i ＝(x i (1),x i (2),...,x i (n)))
X m ＝(x m (1),x m (2),...,x m (n)))
wherein i is 1,2,.. M;
because the carbon emission and the influence factor sequences have different dimensions, the difference between the data is large, and different sequences have no comparability; in order to eliminate the influence of dimension and value range difference among variables on a data result, non-dimensionalization processing is required; mapping the data into an interval [0,1], thereby completing dimensionless processing;
then, the grey correlation analysis was performed using the following steps:
calculating the initial value X of the sequence i ' is X i '＝X i /X i (1)；
Calculating a difference sequence delta i (k) Is Δ i (k)＝|x' 0 (k)-x′ i (k)|；
Calculating a first range M of
Calculating the second poleThe difference m is->
Calculating a correlation coefficient gamma (x) 0 (k),x i (k) Is prepared from
Where xi is a resolution coefficient and xi is an element (0,1);
calculating the gray correlation degree gamma (X) 0 ，X i )
And finally, setting a correlation threshold, and judging the gray correlation of each influence factor:
the degree of association is used for judging the degree of influence between the two factors by comparing the degrees of association, and if the change trends of the two factors are consistent, the synchronous change degree is higher, namely the degree of association between the two factors is higher; otherwise, it is lower;
if the grey correlation degree is greater than the correlation degree threshold value, retaining the corresponding influence factors;
if the grey correlation degree is less than or equal to the correlation degree threshold value, deleting the corresponding influence factors;
finally, obtaining a carbon emission influence factor set after primary screening;
s3, based on the screening result of the step S2, determining a final carbon emission influence factor set by adopting a Lasso regression algorithm; the method specifically comprises the following steps:
because only the variables with a relatively close relation with the carbon emission are obtained by grey correlation analysis, the relation between the influence factors, such as the influence of multiple collinearity, is not considered, and the problem of multiple collinearity can cause large deviation of a prediction result; therefore, on the basis of the result of grey correlation analysis, lasso regression is introduced to solve the problem of multiple collinearity among the influencing factors;
in the Lasso regression process, a penalty term is applied to the coefficients of the model, so that some coefficients tend to be 0 on the basis of least square estimation, the purpose of variable selection is achieved, the problem of multiple collinearity is solved, overfitting can be avoided, and the simplicity and the reasonability of the model are ensured;
the model for the Lasso regression is:
Y＝βX+ε
wherein Y is a vector of n X1, X is a vector of n X p, ε is a vector of n X1, β is a vector of regression coefficients of p X1 and β = (β) 1 ,β 2 ,...,β p )；
Lasso estimation
Is->
Wherein->
The lambda is a harmonic parameter and is used for controlling the size of punishment intensity;
the harmonic parameter λ is determined by a cross validation algorithm: adopting K-fold cross validation, dividing the data set into K parts in equal proportion, training the model by taking K-1 parts as a training set, and fitting the model by taking the rest 1 part as a test set; repeating the operation for K times, and selecting a lambda value which finally enables the residual error square sum to be minimum as a final harmonic parameter;
s4, optimizing a multivariate gray model by adopting a particle swarm algorithm, inputting the carbon emission influence factor set determined in the step S3 into the multivariate gray model, and training to obtain an optimal carbon emission prediction model by taking the average relative error absolute value as a target function; the method specifically comprises the following steps:
A. performing smoothness verification:
before grey modeling, smoothness inspection is carried out; the quasi-smoothness condition is an important criterion for checking whether a gray system model can be established for a sequence, and if the sequence meets the quasi-smoothness ratio condition, the sequence can be used for establishing the gray system model;
carbon emission sequence X 0 Has a smoothness ratio ρ (k) of
k is k =2,3. />
The quasi-smoothness ratio condition is constructed as
B. Generating a first order accumulation sequence
And &>
C. Generating a sequence of close-to-mean values:
constructing a first order additive sequence of carbon emissions
Is immediately adjacent to the generating sequence->
Is composed of
Wherein +>
D. Constructing a multivariate gray model as
Solving a parameter P by adopting a least square method, wherein the expression of the parameter P is P = [ a, b =] T (ii) a Finally, P = (B) is obtained T B) -1 B T Y is, wherein
E. Constructing whitening equations
Thereby obtaining an analog value->
Is composed of
F. Reducing the analog value of the accumulated sequence to the analog value of the original sequence to obtain
Is a reduced value;
the background value of the traditional multivariable grey model is set to 0.5, which is from the k-1 to the k time
Is approximated instead of a curved surface area and does not take into account the times k-1 and k->
Are of different importance, and therefore may cause the calculation result to cause a large error; based on the limitation, the particle swarm algorithm is adopted to improve the background value, so that the error caused by artificially setting the same weight for the background value can be reduced, and the model precision is improved; and (3) introducing a background equation coefficient xi to construct a new background equation: (ii) a
For a multivariate gray model, a background equation is constructed as follows:
optimizing a background equation parameter xi by adopting a self-adaptive particle swarm algorithm: the absolute value of the average relative error is minimized as a target function, the particle swarm optimization algorithm is utilized to carry out parameter estimation, and an optimization model is established as
Wherein
In specific implementation, the particle swarm algorithm uses particles which continuously move in a solution space as an optimization group, each particle has two attributes of position and speed (the dimension of the position and the dimension of the speed are the same as that of the solution space), the position of the particle represents a certain feasible solution, and the speed represents the difference value with the feasible solution found next; each particle adjusts the speed of the particle (according to a certain specific rule) according to the optimal solution searched by the particle and the optimal solution searched by the group at present so as to search a more optimal solution;
the particle swarm optimization algorithm specifically comprises the following steps:
a. initialization: initializing a particle group, setting n particles in the particle group, and endowing each particle with a random initial position xi (i) and a speed v (i);
b. calculating an adaptive value: calculating the adaptive value of each particle according to the fitness function; the absolute value of the average relative error is used as an adaptive value;
c. solving the individual optimal adaptive value: for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical best position: if the adaptive value of the current position is higher, the current position is used as a new historical optimal position;
d. solving the group optimal adaptive value: for each particle, comparing the fitness value of its current position with the fitness value corresponding to its global optimal position: if the adaptive value of the current position is higher, taking the current position as a new global optimal position;
e. update particle position and velocity: the velocity and position of each particle is updated using the following formula:
v(i+1)＝ω×v(i)+c 1 ×rand()×(Pbest(i)-ξ(i))+c 2 ×rand()×(Gbest(i)-ξ(i))
ξ(i+1)＝ξ(i)+v(i)
wherein v (i + 1) is the updated particle velocity; v (i) is the particle velocity before update; omega is the inertial weight; c. C 1 And c 2 Is a learning factor; rand () is a generated random number between 0 and 1; pbest (i) is a local optimum adaptation value; gbest (i) is a global optimal adaptive value; ξ (i + 1) is the updated particle position; ξ (i) is the particle position before update;
f. judging whether the set finishing condition is met:
if the set ending condition is met, ending the algorithm, and finally obtaining the global optimal position as a global optimal solution;
if the set ending condition is not met, repeating the steps b to f until the set ending condition is met;
and S5, adopting the carbon emission prediction model obtained in the step S4 to predict the carbon emission of the target area.
FIG. 2 is a schematic diagram of functional modules of the system of the present invention: the system for realizing the energy carbon emission prediction method based on the multivariable grey model specifically comprises a data acquisition module, a primary screening module, an influence factor determination module, a prediction model training module and a prediction module; the data acquisition module, the preliminary screening module, the influence factor determination module, the prediction model training module and the prediction module are sequentially connected in series; the data acquisition module is used for acquiring data information of a target area as a carbon emission influence factor set, acquiring energy consumption data of the target area as carbon emission source data, and uploading the data to the preliminary screening module; the preliminary screening module is used for preliminarily screening the factors in the carbon emission influence factor set based on a grey correlation analysis algorithm according to the acquired data and uploading the data to the influence factor determining module; the influence factor determining module is used for determining a final carbon emission influence factor set by adopting a Lasso regression algorithm according to the acquired data and based on a screening result, and uploading the data to the prediction model training module; the prediction model training module is used for optimizing the multivariable grey model by adopting a particle swarm algorithm according to the acquired data, inputting the determined carbon emission influence factor set into the multivariable grey model, training by taking the average relative error absolute value as a target function to obtain an optimal carbon emission prediction model, and uploading the data to the prediction module; and the prediction module is used for predicting the carbon emission of the target area by adopting the obtained carbon emission prediction model according to the acquired data.
Claims (7)
1. A multivariable grey model-based energy carbon emission prediction method comprises the following steps:
s1, acquiring data information of a target area as a carbon emission influence factor set, and acquiring energy consumption data of the target area as carbon emission source data;
s2, primarily screening the factors in the carbon emission influence factor set based on a grey correlation analysis algorithm;
s3, determining a final carbon emission influence factor set by adopting a Lasso regression algorithm based on the screening result of the step S2;
s4, optimizing a multivariate gray model by adopting a particle swarm algorithm, inputting the carbon emission influence factor set determined in the step S3 into the multivariate gray model, and training to obtain an optimal carbon emission prediction model by taking the average relative error absolute value as a target function;
and S5, adopting the carbon emission prediction model obtained in the step S4 to predict the carbon emission of the target area.
2. The method for predicting carbon emission from energy source based on multivariate gray model as claimed in claim 1, wherein the step S1 of obtaining data information of target area as carbon emission influencing factor set comprises the following steps:
acquiring socioeconomic data, economic growth data, population data, urbanization level data, industrial structure data, fixed asset investment data, outsourcer direct investment data, energy consumption data, vehicle holding amount data and forest area data related to carbon emission.
3. The method for predicting the carbon emission of the energy source based on the multivariate gray model according to claim 2, wherein the gray correlation analysis algorithm based on the step S2 is used for preliminarily screening the factors in the carbon emission influence factor set, and specifically comprises the following steps:
setting variable sequence as X i Wherein the carbon emission sequence is X 0 ，X 1 ～X i Is a carbon emission influencing factor sequence; n is the number of samples in the sequence; i is the number of variables; total m carbon emission influencing factors;
X 0 ＝(x 0 (1),x 0 (2),...,x 0 (n)))
X i ＝(x i (1),x i (2),...,x i (n)))
X m ＝(x m (1),x m (2),...,x m (n)))
wherein i is 1,2,.. M;
mapping the data into an interval [0,1], thereby completing dimensionless processing;
then, the grey correlation analysis was performed using the following steps:
calculating the initial value X of the sequence i ' is X i '＝X i /X i (1)；
Calculating the difference sequence delta i (k) Is Δ i (k)＝x' 0 (k)-x i '(k)；
Calculating a first range M of
Calculating a second pole difference m as->
Calculating a correlation coefficient gamma (x) 0 (k),x i (k) Is) as
Xi is a resolution coefficient and xi is epsilon (0,1);
calculating the gray correlation degree gamma (X) 0 ,X i ) Is composed of
And finally, setting a correlation threshold, and judging the gray correlation of each influence factor:
if the grey correlation degree is larger than the correlation degree threshold value, retaining the corresponding influence factors;
if the grey correlation degree is less than or equal to the correlation degree threshold value, deleting the corresponding influence factors;
and finally, obtaining the preliminarily screened carbon emission influence factor set.
4. The method according to claim 3, wherein the step S3 of determining the final set of carbon emission influencing factors by using a Lasso regression algorithm based on the screening results of the step S2 comprises the following steps:
the model for the Lasso regression is:
Y＝βX+ε
wherein Y is a vector of n X1, X is a vector of n X p, ε is a vector of n X1, β is a vector of regression coefficients of p X1 and β = (β) 1 ,β 2 ,...,β p )；
Lasso estimation
Is->
Wherein->
The lambda is a harmonic parameter and is used for controlling the size of punishment intensity;
the harmonic parameter λ is determined by a cross validation algorithm: dividing the data set into K parts in equal proportion by adopting K-fold cross validation, taking K-1 part as a training set to train the model, and taking the rest 1 part as a test set to fit the model; and repeating the steps for K times, and selecting the lambda value which finally enables the residual square sum to be minimum as a final harmonic parameter.
5. The multivariable gray model-based energy carbon emission prediction method according to claim 4, wherein the multivariable gray model is optimized by using a particle swarm optimization algorithm in step S4, the set of carbon emission influencing factors determined in step S3 is input into the multivariable gray model, and an average relative error absolute value is used as a target function to train the optimal carbon emission prediction model, which specifically comprises the following steps:
A. performing smoothness verification:
carbon emission sequence X 0 Has a slip ratio of rho (k) of
k is k =2,3.
The quasi-smoothness ratio condition is constructed as
B. Generating a first order accumulation sequence
And &>
C. Generating an adjacent mean sequence:
constructing a first order additive sequence of carbon emissions
Is immediately adjacent to the generating sequence->
Is->
Wherein +>
D. Constructing a multivariate gray model of
Solving a parameter P by adopting a least square method, wherein the expression of the parameter P is P = [ a, b =] T (ii) a Finally, P = (B) is obtained T B) -1 B T Y is, wherein
E. Constructing whitening equations
Thereby obtaining an analog value->
Is composed of
F. The analog value of the accumulated sequence is reduced to the analog value of the original sequence to obtain
Is a reduced value;
for a multivariate gray model, a background equation is constructed as follows:
optimizing a background equation parameter xi by adopting a self-adaptive particle swarm algorithm: targeting minimization of absolute value of average relative errorFunction, using particle swarm optimization algorithm to perform parameter estimation, and establishing an optimization model of
Wherein->
6. The multivariable grey model-based energy carbon emission prediction method according to claim 5, wherein the particle swarm optimization algorithm specifically comprises the following steps:
a. initialization: initializing a particle group, wherein the particle group is provided with n particles, and each particle is endowed with a random initial position xi (i) and a speed v (i);
b. calculating an adaptive value: calculating the adaptive value of each particle according to the fitness function; the absolute value of the average relative error is used as an adaptive value;
c. solving the individual optimal adaptive value: for each particle, comparing the adaptive value of the current position with the adaptive value corresponding to the historical best position: if the adaptive value of the current position is higher, using the current position as a new historical optimal position;
d. solving the group optimal adaptive value: for each particle, comparing the fitness value of its current position with the fitness value corresponding to its global optimal position: if the adaptive value of the current position is higher, taking the current position as a new global optimal position;
e. update particle position and velocity: the velocity and position of each particle is updated using the following formula:
v(i+1)＝ω×v(i)+c 1 ×rand()×(Pbest(i)-ξ(i))+c 2 ×rand()×(Gbest(i)-ξ(i))
ξ(i+1)＝ξ(i)+v(i)
wherein v (i + 1) is the updated particle velocity; v (i) is the particle velocity before update; omega is the inertial weight; c. C 1 And c 2 Is a learning factor; rand () is a generated random number between 0 and 1; pbest (i) is a local optimum adaptation value; gtest (i)) Is a global optimum adaptation value; ξ (i + 1) is the updated particle position; ξ (i) is the particle position before update;
f. judging whether the set finishing condition is met:
if the set ending condition is met, ending the algorithm, and finally obtaining the global optimal position as a global optimal solution;
and if the set ending condition is not met, repeating the steps b to f until the set ending condition is met.
7. A system for implementing the multivariate gray model-based energy carbon emission prediction method as defined in any one of claims 1 to 6, specifically comprising a data acquisition module, a preliminary screening module, an influence factor determination module, a prediction model training module and a prediction module; the data acquisition module, the preliminary screening module, the influence factor determination module, the prediction model training module and the prediction module are sequentially connected in series; the data acquisition module is used for acquiring data information of a target area as a carbon emission influence factor set, acquiring energy consumption data of the target area as carbon emission source data, and uploading the data to the preliminary screening module; the preliminary screening module is used for preliminarily screening the factors in the carbon emission influence factor set based on a grey correlation analysis algorithm according to the acquired data and uploading the data to the influence factor determining module; the influence factor determining module is used for determining a final carbon emission influence factor set by adopting a Lasso regression algorithm according to the acquired data and based on a screening result, and uploading the data to the prediction model training module; the prediction model training module is used for optimizing the multivariable grey model by adopting a particle swarm algorithm according to the acquired data, inputting the determined carbon emission influence factor set into the multivariable grey model, training by taking the average relative error absolute value as a target function to obtain an optimal carbon emission prediction model, and uploading the data to the prediction module; and the prediction module is used for predicting the carbon emission of the target area by adopting the obtained carbon emission prediction model according to the acquired data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211551412.5A CN115952888A (en) | 2022-12-05 | 2022-12-05 | Multivariable grey model-based energy carbon emission prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211551412.5A CN115952888A (en) | 2022-12-05 | 2022-12-05 | Multivariable grey model-based energy carbon emission prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115952888A true CN115952888A (en) | 2023-04-11 |
Family
ID=87295843
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211551412.5A Pending CN115952888A (en) | 2022-12-05 | 2022-12-05 | Multivariable grey model-based energy carbon emission prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115952888A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117373556A (en) * | 2023-12-04 | 2024-01-09 | 山东清控生态环境产业发展有限公司 | Tracing instrument and system based on multidimensional data |
| CN117893077A (en) * | 2024-01-12 | 2024-04-16 | 北京中创碳投科技有限公司 | Regional carbon emission quantification method and system based on multi-source data analysis and electric carbon model |
| CN118353004A (en) * | 2023-07-17 | 2024-07-16 | 华北电力大学 | Hydropower station short-term output prediction method based on AFSA-LSSVM |
-
2022
- 2022-12-05 CN CN202211551412.5A patent/CN115952888A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118353004A (en) * | 2023-07-17 | 2024-07-16 | 华北电力大学 | Hydropower station short-term output prediction method based on AFSA-LSSVM |
| CN117373556A (en) * | 2023-12-04 | 2024-01-09 | 山东清控生态环境产业发展有限公司 | Tracing instrument and system based on multidimensional data |
| CN117373556B (en) * | 2023-12-04 | 2024-02-13 | 山东清控生态环境产业发展有限公司 | Tracing instrument and system based on multidimensional data |
| CN117893077A (en) * | 2024-01-12 | 2024-04-16 | 北京中创碳投科技有限公司 | Regional carbon emission quantification method and system based on multi-source data analysis and electric carbon model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115952888A (en) | Multivariable grey model-based energy carbon emission prediction method and system | |
| CN107967542B (en) | Long-short term memory network-based electricity sales amount prediction method | |
| CN110245801A (en) | A kind of Methods of electric load forecasting and system based on combination mining model | |
| CN109063892A (en) | Industry watt-hour meter prediction technique based on BP-LSSVM combination optimization model | |
| CN116448419A (en) | Zero sample bearing fault diagnosis method based on depth model high-dimensional parameter multi-target efficient optimization | |
| CN107644297B (en) | Energy-saving calculation and verification method for motor system | |
| CN111310990A (en) | Improved gray combination model-based track quality prediction method and system | |
| CN112329990A (en) | User power load prediction method based on LSTM-BP neural network | |
| CN114881323A (en) | Foundation pit dewatering area underground water level prediction and updating method based on deep neural network | |
| CN113011102B (en) | Multi-time-sequence-based Attention-LSTM penicillin fermentation process fault prediction method | |
| CN114595861A (en) | MSTL (modeling, transformation, simulation and maintenance) and LSTM (least Square TM) model-based medium-and-long-term power load prediction method | |
| CN115619028A (en) | Clustering algorithm fusion-based power load accurate prediction method | |
| CN115860211A (en) | Casting blank quality prediction method based on local online modeling | |
| CN112149896A (en) | Attention mechanism-based mechanical equipment multi-working-condition fault prediction method | |
| CN113255963A (en) | Road surface use performance prediction method based on road element splitting and deep learning model LSTM | |
| CN108898257A (en) | A kind of gray prediction method based on particle group optimizing | |
| CN117313795A (en) | Intelligent building energy consumption prediction method based on improved DBO-LSTM | |
| Xu et al. | A novel method of neural network model predictive control integrated process monitoring and applications to hot rolling process | |
| CN113705932B (en) | Short-term load prediction method and device | |
| CN111435471A (en) | Heat supply gas consumption prediction model based on L STM | |
| CN114967625A (en) | Industrial process quality prediction method based on space-time convolution latent variable optimization neighbor component analysis | |
| Kaligambe et al. | Indoor Room Temperature and Relative Humidity Estimation in a Commercial Building Using the XGBoost Machine Learning Algorithm | |
| CN109376957A (en) | A kind of prediction technique of thermal power plant's load | |
| Wang et al. | Operational optimization of copper flotation process based on the weighted Gaussian process regression and index-oriented adaptive differential evolution algorithm | |
| CN114417248B (en) | DCGAN-based linear contour process quality anomaly monitoring method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |