CN106529732A - Carbon emission efficiency prediction method based on neural network and random frontier analysis - Google Patents
Carbon emission efficiency prediction method based on neural network and random frontier analysis Download PDFInfo
- Publication number
- CN106529732A CN106529732A CN201611025929.5A CN201611025929A CN106529732A CN 106529732 A CN106529732 A CN 106529732A CN 201611025929 A CN201611025929 A CN 201611025929A CN 106529732 A CN106529732 A CN 106529732A
- Authority
- CN
- China
- Prior art keywords
- carbon emission
- gdp
- pop
- emission amount
- history
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- 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
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses a carbon emission efficiency prediction method and system based on a neural network and random frontier analysis. The method comprises the steps that carbon emission impact factor GDP and population POP are selected; historical GDP and POP data and regional carbon emission CE in a corresponding period are acquired; historical GDP and POP data are used as input, and the regional carbon emission CE in the corresponding period is used as output to build a self-adaptive fuzzy neural network model; the acquired historical GDP and POP data and a time series model are used to predict GDP and POP in a future period; the predicted GDP and POP are input into the self-adaptive fuzzy neural network model to predict regional carbon emission in the future corresponding period; according to the predicted GDP and POP and the regional carbon emission in the future period, a random frontier analysis model is built; a maximum likelihood method is used to estimate each parameter value of the random frontier analysis model; and the conditional expectations of the technical inefficiencies of GDP and POP are used as the technical efficiency of the regional carbon emission respectively.
Description
Technical field
The invention belongs to region carbon emission forecast analysis field, more particularly to a kind of neutral net and stochastic frontier analysis
Carbon emission efficiency Forecasting Methodology.
Background technology
Global climate change is pushed carbon emission " the teeth of the storm " to, and low-carbon emission reduction becomes every country, area and all considering
Hot issue.In order to realize sustainable development, alleviate greenhouse effect, the research with regard to region carbon emission is subject to increasing
Concern.
China is currently under in the industrialization and urbanization process of high speed, the contradiction that economic development is constrained with energy resources
Increasingly project, add the huge external pressure that Global Green House Gas Emissions Reduction is brought so that low-carbon (LC) development must be walked by China.
The variation tendency and the contribution efficiency of influence condition of carbon emission during Chinese development is found out, is to explore low carbon development path
Premise, not only conforms with the development trend of global " low-carbon (LC) ", and the certainty for implementing the reduction of greenhouse gas discharge target that country proposes
Require.
It can be seen that, there is very big emission reduction in China, will satisfactorily complete emission reduction tasks at this stage, first have to clear and definite China's carbon
The Changing Pattern of discharge and growth trend.The main side for adopting time series models and mathematics to adjust of traditional carbon emission prediction
Method, these methods need accurately data and clear and definite internal relation, but the carbon emission system of entire society is a complexity
, there is too many uncertainty and internal system complex nonlinear relation in system, this is the insurmountable problem of traditional method.
The content of the invention
In order to solve the shortcoming of prior art, the first object of the present invention be to provide it is a kind of based on neutral net with it is random before
Along the carbon emission efficiency Forecasting Methodology of analysis.
The carbon emission efficiency Forecasting Methodology based on neutral net and stochastic frontier analysis of the present invention, including:
Step 1:Choose factor of influence GDP and size of population POP of carbon emission;Wherein, GDP and POP is separate
Variable Factors;
Step 2:Obtain history GDP and POP data and corresponding period inner region carbon emission amount CE;By history GDP and
Used as input, corresponding period inner region carbon emission amount CE builds Adaptive Fuzzy Neural-network mould as output to POP data
Type;
Step 3:Predicted in the following regular period using history GDP and POP data and time series models for obtaining
GDP and POP, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model, and prediction obtains following corresponding period
Interior region carbon emission amount;
Step 4:According to the GDP in the following regular period of prediction and POP and region carbon emission amount, stochastic frontier is built
Analysis model, estimates Stochastic Frontier Analysis Model parameters value by maximum likelihood method, then with technical ineffectiveness rate item
Technical efficiency of the conditional expectation respectively as GDP and POP respectively to region carbon emission amount.
In the step 2, also include history GDP and POP data and the corresponding period inner region carbon emission that will be obtained
Amount CE be divided into two groups, one group as training group, another group used as validation group.
General training group accounts for more than the 70% of total data, and the present invention builds adaptive fuzzy using the data of training group
Neural network model, after the completion of Adaptive Fuzzy Neural-network model training, verifies self adaptation using the data of validation group
The accuracy of fuzzy neural network model.
Stochastic Frontier Analysis Model in the step 4 is determined by Cobb-Douglas production functions.So can be very
The randomness of processing data, can be estimated to past technical efficiency and future is predicted, and analyze knot well
Fruit can provide theoretical foundation as the basis of research carbon emission behavior for reducing discharging policy making.
In the step 2, Adaptive Fuzzy Neural-network model is the fuzzy inference system based on Sugeno models.
Fuzzy reasoning and neutral net are organically combined, its fuzzy membership and fuzzy rule are by mass data
Study is completed, and predicts following region carbon emission amount using adaptive neural network, can avoid complexity in system and
Uncertainty, is constantly corrected by autonomic learning, can obtain optimal solution.
Carbon emission efficiency Forecasting Methodology of the present invention based on neutral net with stochastic frontier analysis adopts adaptive neural network net
Network can be avoided complexity and uncertainty in system, constantly be repaiied by autonomic learning predicting following region carbon emission amount
Just, optimal solution can be obtained;The efficiency analysiss that this method can be carried out over carbon emission in future time period with reference to SFA simultaneously,
SFA is capable of the randomness of processing data well, past technical efficiency can be estimated and future is predicted, and
And analysis result can provide theoretical foundation as the basis of research carbon emission behavior for reducing discharging policy making.
The second object of the present invention there is provided a kind of pre- with the carbon emission efficiency of stochastic frontier analysis based on neutral net
Examining system.
The carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis of the present invention, including:
Factor of influence chooses module, and which is used for choosing the factor of influence GDP of carbon emission and size of population POP;Wherein, GDP
It is separate Variable Factors with POP;
Adaptive Fuzzy Neural-network model construction module, its be used for obtaining history GDP and POP data and it is corresponding when
Phase inner region carbon emission amount CE;Using history GDP and POP data as input, corresponding period inner region carbon emission amount CE conduct
Export to build Adaptive Fuzzy Neural-network model;
Region carbon emission amount prediction module, its be used for using history GDP and POP data and time series models for obtaining come
GDP and POP in the prediction following regular period, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model,
Prediction obtains the region carbon emission amount in following corresponding period;
The technical efficiency computing module of region carbon emission amount, its be used for according to prediction the following regular period in GDP with
POP and region carbon emission amount, build Stochastic Frontier Analysis Model, estimate Stochastic Frontier Analysis Model by maximum likelihood method
Parameters value, then respectively with the conditional expectation of the technical ineffectiveness rate item of GDP and POP as each to region carbon emission amount
Technical efficiency.
The Adaptive Fuzzy Neural-network model construction module, be additionally operable to will obtain history GDP and POP data and
Corresponding period inner region carbon emission amount CE is divided into two groups, one group as training group, another group used as validation group.
Stochastic Frontier Analysis Model in the technical efficiency computing module of the region carbon emission amount is by Cobb-Douglas
Production function is determining.
Carbon emission efficiency prognoses system of the present invention based on neutral net with stochastic frontier analysis adopts adaptive neural network net
Network can be avoided complexity and uncertainty in system, constantly be repaiied by autonomic learning predicting following region carbon emission amount
Just, optimal solution can be obtained;The efficiency analysiss that the system can be carried out over carbon emission in future time period with reference to SFA simultaneously,
SFA is capable of the randomness of processing data well, past technical efficiency can be estimated and future is predicted, and
And analysis result can provide theoretical foundation as the basis of research carbon emission behavior for reducing discharging policy making.
The third object of the present invention there is provided another kind based on neutral net and the carbon emission efficiency of stochastic frontier analysis
Prognoses system.
The carbon emission efficiency prognoses system with stochastic frontier analysis based on neutral net is somebody's turn to do, including:
Data acquisition unit, which is used for obtaining history GDP and POP data and corresponding period inner region from data base
Carbon emission amount CE;Wherein, GDP and POP is separate Variable Factors;
Memorizer, which is used for storing history GDP and POP data and corresponding period inner region carbon emission amount CE for obtaining;
Server, which is configured to:
Obtain history GDP and POP data and corresponding period inner region carbon emission amount CE from memorizer, then by history
Used as input, corresponding period inner region carbon emission amount CE builds adaptive fuzzy nerve net as output to GDP and POP data
Network model;
Predicted using history GDP and POP data and time series models for obtaining the GDP in the following regular period and
POP, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model, and prediction obtains the area in following corresponding period
Domain carbon emission amount;
According to the GDP in the following regular period of prediction and POP and region carbon emission amount, stochastic frontier analysis mould is built
Type, estimates Stochastic Frontier Analysis Model parameters value by maximum likelihood method, then respectively with the technology of GDP and POP without
The conditional expectation of efficiency item is used as technical efficiency each to region carbon emission amount.
The server is additionally configured to history GDP and POP data and the corresponding period inner region carbon emission that will be obtained
Amount CE be divided into two groups, one group as training group, another group used as validation group.
The Stochastic Frontier Analysis Model is determined by Cobb-Douglas production functions.
Beneficial effects of the present invention are:
(1) carbon emission efficiency Forecasting Methodology of the present invention based on neutral net with stochastic frontier analysis adopts adaptive neural network
Network can avoid complexity and uncertainty in system predicting following region carbon emission amount, continuous by autonomic learning
Amendment, can obtain optimal solution;This method can be carried out over and future time period with reference to stochastic frontier analysis method (SFA) simultaneously
The randomness of processing data well is capable of in the efficiency analysiss of interior carbon emission, stochastic frontier analysis method (SFA), can be to the past
Technical efficiency be estimated and future be predicted, and analysis result can as research carbon emission behavior basis,
Theoretical foundation is provided for reducing discharging policy making.
(2) carbon emission efficiency prognoses system of the present invention based on neutral net with stochastic frontier analysis adopts adaptive neural network
Network can avoid complexity and uncertainty in system predicting following region carbon emission amount, continuous by autonomic learning
Amendment, can obtain optimal solution;The efficiency point that the system can be carried out over carbon emission in future time period with reference to SFA simultaneously
The randomness of processing data well is capable of in analysis, SFA, past technical efficiency can be estimated and future be carried out pre-
Survey, and analysis result can provide theoretical foundation as the basis of research carbon emission behavior for reducing discharging policy making.
Description of the drawings
Fig. 1 is the present invention based on neutral net and the carbon emission efficiency Forecasting Methodology flow chart of stochastic frontier analysis;
Fig. 2 is two input single order Fuzzy Inference Models;
Fig. 3 is Adaptive Fuzzy Neural-network model calculation schematic diagram;
Fig. 4 is the technical efficiency of Stochastic Frontier Analysis Model;
Fig. 5 is a kind of carbon emission efficiency prognoses system structural representation based on neutral net and stochastic frontier analysis of the present invention
Figure;
Fig. 6 is that another kind of carbon emission efficiency prognoses system structure based on neutral net with stochastic frontier analysis of the invention is shown
It is intended to.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.
The invention belongs to region Prediction of Greenhouse Gas field, more particularly to a kind of to be based on adaptive neural network
Forecasting Methodology realize prediction to region greenhouse gas emission, and the region greenhouse gas emission based on stochastic frontier analysis
The Analysis of technical efficiency method of influence factor.
Fig. 1 is a kind of carbon emission efficiency Forecasting Methodology flow process based on neutral net and stochastic frontier analysis of the present invention
Figure.The carbon emission efficiency Forecasting Methodology based on neutral net and stochastic frontier analysis of the present invention as shown in Figure 1, including:
Step 1:Choose factor of influence GDP and size of population POP of carbon emission;Wherein, GDP and POP is separate
Variable Factors.
Step 2:Obtain history GDP and POP data and corresponding period inner region carbon emission amount CE;By history GDP and
Used as input, corresponding period inner region carbon emission amount CE builds Adaptive Fuzzy Neural-network mould as output to POP data
Type;
In step 2, also include history GDP and POP data and corresponding period inner region carbon emission amount CE that will be obtained
Be divided into two groups, one group as training group, another group used as validation group.
General training group accounts for more than the 70% of total data, and the present invention builds adaptive fuzzy using the data of training group
Neural network model, after the completion of Adaptive Fuzzy Neural-network model training, verifies self adaptation using the data of validation group
The accuracy of fuzzy neural network model.
In step 2, Adaptive Fuzzy Neural-network model is the fuzzy inference system based on Sugeno models.
Fuzzy reasoning and neutral net are organically combined, its fuzzy membership and fuzzy rule are by mass data
Study is completed, and predicts following region carbon emission amount using adaptive neural network, can avoid complexity in system and
Uncertainty, is constantly corrected by autonomic learning, can obtain optimal solution.
In the specific implementation process of step 2:
The size of population (POP), inputs of the GDP as Adaptive Fuzzy Neural-network model (ANFIS), is output as region carbon
Discharge capacity (CE).ANFIS is a kind of fuzzy inference system based on Sugeno models, will be fuzzy reasoning and neutral net organic
It is to be completed by the study of mass data with reference to, its fuzzy membership and fuzzy rule, two classical input network structures are such as
Shown in Fig. 2.ANFIS structure ground floors are |input paramete and obfuscation, and x, y are input variables;Ai、BiIt is related to the node
Fuzzy variable;O1,i, O2,iIt is fuzzy set A respectivelyi, BiMembership function, wherein:
Now i=1,2;
Now i=3,4;
Determine given input meet volume Ai、BiDegree.Here Ai、BiDegree of membership can be any appropriate parameter
Change membership function, such as Gauss member function, bell membership function etc..
The second layer, calculates dimmed excitation intensity wi, computing formula is as follows:
Third layer, calculates the ratio of i-th excitation density and total excitation density
This node layer carries out the normalization of each rule relevance grade and calculates, i.e.,:Calculate the w of the i-th ruleiWith whole rules
Relevance grade sum ∑ wiRatio.
4th layer, this layer of all of node is all adaptive, calculates output O4, i:
Wherein, pi, qi, riModel parameter collection, the referred to as consequent parameter of fuzzy reasoning, can pass through method of least square or
Possibility predication is obtained;fiFor consequent (conclusion) output function of Sugeno fuzzy systems, when which is that linear function is then referred to as " one
Level is united ", if constant is then referred to as " 0 level system ".
Layer 5 calculates the 4th layer of input signal.
Construction Adaptive Fuzzy Neural-network model as shown in Figure 3.
Step 3:Predicted in the following regular period using history GDP and POP data and time series models for obtaining
GDP and POP, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model, and prediction obtains following corresponding period
Interior region carbon emission amount;
In step 3, as GDP is separate variable with the size of population (POP), carried out using time series models
Prediction GDP and POP, illustrates using simple autoregression model to explain here for convenient.Autoregression model is:
Xt,i=aXt-1,i+b;I=1,2
Wherein, Xt,1For GDP;Xt,2For POP;A and b are constant coefficient;T be time serieses 1,2,3 .... }.
By the use of GDP, the POP in the following regular period of prediction as the input of ANFIS, to following regular period inner region
Carbon emission amount be predicted.
Step 4:According to the GDP in the following regular period of prediction and POP and region carbon emission amount, stochastic frontier is built
Analysis model, estimates Stochastic Frontier Analysis Model parameters value by maximum likelihood method, then respectively with GDP's and POP
The conditional expectation of technical ineffectiveness rate item is used as technical efficiency each to region carbon emission amount.
Stochastic Frontier Analysis Model in step 4 is determined by Cobb-Douglas production functions.So can be fine
The randomness of ground processing data, can be estimated to past technical efficiency and future is predicted, and analysis result
Theoretical foundation can be provided as the basis of research carbon emission behavior for reducing discharging policy making.
In order to carry out the behavior analysiss of greenhouse gas emission, using RANDOM BOUNDARY analytical GDP and POP for temperature
The technical efficiency of room gas discharge.In this invention, technical efficiency is GDP, the POP under optimized situation point under identical carbon emission amount
Not with practical situation under GDP, POP ratio, thus carry out the Emission Reduction Potential of assessment area.
Stochastic Frontier Analysis Model hereinafter referred to as SFA models.
After introducing the concept of time in the present invention, SFA models are allow to carry out efficiency rating to sample data.Concrete model
It is as follows:
Yjt=f (xjt,β)exp(vjt)exp(-ujt), j=1 ... N, t=1 ... T (1)
In formula, YjtIt is the t period carbon outputs (i.e. term area j carbon emission amounts during t) of j-th decision package, xjtIt is jth
Whole influence factors (being GDP, POP in the present invention) in the t periods of individual decision package, f (xjt, β) for input-output model really
Qualitative part, i.e., do not consider the model of random disturbance, and β is model parameter to be estimated, maximums of the T for period t.
By random disturbance εjIt is divided into two parts:For representing statistical error, be otherwise known as a part stochastic error, uses vjt
To represent;Another part is used for the inefficiency of presentation technology, the non-negative that is otherwise known as error term, u in the modeljt=uj exp(-η
(t-T)) it is non-negative error term, η is to estimate parameter, and its size is relevant with the level of disruption of time with error.
The present invention regards carbon emission as the output of social production activity, and the production function in model is selected with Cobb-
As a example by Douglas production functions, formula (1) can be write as following linear forms:
Wherein k represents variable number, and in this example, k=1 is GDP, and k=2 is POP.β0, βkFor model parameter to be estimated, YjtFor
Carbon emission amounts of the region j in t periods, xjtThere is the factor of considerable influence in t periods to carbon emission amount for region j.
Model of the present invention has hypothesis below:
(1) stochastic errorThat is vjtIt is that average is for 0, varianceIndependent normal distribution,
Mainly caused by uncontrollable factor, such as natural disaster etc..(iid:independent identically distributed
Independent same distribution)
(2) non-negative error termWherein, ujtIt is that average is for 0, varianceIndependent normal
Distribution, takes cutting gearbox (N+Expression is clipped<0 part), production efficiency is represent, is the efficiency of carbon emission in this,
And have ujtWith vjtIt is separate.
(3)ujt、vjtWith variable xjtIt is separate.
In the present invention, SFA technical efficiency is:
So, in utDistribution it is known in the case of, meansigma methodss TE of technical efficiency can be calculatedt=E [exp (-
ujt)], i.e. t periods GDP, POP can analyze the trend of following carbon emission for the technical efficiency of carbon emission (CE) based on this, special
It is not the variation tendency affected by GDP and POP, is that policy making is provided fundamental basis.
SFA methods estimate parameters value by maximum likelihood method, are then made with the conditional expectation of technical ineffectiveness rate item
For technical efficiency value.The information and result of calculation that SFA methods take full advantage of each sample is stable, by particular point affected compared with
It is little, have the advantages that comparability is strong, reliability is high.
As shown in figure 4, by taking Cobb-Douglas production functions as an example, it is shown that the advantage of SFA modelling technique efficiency measures
(employ single-throw for convenience of explanation to enter, i.e., x).In figure, by the production proportions face that Cobb-Douglas production functions determine it is:
lnYt=β0+β1lnxt
Wherein, YtFor the output in t periods, it is carbon emission amount in the invention;xtFor the input in t periods, the invention be GDP or
Person POP;β0, β1Parameter is estimated for model, can be obtained by methods such as method of least square, Maximum-likelihood estimations.
And based on this stochastic Frontier Model for determining production proportions face be:lnYt=β0+β1lnxt+vt-ut, it is also possible to table
It is shown as:Yt=exp (β0+β1lnxt+vt-ut).Wherein, vtFor stochastic error, utFor non-negative error term.
Wherein, 2 points of A, B represents that Random Effect is the situation just or for bearing respectively:
A points represent Random Effect for just, then stochastic error vAFor positive number, production proportions are moved on on face:The technical efficiency of sample isB points represent with
Machine affects to be negative, then stochastic error vBFor negative, production proportions face moves down intoSample
Technical efficiency is
Carbon emission efficiency Forecasting Methodology of the present invention based on neutral net with stochastic frontier analysis adopts adaptive neural network net
Network can be avoided complexity and uncertainty in system, constantly be repaiied by autonomic learning predicting following region carbon emission amount
Just, optimal solution can be obtained;The efficiency analysiss that this method can be carried out over carbon emission in future time period with reference to SFA simultaneously,
SFA is capable of the randomness of processing data well, past technical efficiency can be estimated and future is predicted, and
And analysis result can provide theoretical foundation as the basis of research carbon emission behavior for reducing discharging policy making.
Fig. 5 is that a kind of carbon emission efficiency prognoses system structure based on neutral net with stochastic frontier analysis of the present invention is shown
It is intended to.The carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis of the present invention as shown in Figure 5, including:
Factor of influence chooses module, and which is used for choosing the factor of influence GDP of carbon emission and size of population POP;Wherein, GDP
It is separate Variable Factors with POP;
Adaptive Fuzzy Neural-network model construction module, its be used for obtaining history GDP and POP data and it is corresponding when
Phase inner region carbon emission amount CE;Using history GDP and POP data as input, corresponding period inner region carbon emission amount CE conduct
Export to build Adaptive Fuzzy Neural-network model;
Region carbon emission amount prediction module, its be used for using history GDP and POP data and time series models for obtaining come
GDP and POP in the prediction following regular period, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model,
Prediction obtains the region carbon emission amount in following corresponding period;
The technical efficiency computing module of region carbon emission amount, its be used for according to prediction the following regular period in GDP with
POP and region carbon emission amount, build Stochastic Frontier Analysis Model, estimate Stochastic Frontier Analysis Model by maximum likelihood method
Parameters value, then respectively with the conditional expectation of the technical ineffectiveness rate item of GDP and POP as each to region carbon emission amount
Technical efficiency.
Wherein, Adaptive Fuzzy Neural-network model construction module, be additionally operable to will obtain history GDP and POP data with
And corresponding period inner region carbon emission amount CE is divided into two groups, one group as training group, another group used as validation group.
Stochastic Frontier Analysis Model in the technical efficiency computing module of region carbon emission amount is produced by Cobb-Douglas
Function is determining.
Carbon emission efficiency prognoses system of the present invention based on neutral net with stochastic frontier analysis adopts adaptive neural network net
Network can be avoided complexity and uncertainty in system, constantly be repaiied by autonomic learning predicting following region carbon emission amount
Just, optimal solution can be obtained;The efficiency analysiss that the system can be carried out over carbon emission in future time period with reference to SFA simultaneously,
SFA is capable of the randomness of processing data well, past technical efficiency can be estimated and future is predicted, and
And analysis result can provide theoretical foundation as the basis of research carbon emission behavior for reducing discharging policy making.
Fig. 6 is that a kind of carbon emission efficiency prognoses system structure based on neutral net with stochastic frontier analysis of the present invention is shown
It is intended to.The carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis of the present invention as shown in Figure 6, including:
Data acquisition unit, which is used for obtaining history GDP and POP data and corresponding period inner region from data base
Carbon emission amount CE;Wherein, GDP and POP is separate Variable Factors;
Memorizer, which is used for storing history GDP and POP data and corresponding period inner region carbon emission amount CE for obtaining;
Server, which is configured to:
Obtain history GDP and POP data and corresponding period inner region carbon emission amount CE from memorizer, then by history
Used as input, corresponding period inner region carbon emission amount CE builds adaptive fuzzy nerve net as output to GDP and POP data
Network model;
Predicted using history GDP and POP data and time series models for obtaining the GDP in the following regular period and
POP, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model, and prediction obtains the area in following corresponding period
Domain carbon emission amount;
According to the GDP in the following regular period of prediction and POP and region carbon emission amount, stochastic frontier analysis mould is built
Type, estimates Stochastic Frontier Analysis Model parameters value by maximum likelihood method, then respectively with the technology of GDP and POP without
The conditional expectation of efficiency item is used as technical efficiency each to region carbon emission amount.
Wherein, data acquisition unit and memorizer are existing structure.
Server is additionally configured to history GDP and POP data and corresponding period inner region carbon emission amount CE that will be obtained
Be divided into two groups, one group as training group, another group used as validation group.
Stochastic Frontier Analysis Model is determined by Cobb-Douglas production functions.
Carbon emission efficiency prognoses system of the present invention based on neutral net with stochastic frontier analysis adopts adaptive neural network net
Network can be avoided complexity and uncertainty in system, constantly be repaiied by autonomic learning predicting following region carbon emission amount
Just, optimal solution can be obtained;The efficiency analysiss that the system can be carried out over carbon emission in future time period with reference to SFA simultaneously,
SFA is capable of the randomness of processing data well, past technical efficiency can be estimated and future is predicted, and
And analysis result can provide theoretical foundation as the basis of research carbon emission behavior for reducing discharging policy making.
One of ordinary skill in the art will appreciate that all or part of flow process in realizing above-described embodiment method, can be
Instruct related hardware to complete by computer program, described program can be stored in a computer read/write memory medium
In, the program is upon execution, it may include such as the flow process of the embodiment of above-mentioned each method.Wherein, described storage medium can be magnetic
Dish, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random
AccessMemory, RAM) etc..
Although the above-mentioned accompanying drawing that combines is described to the specific embodiment of the present invention, not to present invention protection model
The restriction enclosed, one of ordinary skill in the art should be understood that on the basis of technical scheme those skilled in the art are not
The various modifications made by needing to pay creative work or deformation are still within protection scope of the present invention.
Claims (10)
1. a kind of carbon emission efficiency Forecasting Methodology based on neutral net and stochastic frontier analysis, it is characterised in that include:
Step 1:Choose factor of influence GDP and size of population POP of carbon emission;Wherein, GDP and POP is separate variable
The factor;
Step 2:Obtain history GDP and POP data and corresponding period inner region carbon emission amount CE;By history GDP and POP numbers
According to as input, corresponding period inner region carbon emission amount CE builds Adaptive Fuzzy Neural-network model as output;
Step 3:Predicted using history GDP and POP data and time series models for obtaining the GDP in the following regular period and
POP, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model, and prediction obtains the area in following corresponding period
Domain carbon emission amount;
Step 4:According to the GDP in the following regular period of prediction and POP and region carbon emission amount, stochastic frontier analysis are built
Model, estimates Stochastic Frontier Analysis Model parameters value by maximum likelihood method, then respectively with the technology of GDP and POP
The conditional expectation of inefficiency item is used as technical efficiency each to region carbon emission amount.
2. a kind of carbon emission efficiency Forecasting Methodology based on neutral net and stochastic frontier analysis as claimed in claim 1, its
It is characterised by, in the step 2, also includes history GDP and POP data and the corresponding period inner region carbon emission that will be obtained
Amount CE be divided into two groups, one group as training group, another group used as validation group.
3. a kind of carbon emission efficiency Forecasting Methodology based on neutral net and stochastic frontier analysis as claimed in claim 1, its
It is characterised by, the Stochastic Frontier Analysis Model in the step 4 is determined by Cobb-Douglas production functions.
4. a kind of carbon emission efficiency Forecasting Methodology based on neutral net and stochastic frontier analysis as claimed in claim 1, its
It is characterised by, in the step 2, Adaptive Fuzzy Neural-network model is the fuzzy inference system based on Sugeno models.
5. a kind of carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis, it is characterised in that include:
Factor of influence chooses module, and which is used for choosing the factor of influence GDP of carbon emission and size of population POP;Wherein, GDP and POP
It is separate Variable Factors;
Adaptive Fuzzy Neural-network model construction module, which is used for obtaining history GDP and POP data and in corresponding period
Region carbon emission amount CE;Using history GDP and POP data as input, corresponding period inner region carbon emission amount CE is used as output
To build Adaptive Fuzzy Neural-network model;
Region carbon emission amount prediction module, which is used for using history GDP and POP data and time series models for obtaining predicting
GDP and POP in the following regular period, the GDP and POP of prediction are input into into Adaptive Fuzzy Neural-network model, prediction
Obtain the region carbon emission amount in following corresponding period;
The technical efficiency computing module of region carbon emission amount, its be used for according to prediction the following regular period in GDP and POP with
And region carbon emission amount, build Stochastic Frontier Analysis Model, by maximum likelihood method estimate Stochastic Frontier Analysis Model each
Parameter value, then uses the conditional expectation of technical ineffectiveness rate item of GDP and POP respectively as technology each to region carbon emission amount
Efficiency.
6. a kind of carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis as claimed in claim 5, its
Be characterised by, the Adaptive Fuzzy Neural-network model construction module, be additionally operable to will obtain history GDP and POP data with
And corresponding period inner region carbon emission amount CE is divided into two groups, one group as training group, another group used as validation group.
7. a kind of carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis as claimed in claim 5, its
It is characterised by, the Stochastic Frontier Analysis Model in the technical efficiency computing module of the region carbon emission amount is by Cobb-Douglas
Production function is determining.
8. a kind of carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis, it is characterised in that include:
Data acquisition unit, which is used for obtaining history GDP and POP data and corresponding period inner region carbon row from data base
High-volume CE;Wherein, GDP and POP is separate Variable Factors;
Memorizer, which is used for storing history GDP and POP data and corresponding period inner region carbon emission amount CE for obtaining;
Server, which is configured to:
Obtain history GDP and POP data and corresponding period inner region carbon emission amount CE from memorizer, then by history GDP
With POP data as input, corresponding period inner region carbon emission amount CE builds Adaptive Fuzzy Neural-network as output
Model;
GDP and POP in the following regular period are predicted using history GDP and POP data and time series models for obtaining, will
The GDP and POP of prediction is input into into Adaptive Fuzzy Neural-network model, and prediction obtains the region carbon row in following corresponding period
High-volume;
According to the GDP in the following regular period of prediction and POP and region carbon emission amount, Stochastic Frontier Analysis Model is built,
Stochastic Frontier Analysis Model parameters value is estimated by maximum likelihood method, then respectively with the technical ineffectiveness rate of GDP and POP
The conditional expectation of item is used as technical efficiency each to region carbon emission amount.
9. a kind of carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis as claimed in claim 8, its
It is characterised by, the server is additionally configured to history GDP and POP data and the corresponding period inner region carbon emission that will be obtained
Amount CE be divided into two groups, one group as training group, another group used as validation group.
10. a kind of carbon emission efficiency prognoses system based on neutral net and stochastic frontier analysis as claimed in claim 8, its
It is characterised by, the Stochastic Frontier Analysis Model is determined by Cobb-Douglas production functions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611025929.5A CN106529732A (en) | 2016-11-18 | 2016-11-18 | Carbon emission efficiency prediction method based on neural network and random frontier analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611025929.5A CN106529732A (en) | 2016-11-18 | 2016-11-18 | Carbon emission efficiency prediction method based on neural network and random frontier analysis |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106529732A true CN106529732A (en) | 2017-03-22 |
Family
ID=58351835
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611025929.5A Pending CN106529732A (en) | 2016-11-18 | 2016-11-18 | Carbon emission efficiency prediction method based on neural network and random frontier analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106529732A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109740301A (en) * | 2019-03-14 | 2019-05-10 | 华北电力大学 | A kind of accounting method of the Gas Generator Set carbon emission amount based on BP neural network |
CN111008735A (en) * | 2019-11-27 | 2020-04-14 | 巴斯夫新材料有限公司 | Predictive emission management system and method |
CN111882033A (en) * | 2020-07-15 | 2020-11-03 | 南京航空航天大学 | Keras-based regional civil aviation active and passive carbon emission prediction method |
CN113866387A (en) * | 2021-09-24 | 2021-12-31 | 中国农业科学院农业资源与农业区划研究所 | Regional scale soil carbon emission prediction method and system |
CN114548481A (en) * | 2021-12-26 | 2022-05-27 | 特斯联科技集团有限公司 | Power equipment carbon neutralization processing apparatus based on reinforcement learning |
CN114707774A (en) * | 2022-06-07 | 2022-07-05 | 山东科技大学 | Method and device for predicting carbon emission based on transportation |
CN115496286A (en) * | 2022-09-26 | 2022-12-20 | 重庆德宜高大数据科技有限公司 | Neural network carbon emission prediction method based on big data environment and application |
CN117408394A (en) * | 2023-12-14 | 2024-01-16 | 国网天津市电力公司电力科学研究院 | Carbon emission factor prediction method and device for electric power system and electronic equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201749550U (en) * | 2010-02-04 | 2011-02-16 | 上海辉格科技发展有限公司 | Mobile automatic carbon emission calculation system |
CN103439463A (en) * | 2013-08-16 | 2013-12-11 | 深圳中建院建筑科技有限公司 | Real-time online monitoring system for carbon emission of building |
CN203811206U (en) * | 2014-04-15 | 2014-09-03 | 重庆市计量质量检测研究院 | Enterprise carbon discharge metering and analysis device |
-
2016
- 2016-11-18 CN CN201611025929.5A patent/CN106529732A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201749550U (en) * | 2010-02-04 | 2011-02-16 | 上海辉格科技发展有限公司 | Mobile automatic carbon emission calculation system |
CN103439463A (en) * | 2013-08-16 | 2013-12-11 | 深圳中建院建筑科技有限公司 | Real-time online monitoring system for carbon emission of building |
CN203811206U (en) * | 2014-04-15 | 2014-09-03 | 重庆市计量质量检测研究院 | Enterprise carbon discharge metering and analysis device |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109740301A (en) * | 2019-03-14 | 2019-05-10 | 华北电力大学 | A kind of accounting method of the Gas Generator Set carbon emission amount based on BP neural network |
CN111008735A (en) * | 2019-11-27 | 2020-04-14 | 巴斯夫新材料有限公司 | Predictive emission management system and method |
CN111008735B (en) * | 2019-11-27 | 2023-06-02 | 巴斯夫新材料有限公司 | Predictive emissions management system and method |
CN111882033A (en) * | 2020-07-15 | 2020-11-03 | 南京航空航天大学 | Keras-based regional civil aviation active and passive carbon emission prediction method |
CN111882033B (en) * | 2020-07-15 | 2024-04-05 | 南京航空航天大学 | Keras-based regional civil aviation main passive carbon emission prediction method |
CN113866387A (en) * | 2021-09-24 | 2021-12-31 | 中国农业科学院农业资源与农业区划研究所 | Regional scale soil carbon emission prediction method and system |
CN114548481A (en) * | 2021-12-26 | 2022-05-27 | 特斯联科技集团有限公司 | Power equipment carbon neutralization processing apparatus based on reinforcement learning |
CN114707774A (en) * | 2022-06-07 | 2022-07-05 | 山东科技大学 | Method and device for predicting carbon emission based on transportation |
CN115496286A (en) * | 2022-09-26 | 2022-12-20 | 重庆德宜高大数据科技有限公司 | Neural network carbon emission prediction method based on big data environment and application |
CN117408394A (en) * | 2023-12-14 | 2024-01-16 | 国网天津市电力公司电力科学研究院 | Carbon emission factor prediction method and device for electric power system and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Dong et al. | Hourly energy consumption prediction of an office building based on ensemble learning and energy consumption pattern classification | |
CN106529732A (en) | Carbon emission efficiency prediction method based on neural network and random frontier analysis | |
CN108846517B (en) | Integration method for predicating quantile probabilistic short-term power load | |
CN109902801B (en) | Flood collective forecasting method based on variational reasoning Bayesian neural network | |
Lu et al. | US natural gas consumption prediction using an improved kernel-based nonlinear extension of the Arps decline model | |
Zolfaghari et al. | Modeling and predicting the electricity production in hydropower using conjunction of wavelet transform, long short-term memory and random forest models | |
Pousinho et al. | A hybrid PSO–ANFIS approach for short-term wind power prediction in Portugal | |
Moeini et al. | Fuzzy rule-based model for hydropower reservoirs operation | |
Semero et al. | A PSO-ANFIS based hybrid approach for short term PV power prediction in microgrids | |
Awan et al. | Improving ANFIS based model for long-term dam inflow prediction by incorporating monthly rainfall forecasts | |
CN106022614A (en) | Data mining method of neural network based on nearest neighbor clustering | |
CN106991507A (en) | A kind of SCR inlet NOx concentration on-line prediction method and device | |
Zou et al. | Wind turbine power curve modeling using an asymmetric error characteristic-based loss function and a hybrid intelligent optimizer | |
Lv et al. | Novel deterministic and probabilistic combined system based on deep learning and self-improved optimization algorithm for wind speed forecasting | |
CN105046453A (en) | Construction engineering project cluster establishment method introducing cloud model for evaluation and selection | |
Wang et al. | Short-term load forecasting of power system based on time convolutional network | |
CN109858665A (en) | Photovoltaic short term power prediction technique based on Feature Selection and ANFIS-PSO | |
Niu et al. | Short term load forecasting model using support vector machine based on artificial neural network | |
CN111680712A (en) | Transformer oil temperature prediction method, device and system based on similar moments in the day | |
Shi et al. | A fuzzy time series forecasting model with both accuracy and interpretability is used to forecast wind power | |
Gorbatiuk et al. | Application of fuzzy time series forecasting approach for predicting an enterprise net income level | |
Zhu | Research on adaptive combined wind speed prediction for each season based on improved gray relational analysis | |
Hernandez-Ambato et al. | Multistep-ahead streamflow and reservoir level prediction using ANNs for production planning in hydroelectric stations | |
CN106655266A (en) | Method for configuring flexibly adjustable power supplies of new energy-integrated regional power network | |
Du | Prediction of consumer price index based on RBF neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170322 |