CN104484701A - Neural network evaluation method for environment issues - Google Patents
Neural network evaluation method for environment issues Download PDFInfo
- Publication number
- CN104484701A CN104484701A CN201410700572.0A CN201410700572A CN104484701A CN 104484701 A CN104484701 A CN 104484701A CN 201410700572 A CN201410700572 A CN 201410700572A CN 104484701 A CN104484701 A CN 104484701A
- Authority
- CN
- China
- Prior art keywords
- evaluation
- environmental
- neural network
- neuroid
- environmental problem
- 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.)
- Granted
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a neural network evaluation method for environment issues. The neural network evaluation method comprises the following steps of S1, collecting the data for evaluation of various environment issues; S2, establishing a database for the evaluation of the environment issues; S3, obtaining objective and subjective factors with significant effects on the evaluation of the environment issues; S4, analyzing main components of the factors with significant effects on the evaluation of the environment issues, obtaining independent factors with significant effects on the evaluation of the environment issues, using the independent factors with significant effects as input variables of a neural network model for the evaluation of the environment issues, and establishing a primary model of a neural network for the evaluation of the environment issues; S5, obtaining the optimum neural network model to evaluate the environment issues. The neural network evaluation method has the beneficial effects that protection schemes can be provided for environment engineers in advance, and the important quantitative basis is provided for establishing corresponding protection measures in city planning phases.
Description
Technical field
The present invention relates to the evaluation method of environmental problem, particularly relate to a kind of neuroid evaluation method of environmental problem.
Background technology
Along with High-speed Urbanization, China's environmental problem is more and more outstanding.At present, China, for the processing mode of environmental problem, is mainly solved by Environmental Engineer, takes the passive treatment method of one " treat the head when the head aches, pin pain cure pin ", lacks systematization, integration, prospective countermeasure.Mainly owing to lacking, evaluation means that is scientific, architecture causes for this.
Current environmental evaluation technology carries out field investigation, actual measurement mainly for the environmental problem existed, and according to the statistical study to investigation, measured result, finds environmental problem crux, thus proposes solution.
This environmental assessment techniques remedying formula can only be the evaluation for a certain specific region, specific environment.Be the environmental assessment techniques of a kind of " thing one is discussed ", direct help is not had to other similar environmental problem evaluations.Meanwhile, such method is the repairing to destroyed environment, and does not have direct reference function to the development in future.
Current, Fast Urbanization makes the speed of speed less than urban development of solution environmental problem.This mainly traditional environmental assessment techniques cannot be evaluated the environmental problem that Urban Planning Stage is hidden, thus causes the environmental problem after building to continue to bring out.
Existing environmental evaluation technology is comparatively simple, is not suitable for solving Universal Problems.The environmental evaluation information obtained is not made full use of, wastes the manpower and materials investigated on the spot.This method does not make full use of the advantage of information resource database, causes the limitation solving environmental problem.
In general, the greatest drawback of prior art is to provide perspective counter-measure for environmental problem.The fundamental purpose of environmental evaluation is the main crux understanding environmental problem.A lot of environmental problem has very similar problem crux, by the investigation and analysis to a certain class problem, should be able to play forewarning function to the generation of Similar Problems from now on.Therefore, utilize the mass data that environment is investigated, adopt computer science and technology, change the environmental assessment techniques of existing " thing one is discussed ", propose a kind of can just can the method for anticipation environmental problem in the planning construction stage, it is fundamental purpose of the present invention that contingent environmental problem was solved in the city planning design stage.
Summary of the invention
In order to solve the problems of the prior art; the invention provides a kind of neuroid evaluation method of environmental problem of scientific, architecture; urban planner can be made, designer understands urban construction to the influence degree of environment in the Urban Planning and Design stage, is the quantitative basis that Environmental Engineer provides protection scheme in advance, sets up corresponding safeguard measure to provide important for Urban Planning Stage.
The invention provides a kind of neuroid evaluation method of environmental problem, comprise the following steps:
S1, gather all kinds of environmental problem evaluate data;
S2, set up all kinds of environmental problem evaluate database;
The impact that S3, the data understanding all kinds of environmental problem evaluation by correlation analysis are evaluated all kinds of environmental problem, obtains the factor had a significant impact environmental problem evaluation;
The correlation analysis result of S4, Corpus--based Method model, the factor had a significant impact environmental problem evaluation is carried out principal component analysis (PCA), obtain there is significantly independent sex factor to environmental evaluation, this there is is significantly independent sex factor as the input variable of environmental evaluation neural network model to environmental evaluation, sets up the rudimentary model of environmental evaluation neural network model;
S5, the database utilizing step S2 to set up, train the rudimentary model of the environmental evaluation neural network model that step S4 sets up, check, check, optimize, and obtains optimum neural network model.
As a further improvement on the present invention, the data of all kinds of environmental problem evaluations in step S1 comprise air, water, noise, soil, the objective environment data of bio-diversity and the subjective assessment data of people.
As a further improvement on the present invention, step S1 is: the historical data and the as-is data that gather the evaluation of all kinds of environmental problem.
As a further improvement on the present invention, step S3, by adopting the method for Statistic analysis models, solves the correlativity of input variable and output variable, obtains the factor had a significant impact environmental problem evaluation.
As a further improvement on the present invention, step S4 will have significantly independent sex factor as environmental evaluation neural network model input variable to environmental evaluation, set up the rudimentary model of environmental evaluation neuroid, the rudimentary model of this environmental evaluation neuroid determines middle number of plies n according to the difference of the number M of input variable and output variable number N, and be used for being the neuron number m of computing, based on the number difference of input variable and output variable, arrange with the number of plies of difference number few 1 to 2 as neuron models structural sheet, every layer of neuron number is set to m-2 of upper strata neuron m number, and the rudimentary model of multiple environmental evaluation neuroid is set up with such set-up mode.
As a further improvement on the present invention, in step S5, the rudimentary model of environmental evaluation neural network model is optimized and comprises: according to neuroid back-propagation algorithm, under the support of database, determines optimum neural network model.
As a further improvement on the present invention, in step S5, the rudimentary model of environmental evaluation neural network model carries out in the process of training, the relatively rudimentary model result of environmental evaluation neural network model and the difference of legitimate reading, utilize and observe the self-adjusting parameter of back-propagation algorithm, the rudimentary model structure of environmental evaluation neural network model is finely tuned, obtains optimum neural network model.
The invention has the beneficial effects as follows: pass through such scheme, make full use of a large amount of actual investigational data, set up magnanimity environmental information storehouse, and then utilize neuroid to the advanced learning ability of large data, develop the evaluation method of the anticipation formula to " FUTURE ENVIRONMENT " that city planning embodies.Environmental protection work can be advanceed to the preliminary stage of urban construction by such technology; stop the phenomenon generation that recent environmental problems " first destroys; administer afterwards "; urban planner can be made, designer understands urban construction to the influence degree of environment in the Urban Planning and Design stage, is the quantitative basis that Environmental Engineer provides protection scheme in advance, sets up corresponding safeguard measure to provide important for Urban Planning Stage.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the neuroid evaluation method of a kind of environmental problem of the present invention;
Fig. 2 is the structural representation of the neural network model of the neuroid evaluation method of a kind of environmental problem of the present invention;
Fig. 3 is the optimizing process schematic diagram of the neural network model of the neuroid evaluation method of a kind of environmental problem of the present invention.
Embodiment
Illustrate below in conjunction with accompanying drawing and embodiment the present invention is further described.
As shown in Figure 1 to Figure 3, the present invention, according to the mass data of actual environment problem, builds the database of environmental evaluation under different scene, thus sets up the magnanimity case of neuroid Learning demands.By affecting the condition element of environmental evaluation in statistics correlation analysis model investigation society, determine the input variable of environmental assessment neuroid, neuroid structure is adjusted again by model calculation, and choose the complicacy that optimized parameter reduces network model, obtain evaluation result accurately.
In model calculation process, need the mass data of grasp to be divided into learning data, to check data and check data.In model construction process, adopting model structure by the principle of letter to difficulty, by adjusting the mode of neuron node number and the hiding number of plies, based on the data of magnanimity, carrying out Optimized model, avoid " over-fitting " phenomenon occurs, choose best model.
A neuroid evaluation method for environmental problem, comprises the following steps:
(1) gather historical data and as-is data that all kinds of environmental problem evaluates, comprise the objective environment data such as air, water, noise, soil, bio-diversity, the subjective data of people and subjective assessment data;
(2) magnanimity " database " that all kinds of environmental problem is evaluated is set up;
(3) the subjective and objective data of environment are understood on the impact of all kinds of environmental evaluation by correlation analysis.Due to " robustness " of neuroid, general neural network model does not consider the relation of constrained input variable, but the factor likely affecting output variable is all thought of as input variable.But, evaluate this problem for environmental problem, because the factor affecting environmental evaluation is very extensive, causes needing the variable of input neuron network very huge, thus make neural network structure too complicated, add the difficulty of computing; Also the result that computing cannot restrain may be caused.This difficulty is also cause the less reason for studying subjective assessment problem of neuroid.The present invention then by adopting the method for Statistic analysis models, solves the correlativity of constrained input variable; Using conspicuousness correlative factor as input variable, thus considerably reduce the quantity of input variable, simplify, optimize neuroid structure, and then can obtain better, more accurately neuroid predict the outcome, improve feasibility and the accuracy of neural network model prediction environment subjective assessment problem.Thisly utilize statistical model, determine environmental impact input variable, simplify neuroid structure, make it more effectively, experimental technique more accurately, be one of main innovate point of the present invention;
(4) the correlation analysis result of Corpus--based Method model, by the subjective and objective factor principal component analysis again had a significant impact environmental evaluation, determines there is significantly independent sex factor to environmental evaluation; Using the input variable of these factors as environmental evaluation neural network model, set up environmental evaluation neuroid rudimentary model.Rudimentary model structure carrys out the middle number of plies n of Confirming model according to the difference of the number M of input variable and output variable number N, and is used for being the neuron number m of computing.Based on the number difference of input variable and output variable, arrange with the number of plies of difference number few 1 to 2 as neuron models structural sheet; Every layer of neuron number is set to m-2 (see figure 2) of upper strata neuron m number, and sets up multiple neuroid rudimentary model that may exist with such set-up mode;
(5) mass data of environmental problem rating database is utilized, neuroid rudimentary model is trained, checks and check, according to the advantage of neuroid back-propagation algorithm, under mass data supports, determine optimum neural network model, eliminate the model that the used time is many, accuracy is lower; In model training process, the difference of comparison model result of calculation and legitimate reading, utilizes and observes the self-adjusting parameter of back-propagation algorithm, finely tune model structure, makes it to reach optimum, the most accurate compute mode (see figure 3).The present invention utilizes neuroid to have the advantage analyzed linearity and non-linearity problem, realize utilizing " environment subjective and objective factor " to predict " environment subjective assessment " this complexity problem, thus realize the object of anticipation latency environment problem in city planning design process.The accuracy of neural network model prediction is the depth and broadness of its study case.Because environmental impact issues has chronicity, multiple feature, therefore there are long-term environmental monitoring data and enquiry data in each urban environmental protection department general, carries out environmental problem evaluation to provide the foundation condition to utilizing neural network model.Meanwhile, due in Chinese Urbanization evolution, the environmental problem constantly occurred has the feature of repeatability, therefore the neural network model of development based on long-term mass data accumulation, can Similar Problems accurately in the following construction in anticipation city; Thus make full use of time and again the data information of environmental problem investigation, play environmental data maximal efficiency.
The role of evaluation that current environmental assessment technology is not forward-looking, just for the discussion of practical problems, the method for proposition also can only solve the environmental problem of a certain concrete generation.And the main contributions of the neuroid evaluation method of a kind of environmental problem provided by the invention is to give full play to the large data message advantage that environmental problem is investigated on the spot, utilize artificial neuron's anticipation formula function of computer science, environmental data is carried out deep excavation, for on the study analysis basis of a large amount of existing data, by " robustness " function of neuron models, anticipation is carried out to the like environment problem of " being about to occur ", thus can at the preliminary stage of urban construction, the impact that abundant assessment construction may be caused environment, thus better can solve urban development and environmental harmony problem, be conducive to the developing goal realizing China's Ecological Civilization Construction.The present invention can carry out the globality of environmental problem, universality, anticipation assessment in the first stage of construction stage, breach the limitation of existing method " thing one is discussed ".
The key job of the neuroid evaluation method of a kind of environmental problem provided by the invention is the Pre-Evaluation of intellectual technology development to built environment utilizing neural network model.Main innovate point is the application of computer technology in environmental engineering and city planning design crossing domain.Therefore, invention needs the main points of protection to be the construction method of such neuron models of development, the structure choice method of neuron models, and neuron models calculate in a series of technical essential such as parameter selection method.
Above content is in conjunction with concrete preferred implementation further description made for the present invention, can not assert that specific embodiment of the invention is confined to these explanations.For general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, some simple deduction or replace can also be made, all should be considered as belonging to protection scope of the present invention.
Claims (7)
1. a neuroid evaluation method for environmental problem, is characterized in that, comprise the following steps:
S1, gather all kinds of environmental problem evaluate data;
S2, set up all kinds of environmental problem evaluate database;
The impact that S3, the data understanding all kinds of environmental problem evaluation by correlation analysis are evaluated all kinds of environmental problem, obtains the factor had a significant impact environmental problem evaluation;
The correlation analysis result of S4, Corpus--based Method model, the factor had a significant impact environmental problem evaluation is carried out principal component analysis (PCA), obtain there is significantly independent sex factor to environmental evaluation, this there is is significantly independent sex factor as the input variable of environmental evaluation neural network model to environmental evaluation, sets up the rudimentary model of environmental evaluation neural network model;
S5, the database utilizing step S2 to set up, train the rudimentary model of the environmental evaluation neural network model that step S4 sets up, check, check, optimize, and obtains optimum neural network model.
2. the neuroid evaluation method of environmental problem according to claim 1, is characterized in that: the data of all kinds of environmental problem evaluations in step S1 comprise air, water, noise, soil, the objective environment data of bio-diversity and the subjective assessment data of people.
3. the neuroid evaluation method of environmental problem according to claim 1, is characterized in that: step S1 is: the historical data and the as-is data that gather the evaluation of all kinds of environmental problem.
4. the neuroid evaluation method of environmental problem according to claim 1, it is characterized in that: step S3 is by adopting the method for Statistic analysis models, solve the correlativity of input variable and output variable, obtain the factor that environmental problem evaluation is had a significant impact.
5. the neuroid evaluation method of environmental problem according to claim 1, it is characterized in that: step S4 will have significantly independent sex factor as environmental evaluation neural network model input variable to environmental evaluation, set up the rudimentary model of environmental evaluation neuroid, the rudimentary model of this environmental evaluation neuroid determines middle number of plies n according to the difference of the number M of input variable and output variable number N, and be used for being the neuron number m of computing, based on the number difference of input variable and output variable, arrange with the number of plies of difference number few 1 to 2 as neuron models structural sheet, every layer of neuron number is set to m-2 of upper strata neuron m number, and the rudimentary model of multiple environmental evaluation neuroid is set up with such set-up mode.
6. the neuroid evaluation method of environmental problem according to claim 1, it is characterized in that: in step S5, the rudimentary model of environmental evaluation neural network model is optimized and comprises: according to neuroid back-propagation algorithm, under the support of database, determine optimum neural network model.
7. the neuroid evaluation method of environmental problem according to claim 1, it is characterized in that: in step S5, the rudimentary model of environmental evaluation neural network model carries out in the process of training, the relatively rudimentary model result of environmental evaluation neural network model and the difference of legitimate reading, utilize and observe the self-adjusting parameter of back-propagation algorithm, the rudimentary model structure of environmental evaluation neural network model is finely tuned, obtains optimum neural network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410700572.0A CN104484701B (en) | 2014-11-28 | 2014-11-28 | The neuroid evaluation method of environmental problem |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410700572.0A CN104484701B (en) | 2014-11-28 | 2014-11-28 | The neuroid evaluation method of environmental problem |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104484701A true CN104484701A (en) | 2015-04-01 |
| CN104484701B CN104484701B (en) | 2017-11-21 |
Family
ID=52759242
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410700572.0A Active CN104484701B (en) | 2014-11-28 | 2014-11-28 | The neuroid evaluation method of environmental problem |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104484701B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107330608A (en) * | 2017-06-27 | 2017-11-07 | 宁波创元信息科技有限公司 | A kind of production scheduling method based on Techniques of Neural Network |
| CN111861282A (en) * | 2020-08-05 | 2020-10-30 | 深圳博沃智慧科技有限公司 | Method and device for evaluating environmental assessment results and computer-readable storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101908197A (en) * | 2009-11-16 | 2010-12-08 | 北京师范大学 | Analysis method of urban ecosystem |
-
2014
- 2014-11-28 CN CN201410700572.0A patent/CN104484701B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101908197A (en) * | 2009-11-16 | 2010-12-08 | 北京师范大学 | Analysis method of urban ecosystem |
Non-Patent Citations (4)
| Title |
|---|
| 古今今等: "基于综合主成分及径向基网络的环境质量评价", 《环境科学与技术》 * |
| 吴建新等: "吕泗大洋港近岸海域浮游动物生态特征及其环境评价", 《海洋科学》 * |
| 张莹等: "基于主成分分析-神经网络法的湛江海域水质评价研究", 《热带海洋科学学术研讨会暨广东海洋学会会员***论文及2013》 * |
| 雷璐宁等: "基于改进的SOM神经网络在水质评价分析中的应用", 《仪器仪表学报》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107330608A (en) * | 2017-06-27 | 2017-11-07 | 宁波创元信息科技有限公司 | A kind of production scheduling method based on Techniques of Neural Network |
| CN107330608B (en) * | 2017-06-27 | 2021-04-23 | 宁波创元信息科技有限公司 | Production scheduling method based on neuron network technology |
| CN111861282A (en) * | 2020-08-05 | 2020-10-30 | 深圳博沃智慧科技有限公司 | Method and device for evaluating environmental assessment results and computer-readable storage medium |
| CN111861282B (en) * | 2020-08-05 | 2024-01-12 | 深圳博沃智慧科技有限公司 | Criticizing result evaluation method, criticizing result evaluation device and computer-readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104484701B (en) | 2017-11-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guo et al. | A multi-variable grey model with a self-memory component and its application on engineering prediction | |
| Hu et al. | Modern machine learning techniques for univariate tunnel settlement forecasting: A comparative study | |
| Wang et al. | Evaluation of compaction quality based on SVR with CFA: case study on compaction quality of earth-rock dam | |
| Wang et al. | Correction of soil parameters in calculation of embankment settlement using a BP network back-analysis model | |
| Jia et al. | Seismic fragility analysis of RC frame-shear wall structure under multidimensional performance limit state based on ensemble neural network | |
| Teng et al. | Decision-making tools for evaluation the impact on the eco-footprint and eco-environmental quality of green building development policy | |
| Liu et al. | A systems dynamic model of a coal-based city with multiple adaptive scenarios: A case study of Ordos, China | |
| Shi et al. | A novel seepage safety monitoring model of CFRD with slab cracks using monitoring data | |
| Xu et al. | A hybrid approach for dynamic simulation of safety risks in mega construction projects | |
| Yi et al. | Research methods of water resources carrying capacity: Progress and prospects | |
| Peng et al. | Research on the prediction of the water demand of construction engineering based on the BP neural network | |
| Zhou et al. | Scenario prediction and critical factors of CO2 emissions in the Pearl River Delta: A regional imbalanced development perspective | |
| CN104063609B (en) | Method of assisting in judging pollution source monitoring data validity by utilizing neural network | |
| CN106759546B (en) | Based on the Deep Foundation Distortion Forecast method and device for improving multivariable grey forecasting model | |
| CN104484701A (en) | Neural network evaluation method for environment issues | |
| Wang | Application of deep learning model in building energy consumption prediction | |
| Yi et al. | A deep leaning method for dynamic vibration analysis of bridges subjected to uniform seismic excitation | |
| Shi et al. | Forecasting and advancing water carrying capacity in Henan Province in China: Application of ‘four determinations with water’in AHP and SD modeling | |
| Guo et al. | How efficient is the environmental pollution control in China? | |
| Francis et al. | Integrating lean construction and sustainability via a system dynamics framework | |
| Wang et al. | A new framework for distributed storage tanks placement based on a resilience characteristic metric and reduced modelling | |
| Song et al. | Fast inversion method for seepage parameters of core earth-rock dam based on LHS-SSA-MKELM fusion surrogate model | |
| Li | Construction of a Smart City Network Information Security Evaluation Model Based on GRA-BPNN | |
| Zhang et al. | Buffer Monitoring of Critical Chain Projects Based on Support Vector Machine Prediction | |
| Liu et al. | Crack prediction based on wavelet correlation analysis least squares support vector machine for stone cultural relics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |