CN106874960A - Based on the fisheries environment class of pollution appraisal procedure for improving SVMs - Google Patents
Based on the fisheries environment class of pollution appraisal procedure for improving SVMs Download PDFInfo
- Publication number
- CN106874960A CN106874960A CN201710122690.1A CN201710122690A CN106874960A CN 106874960 A CN106874960 A CN 106874960A CN 201710122690 A CN201710122690 A CN 201710122690A CN 106874960 A CN106874960 A CN 106874960A
- Authority
- CN
- China
- Prior art keywords
- pollution
- fisheries
- environment
- index
- svms
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Mining & Mineral Resources (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- General Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Agronomy & Crop Science (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Farming Of Fish And Shellfish (AREA)
Abstract
A kind of fisheries water environmental pollution grade appraisal procedure based on improvement SVMs.The invention belongs to environmental evaluation technical field, its the step of is the correlative factor for analyzing influence fisheries environment quality in marine engineering first, then evaluation index is built according to correlative factor, and for index sample data adds label, then evaluation index is combined with supporting vector machine model, carry out Support Vector Machines Optimized model with reference to grid search and cross validation method, establish the complex nonlinear relation between evaluation index and fisheries environment credit rating, make evaluation result more scientific and reasonable, meet objective reality, preferable environmental quality grade assessment can be also obtained in the case of Finite Samples.
Description
Technical field
The invention belongs to Environmental Quality Evalution technical field, more particularly to a kind of fishery water based on improvement SVMs
Domain environmental pollution grade appraisal procedure.
Technical background
Because marine engineering is in process of construction, substantial amounts of suspension bed sediment can be produced, and poisonous having of containing in suspension
Evil material may dissolution produces secondary pollution again in the seawater, have impact on the seawater quality in construction marine site and surrounding waters and sinks
Product environment, and then have impact on the existence of fishery aquatile.Therefore to relating to the fisheries environment investigation before and after extra large engineering construction
And environmental pollution grade assessment is carried out, and marine environment situation of change before and after construction can be fully grasped, provided for follow-up fishery
Source and ecological environmental protection are significant.
The grade assessment of Marine fishery eco-environment quality is influenceed by many factors, dirty for fisheries environment in relating to fisherfolk's journey
The grade evaluation studies of dye, be more by the way that some Mathematical Modelings are quantitative or qualitative analysis obtains the infringement of fishery, and this
Invention considers the briny environment of reflection Marine fishery eco-environment pollution level and the indices of depositional environment, introduces engineering
Supporting vector machine model in learning method is merged to the evaluation index of reflection seawater quality, depositional environment situation, is established
Complex nonlinear relation between evaluation index and fisheries environment credit rating, while the finiteness for solving sample collection is asked
Topic.
The content of the invention
The technical problem to be solved in the present invention is to provide a kind of based on fisheries environment pollution for improving SVMs etc.
Level appraisal procedure, the multiple reflection seawater qualities of methods described fusion, the index of depositional environment and improvement supporting vector machine model, carries
The quick degree and accuracy of fisheries environment class of pollution assessment high, with engineering application value higher.
The present invention is achieved by the following technical solution:
A kind of fisheries environment class of pollution appraisal procedure based on improvement SVMs, it is comprised the following steps:
A kind of fisheries environment class of pollution appraisal procedure based on improvement SVMs, it is comprised the following steps:
1) investigate and gather the index for reflecting seawater quality and depositional environment before and after marine engineering is implemented, be deposited into data
In storehouse;
2) each achievement data is pre-processed using ZCA albefactions;
3) according to《Sea water quality standard》In water grade regulation and computing formula, be index sample data addition correspondence
Water grade label;
Further, the step 3) in computing formula be AIt is comprehensive=AIt is organic+AIt is poisonous+AIt is inorganic+AOther;AIt is organic=petroleum-type+suspension
Material+floating substance+organic carbon;AIt is poisonous=(chromium+cadmium+lead+copper+mercury+arsenic)/5;AIt is inorganic=inorganic nitrogen+reactive phosphate+vulcanization
Thing;AOther=salinity+dissolved oxygen+COD+pH.
4) the achievement data matrix of the 3) tape label that step is obtained, is randomly divided into training set, test set and cross validation collection,
Optimal parameter combination C and g is searched in cross validation collection using grid search and right-angled intersection verification method, then training
Collection is input to the improvement supporting vector machine model training of setting optimized parameter, obtains class of pollution assessment models, recently enters survey
Examination collection test, obtains the accuracy of fisheries environment class of pollution assessment, demonstrates the validity of institute's extracting method.
In above-mentioned steps 1) in:The factor of influence fisheries environment pollution mainly has:The suspension that marine engineering is produced in building
Silt, and the poisonous and harmful substance brought by suspension bed sediment.Therefore water environment, deposition ring after investigation marine engineering is implemented
Border, wherein water environment survey item are salinity, pH, dissolved oxygen, floating substance, suspended material, inorganic nitrogen, reactive phosphate, change
Oxygen demand totally 8 indexs are learned, depositional environment includes sulfide, organic carbon, petroleum-type, heavy metal.Further, described inorganic nitrogen
Including nitrate, nitrite and ammonia nitrogen;Described heavy metal includes chromium, cadmium, lead, copper, mercury and arsenic.
In above-mentioned steps 2) in:Because the dimension and span of each index in sample set are differed, while considering
Independence between each index, the correlation between index should be as small as possible, can so reduce the training error of model.Cause
This uses ZCA whitening techniques preprocessed datas.
In above-mentioned steps 3) in:According to《Sea water quality standard》With《Marine sediment quality standard》In credit rating rule
Fixed and computing formula (1), is that index sample data adds corresponding water pollution grade label.
In above-mentioned steps 4) in:The data set by pretreatment, according to being randomly divided into training set and test set and intersection
Checking collection, optimal parameter combination C and g is searched using grid search and right-angled intersection verification method in cross validation collection,
Training set is input to the improvement supporting vector machine model training of setting optimized parameter, obtains class of pollution assessment models, last defeated
Enter test set test, obtain the accuracy of fisheries environment class of pollution assessment, demonstrate the validity of institute's extracting method.
The present invention solves the problems, such as that environmental index sample is small and the difficult fusion of each index, and people is set up using SVMs
Work model of mind realizes the assessment of environmental pollution grade.During supporting vector machine model is built, calculated using grid search
Method is supported the optimization of vector machine inner parameter, sets up the fisheries environment class of pollution assessment models based on Optimal Parameters, compares
With traditional fisheries environment class of pollution appraisal procedure, the present invention is with accuracy higher and intelligent.
Brief description of the drawings
Fig. 1 is based on improving SVMs to fisheries environment class of pollution appraisal procedure flow chart;
Fig. 2 SVC (Support Vector Classification) parameter selects grid optimization contour map;
Fig. 3 SVC parameters selection optimization grid graphics.
Specific embodiment
Technical scheme is described in detail with case study on implementation below in conjunction with the accompanying drawings, Fig. 1 gives ocean
The evaluation process of fisheries environment pollution level, comprises the following steps that:
Step 1:Marine fishery eco-environment pollution level analysis of Influential Factors.
Influenceing the factor of fisheries water environmental quality mainly has:The suspension bed sediment that marine engineering is produced in building, Yi Jiyou
The poisonous and harmful substance that suspension bed sediment brings.Therefore water environment, depositional environment after investigation marine engineering is implemented, wherein water environment
Survey item includes salinity, pH, dissolved oxygen, floating substance, suspended material, inorganic nitrogen, reactive phosphate, COD totally 8
Item index, depositional environment includes sulfide, organic carbon, petroleum-type, heavy metal.Further, described inorganic nitrogen include nitrate,
Nitrite and ammonia nitrogen;Described heavy metal includes chromium, cadmium, lead, copper, mercury and arsenic.
Step 2:The pretreatment of evaluation index data.
Because the dimension and span of each index in sample set are differed, while in view of the independence of each index
Property, the correlation between index should be as small as possible, can so reduce the training error of model.Therefore it is pre- using ZCA whitening techniques
Processing data.
Step 3:For index sample data adds water pollution grade label.
According to《Sea water quality standard》With《Marine sediment quality standard》In credit rating regulation and computing formula
(1) it is, that index sample data adds corresponding environmental pollution grade label.
Step 4:With reference to achievement data and supporting vector machine model, fisheries environment class of pollution evaluation model is obtained.
The data set by pretreatment, according to training set and test set and cross validation collection is randomly divided into, using grid
Search and right-angled intersection verification method search optimal parameter combination C and g in cross validation collection, and training set is input to setting
The improvement supporting vector machine model training of optimized parameter, obtains class of pollution assessment models, recently enters test set test, obtains
The accuracy of fisheries environment class of pollution assessment, carrys out the quality of assessment models.
Described step 1 is described as follows:
The factor of influence fishery marine environment pollution mainly has:The suspension such as sandstone in marine engineering construction, and by hanging
The harmful substances such as the heavy metal that mud scum sand ribbon is come.In order to reflect seawater quality, Sediment environment quality condition, structure of the present invention comprehensively
Following index system is built.It is as shown in Table 1 below:
The influence factor and index system of the fisheries environment pollution level of table 1
In the evaluation index of Marine fishery eco-environment pollution level, suspended material is one of important observation index, and it contains
The size of amount determines pollution level of the marine engineering to briny environment, and it is below sample data addition label emphasis consideration to be also
Index.Marine engineering sea fishery is caused potential impact other virulence factors include sulfide, petroleum-type, heavy metal.
Described step 2 is described as follows:
Present case uses the Monitoring Data in Bohai Offshore bay, because the dimension difference between parameters is larger, because
This is normalized to sample, by effectively have adjusted the scope of index and the difference of dimension after pretreatment, it is to avoid numerical value
The appearance of improper situation is accepted or rejected, precision of prediction is improved.Make the standard of each index to reduce the redundancy of achievement data simultaneously
It is unified so that all of property variance is consistent, uncorrelated or correlation is relatively low as far as possible between different attribute.Therefore in sample number
According to entering whitening processing to initial data before being input to SVMs.Had the following properties that through the input data after whitening processing:
1) correlation is relatively low between feature;
2) all features have identical variance.
Common whitening operation has PCA (Principal Component Analysis) albefactions and ZCA albefactions, wherein
PCA albefactions can ensure that the variance of each dimension of data is 1, and can be used for dimensionality reduction can also decorrelation;ZCA albefactions are to ensure that data are each
The variance of dimension is identical, is mainly used in decorrelation, and make the data after albefaction close to original input data as far as possible.Assuming that original
Achievement data is concentrated with n sample, and each sample has d index, and ZCA whitening process is as follows:
1) zero averaging first is carried out to original index matrix X before using albefaction, this is subtracted with per one-dimensional sample data
The average of dimension is worth to Xi, sample data value is 0-1 after normalization, so as to obtain normalization matrix A.
2) the 2nd requirement according to albefaction, is to reduce the correlation between feature first, is utilized
The corresponding sample covariance matrix ∑s of calculating matrix A, and singular value decomposition is done to covariance matrix, obtain corresponding characteristic value
λ1>λ2>…λnWith corresponding eigenvectors matrix U, wherein U=[u1,…,un], then using eigenvectors matrix U as base vector
New sample data X is obtained after conversionrot, wherein Xrot=UTX。
3) in order that postrotational each input pointer has unit variance, directly useContracted as zoom factor
Put each index Xrot,i, now obtain PCA whitening processing results
4) ZCA albefactions are that a rotation process has been done on the basis of PCA albefactions so that the data after albefaction are more nearly
Initial data, by XPCAwhite,iPremultiplication matrix U obtains the result X after initial data ZCA albefactionsZCAwhite,i=UXPCAwhite,i。
During actually whitening processing is carried out, it may appear that some eigenvalue λsiNumerically close to 0 situation, this meeting
Cause to occur being approximately equal to 0 divided by one during scalingData overflow or the unstable phenomenon of numerical value are in turn resulted in,
In order to overcome these problems, realize scaling process using regularization in actual applications, i.e., make even root and inverse before to
Characteristic value adds a constant ε for very little, finally utilizesProcessed, usual ε goes one very
Small positive number, the inventive method ε=0.1 due to there is negative in the data after albefaction, therefore uses minimax normalizing again
Change data are 0-1.
The sub-step of described step 3 is described as follows:
S3.1:The determination of Marine fishery eco-environment pollution level grade scale
The present invention evaluate Marine fishery eco-environment pollution level when, according to GB《Sea water quality standard》With《Ocean sinks
Product amount of substance standard》In regulation, show that sea fishery water pollution degree can be divided into six grades, it is as shown in Table 2 below.
The Marine fishery eco-environment comprehensive pollution indexes grade classification of table 2
According to the Marine fishery eco-environment quality evaluation computing formula that pertinent literature is proposed, AIt is comprehensive=AIt is organic+AIt is poisonous+AIt is inorganic+AOther
(1) A in formulaIt is comprehensiveIt is seawater quality comprehensive pollution indexes, AIt is organic、AIt is poisonous、AIt is inorganic、AOtherRespectively organic contamination index, toxic pollutant
Index, inorganic matter index and other monitoring indexes;AIt is organic=petroleum-type+suspended material+floating substance+organic carbon;AIt is poisonous=(chromium+
Cadmium+lead+copper+mercury+arsenic)/5;AIt is inorganic=inorganic nitrogen+reactive phosphate+sulfide;AOther=salinity+dissolved oxygen+COD+
pH.The desired value that above-mentioned formula is mentioned is standardized data.
S3.2:It is index sample addition environmental pollution grade label according to the monitoring index of each index
According to the data after standardization, corresponding formula (1) is updated to, is divided into according to complex pollution indices of environmental quality
Corresponding grade, the achievement data after which part normalization is shown in Table 3, and wherein sample sequence number 1-4 is slight pollution, 4-8
It is serious pollution.
The part sample index data of table 3
Described step 4 is described as follows:
S4.1:SVMs is mainly based upon following thought:By the Nonlinear Mapping of prior selection by input vector
High-dimensional feature space is mapped to, optimal decision function is constructed in this space.When optimal decision function is constructed, using finishing
Structure principle of minimization risk, while introducing the concept at interval.And the kernel function using former space instead of high-dimensional feature space
Dot-product operation, it is to avoid complicated calculations.Current algorithm of support vector machine mainly has classification and recurrence etc..The present invention uses non-thread
Property sorting algorithm.
If original index data are concentrated with n sample, each sample has d index, by Nonlinear MappingSample
(x1,y1),…(xn,yn) it is mapped to high-dimensional feature spaceIn this high-dimensional feature space
Middle construction hyperplaneSo Nonlinear Classification is converted into linear in high-dimensional feature space
Classification, that is, meetingUnder conditions of (ζiIt is to allow the wrong slack variable divided), minimize |
| w | | problem, optimization problem is:
ζi>=0, i=1 ..., n, c are given constant.
This optimization problem is solved using Lagrangian method:
Wherein αi,riIt is Lagrange multiplier αi≥0,ri>=0, i=1 ..., n.
According to optimal conditions:
Obtain:
(5) are substituted into (3), problem is converted into maximization expression
The dot product of SVMs not direct solution high-dimensional feature spaceBut with former space
Kernel function K (xi,xj) replace it, kernel function K (xi,xj) it is the symmetric function for meeting Mercer conditions.
So optimization problem is converted into, and maximal function W (a) under constraints, according to KT conditions, hyperplane is:
Finally grader is:F (x, w, b)=sgn [∑sSupporting vectorαiyiK(x,xi)+b=0].
S4.2:Kernel function
Selection different kernel function K (x, xi), different vector machines can be constructed, the present invention is considered from radial direction base core
Function:
Wherein γ is regulation coefficient, and σ is core width.Combined using grid search and cross validation and search out optimal parameter C
And g, wherein C=22.6274, g=0.25, parameter selection result is shown in Fig. 2 and Fig. 3.
Input training set obtains class of pollution assessment models to supporting vector machine model training is improved, and test set is input to
Class of pollution assessment models, (wherein 36 31, samples are commented for 86.12% to obtain the accuracy of fisheries environment Contamination Assessment grade
Level is correct), corresponding fisheries environment quality-monitoring report is compareed, Contamination Assessment level results are essentially identical with report content, test
The feasibility of this method is demonstrate,proved.
In summary analyze, the present invention is by merging multiple Seawater environmental quality indexs for reflecting fisheries environment pollution levels
And depositional environment quality index, and ZCA whitening pretreatments are carried out to achievement data matrix, it is input to improvement supporting vector machine model
Middle training, has obtained fisheries environment class of pollution assessment models, and the evaluation result of this method is scientific and reasonable, with engineering higher
Actual application value.
Claims (4)
1. it is a kind of based on the fisheries environment class of pollution appraisal procedure for improving SVMs, it is characterised in that it includes following step
Suddenly:
1) investigate and gather the index for reflecting seawater quality and depositional environment before and after marine engineering is implemented, be deposited into database
In;
2) each achievement data is pre-processed using ZCA albefactions;
3) according to《Sea water quality standard》In water grade regulation and computing formula, be that index sample data adds corresponding water
Matter grade label;
4) the achievement data matrix of the 3) tape label that step is obtained, is randomly divided into training set, test set and cross validation collection, uses
Grid search and right-angled intersection verification method search optimal parameter combination C and g in cross validation collection, then that training set is defeated
Enter the improvement supporting vector machine model training to setting optimized parameter, obtain class of pollution assessment models, recently enter test set
Test, obtains the accuracy of fisheries environment class of pollution assessment, demonstrates the validity of institute's extracting method.
2. according to claim 1 a kind of based on the fisheries environment class of pollution appraisal procedure for improving SVMs, its
It is characterised by the step 1) in:Water environment survey item is salinity, pH, dissolved oxygen, floating substance, suspended material, inorganic
Nitrogen, reactive phosphate, COD totally 8 indexs, depositional environment include sulfide, organic carbon, petroleum-type, heavy metal.
3. according to claim 2 a kind of based on the fisheries environment class of pollution appraisal procedure for improving SVMs, its
It is characterised by that described inorganic nitrogen includes nitrate, nitrite and ammonia nitrogen;Described heavy metal includes chromium, cadmium, lead, copper, mercury
And arsenic.
4. according to claim 1 a kind of based on the fisheries environment class of pollution appraisal procedure for improving SVMs, its
Be characterised by the step 3) in computing formula be AIt is comprehensive=AIt is organic+AIt is poisonous+AIt is inorganic+AOther;AIt is organic=petroleum-type+suspended material+drift
Float matter+organic carbon;AIt is poisonous=(chromium+cadmium+lead+copper+mercury+arsenic)/5;AIt is inorganic=inorganic nitrogen+reactive phosphate+sulfide;AOther=
Salinity+dissolved oxygen+COD+pH.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710122690.1A CN106874960A (en) | 2017-03-03 | 2017-03-03 | Based on the fisheries environment class of pollution appraisal procedure for improving SVMs |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710122690.1A CN106874960A (en) | 2017-03-03 | 2017-03-03 | Based on the fisheries environment class of pollution appraisal procedure for improving SVMs |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106874960A true CN106874960A (en) | 2017-06-20 |
Family
ID=59169578
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710122690.1A Pending CN106874960A (en) | 2017-03-03 | 2017-03-03 | Based on the fisheries environment class of pollution appraisal procedure for improving SVMs |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106874960A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107357840A (en) * | 2017-06-23 | 2017-11-17 | 广东开放大学(广东理工职业学院) | A kind of fishery big data determination method and system |
CN109190979A (en) * | 2018-09-03 | 2019-01-11 | 深圳市智物联网络有限公司 | A kind of industry internet of things data analysis method, system and relevant device |
CN111582734A (en) * | 2020-05-12 | 2020-08-25 | 上海海洋大学 | Ocean pollution comparative analysis and risk assessment intelligent method based on python crawler system and SVM |
CN112101789A (en) * | 2020-09-16 | 2020-12-18 | 清华大学合肥公共安全研究院 | Water pollution alarm grade identification method based on artificial intelligence |
CN112766739A (en) * | 2021-01-22 | 2021-05-07 | 北京工商大学 | BWM-E model-based evaluation method for heavy metal pollution in meat products |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102902985A (en) * | 2012-10-10 | 2013-01-30 | 常州大学 | Coastal water quality evaluation method based on two-classification support vector machines and particle swarm algorithm |
CN102999709A (en) * | 2012-12-20 | 2013-03-27 | 中国环境科学研究院 | Underground water grading and zoning evaluation method in agricultural activity area |
-
2017
- 2017-03-03 CN CN201710122690.1A patent/CN106874960A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102902985A (en) * | 2012-10-10 | 2013-01-30 | 常州大学 | Coastal water quality evaluation method based on two-classification support vector machines and particle swarm algorithm |
CN102999709A (en) * | 2012-12-20 | 2013-03-27 | 中国环境科学研究院 | Underground water grading and zoning evaluation method in agricultural activity area |
Non-Patent Citations (5)
Title |
---|
冯帅等: ""太湖流域上游河网污染物降解系数研究"", 《环境科学学报》 * |
李赢等: ""基于模糊聚类和完全二叉树支持向量机的变压器故障诊断"", 《电工技术学报》 * |
李雪等: ""基于BP人工神经网络的海水水质综合评价"", 《海洋通报》 * |
纪昌明等: ""基于网格搜索和交叉验证的支持向量机在梯级水电系统隐随机调度中的应用"", 《电力自动化设备》 * |
谢霖铨等: ""基于降噪自编码的推荐算法"", 《计算机与现代化》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107357840A (en) * | 2017-06-23 | 2017-11-17 | 广东开放大学(广东理工职业学院) | A kind of fishery big data determination method and system |
CN109190979A (en) * | 2018-09-03 | 2019-01-11 | 深圳市智物联网络有限公司 | A kind of industry internet of things data analysis method, system and relevant device |
CN111582734A (en) * | 2020-05-12 | 2020-08-25 | 上海海洋大学 | Ocean pollution comparative analysis and risk assessment intelligent method based on python crawler system and SVM |
CN112101789A (en) * | 2020-09-16 | 2020-12-18 | 清华大学合肥公共安全研究院 | Water pollution alarm grade identification method based on artificial intelligence |
CN112766739A (en) * | 2021-01-22 | 2021-05-07 | 北京工商大学 | BWM-E model-based evaluation method for heavy metal pollution in meat products |
CN112766739B (en) * | 2021-01-22 | 2023-08-11 | 北京工商大学 | Method for evaluating heavy metal pollution in meat product based on BWM-E model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106874960A (en) | Based on the fisheries environment class of pollution appraisal procedure for improving SVMs | |
Zhang et al. | Application of multivariate statistical techniques in the assessment of water quality in the Southwest New Territories and Kowloon, Hong Kong | |
Lang et al. | A framework for early-warning modeling with an application to banks | |
CN107341497A (en) | The unbalanced weighting data streams Ensemble classifier Forecasting Methodology of sampling is risen with reference to selectivity | |
Hasan et al. | Investigating the causal relationship between financial development and carbon emission in the emerging country | |
CN107368542A (en) | A kind of concerning security matters Classified Protection of confidential data | |
CN112785450A (en) | Soil environment quality partitioning method and system | |
Wang et al. | Source identification of mine water inrush: a discussion on the application of hydrochemical method | |
Bouguila et al. | Mml-based approach for finite dirichlet mixture estimation and selection | |
Li et al. | Predicting seabed sand content across the Australian margin using machine learning and geostatistical methods | |
Xin et al. | Research on the application of multimodal-based machine learning algorithms to water quality classification | |
Vijay et al. | Ground water quality prediction using machine learning algorithms in R | |
Malek et al. | Dissolved oxygen prediction using support vector machine | |
Schneider et al. | Unravelling the effect of flow regime on macroinvertebrates and benthic algae in regulated versus unregulated streams | |
Díaz-González et al. | Development and comparison of machine learning models for water multidimensional classification | |
Haining et al. | Design of teaching quality evaluation model based on fuzzy mathematics and SVM algorithm | |
Zhang et al. | Dbiecm-an evolving clustering method for streaming data clustering | |
Dickson et al. | An evaluation of methods for imputation of missing trace element data in groundwaters | |
Fu et al. | Development of modified integrated water quality index to assess the surface water quality: a case study of Tuo River, China | |
Huang et al. | A feature extraction method based on the entropy-minimal description length principle and GBDT for common surface water pollution identification | |
Lerche | Estimates of worldwide gas hydrate resources | |
Jun-e et al. | Gas outburst risk analysis based on pattern recognition of RSSVM model | |
Ghaffar et al. | Cyanobacteria dominance in lakes and evaluation of its predictors: A study of Southern Appalachians Ecoregion, USA | |
Zhong et al. | Positive and Inverse Degree of Grey Incidence Estimation Model of Soil Organic Matter Based on Hyper-spectral Data. | |
CN113159419A (en) | Group feature portrait analysis method, device and equipment and readable storage medium |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170620 |
|
RJ01 | Rejection of invention patent application after publication |