CN109785912A - A kind of factor method for quickly identifying and device for target contaminant source resolution - Google Patents
A kind of factor method for quickly identifying and device for target contaminant source resolution Download PDFInfo
- Publication number
- CN109785912A CN109785912A CN201910112915.4A CN201910112915A CN109785912A CN 109785912 A CN109785912 A CN 109785912A CN 201910112915 A CN201910112915 A CN 201910112915A CN 109785912 A CN109785912 A CN 109785912A
- Authority
- CN
- China
- Prior art keywords
- factor
- data
- target contaminant
- training
- source resolution
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of factor method for quickly identifying for target contaminant source resolution, has delivered document including target contaminant source resolution involved in collection content；From the figure and data for having delivered and having been extracted in document and being related to target contaminant source resolution factor pair and answer source, constructs training set and test set is trained study, obtain preliminary training pattern；The training pattern is verified using the cross-validation method based on artificial neural network, obtains final factor identification model；Sample to be tested is parsed, the unidentified factor is calculated；The unidentified factor is calculated using the factor identification model, parses the corresponding contamination sources of the factor.The present invention quickly knows method for distinguishing for the PMF factor by a kind of, by the corresponding source of artificial neural network deep learning algorithm automatic identification each factor, greatly improves the speed and accuracy of factor identification, while avoiding the subjectivity artificially chosen.
Description
Technical field
The present invention relates to pollution Source apportionment fields, quickly identify progress for the PMF factor in particular to one kind
The method and device of source resolution.
Background technique
Source resolution work is to formulate the essential condition of scientific and reasonable relevant policies regulation, it has also become carries out ring in China various regions
One of the core content of border prevention and cure of pollution.
Positive definite matrix factorized model (Positive Matrix Factorization, PMF), PMF is a kind of receptor
Origin analysis model is based on factorial analysis principle, emission source component spectrum is not relied on excessively, in available different time sequence
The variation of the contribution margin in source is widely used in source of atmospheric particulate matter parsing, atmosphere VOCs, pollution entering the water parsing, soil dirt
Contaminate the research such as source resolution.But since in currently used receptor model PMF method, there are very big uncertainties, i.e., in PMF
After the calculated n factor, the corresponding source of these factors how is identified, such as coal-fired source, motor vehicle source, the determination in fugitive dust source etc.
Aspect fully relies on the experience and subjective judgement of PMF operator；Second is that it should be finally several factors that PMF, which not can determine that,
It is to be combined by operator by calculating a variety of schemes because of subnumber repeatedly, determines by experience and subjective judgement and finally answer
This selects several factors.And the experience of operator is from the general of the document for reading forefathers and the chemical component for including to pollution sources
Cognition, therefore the source resolution result that different operators obtains varies with each individual, the correct result do not sought unity of standard and generally acknowledged.This
Cause the uncertainty of source resolution result very big, to influence emission reduction and the treatment decision-making of the user of ultimate source parsing result
Validity.
Artificial neural network (ANN) is a kind of information processing system constructed on the basis of simulating biological neural network.
It is a kind of non-classical numerical algorithm with powerful information storage capacity and computing capability.Deep learning has been used to now
Refer to the various machine learning models based on multitiered network structure, common deep learning model is multilayer neural network, is passed through
More complicated functional relation may be implemented in multilayered model.
In view of this, there is an urgent need to design a kind of new factor method for quickly identifying for target contaminant source resolution and
Device.
Summary of the invention
The object of the present invention is to provide a kind of speed that can be improved factor identification and accuracy for target contaminant
The factor method for quickly identifying and device of source resolution.
To achieve the above object, a kind of technical solution of the invention be to provide it is a kind of for target contaminant source resolution because
Sub- method for quickly identifying, comprising: that collects target contaminant source resolution involved in content has delivered document；Text has been delivered from described
It offers middle extraction and is related to figure and data that target contaminant source resolution factor pair answers source, construct training set and test set is instructed
Practice study, obtains preliminary training pattern；The training pattern is verified using the cross-validation method based on artificial neural network, is obtained
Obtain factor identification model finally；Sample to be tested is parsed, the unidentified factor is calculated；Use the factor identification model
The unidentified factor is calculated, the corresponding contamination sources of the factor are parsed.
Further, described to collect specifically wrapping the step of having delivered document for target contaminant source resolution involved in content
Include: that collects target contaminant source resolution involved at least 100 contents has delivered document.
Further, it is described from it is described delivered to extract in document be related to target contaminant source resolution factor pair and answer source
The step of figure and data, building training set and test set are trained study, obtain preliminary training pattern, specifically includes: will
Collected document is divided into two parts data in the ratio of 3:1, wherein 3/4ths data are as training set, a quarter
Data are trained study as test set, obtain preliminary training pattern.
Further, described that collected document is divided into two parts data in the ratio of 3:1, wherein 3/4ths number
According to the step of as training set, the data of a quarter are trained study as test set, obtain preliminary training pattern tool
Body includes: that collected document is divided into two parts data in the ratio of 3:1, wherein 3/4ths data are as training set,
The data of a quarter carry out repetition learning training, to the classification results of output, with true tag contrast conting as test set
It is bigger that error or loss function value output result differ bigger loss function value with true tag, when exporting result and true tag
Loss is zero when equal, is optimized with gradient descent method iteration undated parameter, obtains preliminary training pattern.
Further, described that the training pattern is verified using the cross-validation method based on artificial neural network, it obtains most
The step of factor identification model at end, specifically includes:
Training data is divided into k parts at random, S1,S2,…,Sk；
For each model, algorithm is executed k times, selects a S every timejCollect as verifying, and it is other as training set
Carry out training pattern, the model that training is obtained is in SjOn tested, so, an error E can be obtained every time, finally
The error obtained to k times is averaging, so that it may obtain the extensive error of model；Select the model with minimum extensive error as
Final mask, and train the model again on entire training set, to obtain final factor identification model.
To achieve the above object, another technical solution of the invention is to provide a kind of for target contaminant source resolution
The quick identification device of the factor, including collection module, the collection module is for collecting target contaminant source resolution involved in content
Delivered document；Construct module, the building module is used to deliver in document extraction from described and be related to target stains material resource
Parsing factor pair answers the figure and data in source, constructs training set and test set is trained study, obtain preliminary training mould
Type；Authentication module, the authentication module are used to verify the training pattern using the cross-validation method based on artificial neural network,
Obtain final factor identification model；Computing module, the computing module calculate unidentified for parsing sample to be tested
The factor；Parsing module, the parsing module are used to calculate the unidentified factor using the factor identification model, parse
The corresponding contamination sources of the factor out.
The invention has the following advantages: the present invention quickly knows method for distinguishing for the PMF factor by a kind of, realize big
Gas VOCs, particulate matter, water, soil etc. apply the quick identification of the factor in PMF source resolution.It is calculated by artificial neural network deep learning
The corresponding source of each factor of method automatic identification, greatly improves the speed and accuracy of factor identification, while avoiding artificial
The subjectivity of selection.
Detailed description of the invention
In order to illustrate the embodiments of the present invention more clearly or technical solution, required use in being described below to embodiment
Attached drawing be briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill of field, without creative efforts, it can also be obtained according to these attached drawings others
Attached drawing, in which:
Fig. 1 is flow diagram of the present invention for the factor method for quickly identifying of target contaminant source resolution.
Fig. 2 is flow diagram of the present invention for the quick identification device of the factor of target contaminant source resolution.
Specific embodiment
The technical scheme in the embodiments of the invention will be clearly and completely described below, it is clear that described implementation
Example is only a part of the embodiments of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, this field is common
Technical staff's all other embodiment obtained without making creative work belongs to the model that the present invention protects
It encloses.
Refering to fig. 1, a kind of factor method for quickly identifying for target contaminant source resolution of the present invention the following steps are included:
Step 1: target contaminant source resolution involved in collection content has delivered document
Specifically, all Chinese and English documents of the PMF source resolution for the related target contaminant delivered are collected, it will wherein
Factor pair answer the figure in source and data to extract, formed one group of chemistry including in the factor of each determining pollution sources at
Point title and corresponding concentration and total concentration ratio data library.
More specifically, 100 or so the documents to PM2.5 source resolution are first investigated, factor pair in document is refined and answers source
Figure and data, each piece can form a study case, i.e., one group of chemistry for including in the factor of each determining pollution sources at
Point title and corresponding concentration and total concentration ratio data, this group of data are correspondingly formed one group of study label, at subsequent
According to network actual conditions by document number be supplemented to 500-1000 it is even more.
Step 2: from it is described delivered in document extract be related to target contaminant source resolution factor pair answer source figure and
Data, construct training set and test set is trained study, obtain preliminary training pattern
Specifically, all samples collected by step 1 are divided by two parts using deep learning arithmetic programming in proportion,
The big a part of accounting is training set, and few a part forms stable calculation as test set, automatic progress repetition training study
Method and output, finding makes the smallest optimal function of loss function, obtains training pattern.
More specifically, by all samples collected by step 1,3:1 divides for two parts in proportion, wherein 3/4ths number
According to as training set, the data of a quarter are as test set, automatic to carry out repetition training study, the classification knot exported to network
Fruit, with true tag contrast conting error or loss function value, when output result is equal with true tag, loss is zero, the two
Difference is bigger, and loss function value is bigger, and the total losses on training sample is the optimization aim in supervised learning, uses gradient descent method
Iteration undated parameter eventually forms preliminary training pattern to optimize this target.
Step 3: verify the training pattern using the cross-validation method based on artificial neural network, obtain it is final because
Sub- identification model is specifically, be divided into k parts for training data at random, S1,S2,…,Sk；For each model, algorithm executes k
It is secondary, a S is selected every timejCollect as verifying, and other conduct training sets carry out training pattern, the model that training is obtained is in SjOn
It is tested, so, an error E can be obtained every time, the error finally obtained to k times is averaging, so that it may be obtained
The extensive error of model；Select the model with minimum extensive error as final mask, and on entire training set again
Training model, to obtain final factor identification model.
Step 4: sample to be tested is parsed, the unidentified factor is calculated
Specifically, with a kind of PM2.5For be illustrated, step specifically includes:
Determine principal component because of subnumber；
Factorization；
Nonnegativity restrictions factor rotation
Using flow (TSP) sampler in TH-150C intelligence and PM2.5 cutter, PM2.5 sample, sampling instrument point are acquired
It Yong not polypropylene filter and quartz filter acquisition PM2.5 sample.Wherein sampler installation polypropylene screen is for measuring inorganic elements；
Quartz filter is for measuring ion and carbon component.Chemical component detection is carried out for PM2.5 sample, specific chemical composition includes
Li、Be、Na、Mg、Ti、Ca、Fe、Ba、P、K、Sc、As、Rb、Y、Mo、Cd、Sn、Sb、Cs、La、V、Cr、Mn、Co、Ni、Cu、Zn、
Ce, Sm, W, Tl, Pb, Bi, Th, U, Na+, Mg2+, Ca2+, K+, NH4+, SO42-, Cl-, NO3-, TC, OC and EC.
PM2.5Chemical concentration matrix and uncertainty data input PMF model, calculate the unidentified factor.
Step 5: calculating the unidentified factor using the factor identification model, the corresponding pollution of the factor is parsed
Material resource
Specifically, the sample chemical compositional data that experiment is obtained is PM2.5Chemical component matrix inputs PMF model, meter
The unidentified factor is calculated, neural network model is inputted, calculates the corresponding source of recognition factor automatically, such as: 60% sulfate,
50% nitrate, 40%Cl- can be identified as coal combustion.
Referring to fig. 2, the present invention also provides a kind of quick identification devices of the factor for target contaminant source resolution, comprising:
Collection module, what the collection module was used to collect target contaminant source resolution involved in content has delivered document；Module is constructed,
The building module is used for from the figure for having delivered and having extracted in document and being related to target contaminant source resolution factor pair and answer source
And data, it constructs training set and test set is trained study, obtain preliminary training pattern；Authentication module, the verifying mould
Block is used to verify the training pattern using the cross-validation method based on artificial neural network, obtains final factor identification mould
Type；Computing module, the computing module calculate the unidentified factor for parsing sample to be tested；Parsing module, it is described
Parsing module is used to calculate the unidentified factor using the factor identification model, parses the corresponding pollutant of the factor
Source.The more detailed working method of the device sees Fig. 1 and corresponding explanation, and details are not described herein again.
The present invention also provides a kind of device with store function, this, which has, is stored with program number on the device of store function
According to realization is previously used for the factor method for quickly identifying of target contaminant source resolution, phase when the program data is executed by processor
The detailed description held inside the Pass refers to above method part, and details are not described herein.
Wherein, there is the device of store function can read for server, floppy disk drive, hard disk drive, CD-ROM for this
Take at least one of device, magneto-optic disk reader etc..
The above description is only an embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright description is applied directly or indirectly in other relevant technology necks
Domain is included within the scope of the present invention.
Claims (10)
1. a kind of factor method for quickly identifying for target contaminant source resolution characterized by comprising
That collects target contaminant source resolution involved in content has delivered document；
From the figure and data for having delivered and having been extracted in document and being related to target contaminant source resolution factor pair and answer source, building instruction
Practice collection and test set is trained study, obtains preliminary training pattern；
The training pattern is verified using the cross-validation method based on artificial neural network, obtains final factor identification model；
Sample to be tested is parsed, the unidentified factor is calculated；
The unidentified factor is calculated using the factor identification model, parses the corresponding contamination sources of the factor.
2. a kind of factor method for quickly identifying for target contaminant source resolution as described in claim 1, it is characterised in that:
Target contaminant source resolution involved in the collection content specifically includes the step of having delivered document: collecting at least 100
Target contaminant source resolution has delivered document involved in appearance.
3. a kind of factor method for quickly identifying for target contaminant source resolution as described in claim 1, it is characterised in that:
It is described from it is described delivered to extract in document be related to the figure and data that target contaminant source resolution factor pair answers source, building instruction
Practice the step of collection and test set are trained study, obtain preliminary training pattern to specifically include: by collected document by 3:1
Ratio be divided into two parts data, wherein 3/4ths data are as training set, the data of a quarter as test set into
Row training study, obtains preliminary training pattern.
4. a kind of factor method for quickly identifying for target contaminant source resolution as claimed in claim 3, it is characterised in that:
It is described that collected document is divided into two parts data in the ratio of 3:1, wherein 3/4ths data are as training set, four points
One of data the step of being trained study as test set, obtaining preliminary training pattern specifically include: will be collected
Document is divided into two parts data in the ratio of 3:1, wherein 3/4ths data are as training set, the data conduct of a quarter
Test set carries out repetition learning training, defeated with true tag contrast conting error or loss function value to the classification results of output
It is bigger to differ bigger loss function value with true tag for result out, when export result it is equal with true tag when loss be zero, use
Gradient descent method iteration undated parameter optimizes, and obtains preliminary training pattern.
5. a kind of factor method for quickly identifying for target contaminant source resolution as described in claim 1, it is characterised in that:
It is described that the training pattern is verified using the cross-validation method based on artificial neural network, obtain final factor identification model
Step specifically includes:
Training data is divided into k parts at random, S1,S2,…,Sk；
For each model, algorithm is executed k times, selects a S every timejCollect as verifying, and other is trained as training set
Model, the model that training is obtained is in SjOn tested, so, an error E can be obtained every time, finally to k times
Obtained error is averaging, so that it may obtain the extensive error of model；Select the model with minimum extensive error as final
Model, and train the model again on entire training set, to obtain final factor identification model.
6. a kind of quick identification device of factor for target contaminant source resolution characterized by comprising
Collection module, what the collection module was used to collect target contaminant source resolution involved in content has delivered document；
Construct module, the building module is used to deliver in document extraction from described and be related to target contaminant source resolution factor pair
The figure and data in source are answered, training set is constructed and test set is trained study, obtain preliminary training pattern；
Authentication module, the authentication module are used to verify the trained mould using the cross-validation method based on artificial neural network
Type obtains final factor identification model；
Computing module, the computing module calculate the unidentified factor for parsing sample to be tested；
Parsing module, the parsing module are used to calculate the unidentified factor using the factor identification model, parse
The corresponding contamination sources of the factor.
7. a kind of quick identification device of factor for target contaminant source resolution as claimed in claim 6, it is characterised in that:
Target contaminant source resolution involved in the collection content specifically includes the step of having delivered document: collecting at least 100
Target contaminant source resolution has delivered document involved in appearance.
8. a kind of quick identification device of factor for target contaminant source resolution as claimed in claim 6, it is characterised in that:
It is described from it is described delivered to extract in document be related to the figure and data that target contaminant source resolution factor pair answers source, building instruction
Practice the step of collection and test set are trained study, obtain preliminary training pattern to specifically include: by collected document by 3:1
Ratio be divided into two parts data, wherein 3/4ths data are as training set, the data of a quarter as test set into
Row training study, obtains preliminary training pattern.
9. a kind of quick identification device of factor for target contaminant source resolution as claimed in claim 8, it is characterised in that:
It is described that collected document is divided into two parts data in the ratio of 3:1, wherein 3/4ths data are as training set, four points
One of data the step of being trained study as test set, obtaining preliminary training pattern specifically include: will be collected
Document is divided into two parts data in the ratio of 3:1, wherein 3/4ths data are as training set, the data conduct of a quarter
Test set carries out repetition learning training, defeated with true tag contrast conting error or loss function value to the classification results of output
It is bigger to differ bigger loss function value with true tag for result out, when export result it is equal with true tag when loss be zero, use
Gradient descent method iteration undated parameter optimizes, and obtains preliminary training pattern.
10. a kind of quick identification device of factor for target contaminant source resolution as claimed in claim 6, feature exist
In: it is described that the training pattern is verified using the cross-validation method based on artificial neural network, obtain final factor identification mould
The step of type, specifically includes:
Training data is divided into k parts at random, S1,S2,…,Sk；
For each model, algorithm is executed k times, selects a S every timejCollect as verifying, and other is trained as training set
Model, the model that training is obtained is in SjOn tested, so, an error E can be obtained every time, finally to k times
Obtained error is averaging, so that it may obtain the extensive error of model；
It selects the model with minimum extensive error as final mask, and trains the model again on entire training set,
To obtain final factor identification model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910112915.4A CN109785912A (en) | 2019-02-13 | 2019-02-13 | A kind of factor method for quickly identifying and device for target contaminant source resolution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910112915.4A CN109785912A (en) | 2019-02-13 | 2019-02-13 | A kind of factor method for quickly identifying and device for target contaminant source resolution |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109785912A true CN109785912A (en) | 2019-05-21 |
Family
ID=66503243
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910112915.4A Pending CN109785912A (en) | 2019-02-13 | 2019-02-13 | A kind of factor method for quickly identifying and device for target contaminant source resolution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109785912A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111739588A (en) * | 2020-06-19 | 2020-10-02 | 中科三清科技有限公司 | Method and device for analyzing atmospheric pollutant source, storage medium and terminal |
| CN112614553A (en) * | 2020-12-29 | 2021-04-06 | 南开大学 | VOCs source analysis method based on chemical kinetics-receptor model fusion technology |
| CN113311081A (en) * | 2021-05-17 | 2021-08-27 | 清华大学 | Pollution source identification method and device based on three-dimensional liquid chromatography fingerprint |
| CN116662853A (en) * | 2023-05-29 | 2023-08-29 | 新禾数字科技(无锡)有限公司 | Method and system for automatically identifying analysis result of pollution source |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105158353A (en) * | 2015-07-29 | 2015-12-16 | 北华航天工业学院 | Source apportionment method for polycyclic aromatic hydrocarbon pollution in soil |
| CN107239613A (en) * | 2017-06-05 | 2017-10-10 | 南开大学 | A kind of intelligent source class recognition methods based on online data and Factor Analysis Model |
| CN107341576A (en) * | 2017-07-14 | 2017-11-10 | 河北百斛环保科技有限公司 | A kind of visual air pollution of big data is traced to the source and trend estimate method |
| CN108595414A (en) * | 2018-03-22 | 2018-09-28 | 浙江大学 | Heavy metal-polluted soil enterprise pollution source discrimination based on source remittance space variable reasoning |
| CN109145296A (en) * | 2018-08-09 | 2019-01-04 | 新华智云科技有限公司 | A kind of general word recognition method and device based on monitor model |
-
2019
- 2019-02-13 CN CN201910112915.4A patent/CN109785912A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105158353A (en) * | 2015-07-29 | 2015-12-16 | 北华航天工业学院 | Source apportionment method for polycyclic aromatic hydrocarbon pollution in soil |
| CN107239613A (en) * | 2017-06-05 | 2017-10-10 | 南开大学 | A kind of intelligent source class recognition methods based on online data and Factor Analysis Model |
| CN107341576A (en) * | 2017-07-14 | 2017-11-10 | 河北百斛环保科技有限公司 | A kind of visual air pollution of big data is traced to the source and trend estimate method |
| CN108595414A (en) * | 2018-03-22 | 2018-09-28 | 浙江大学 | Heavy metal-polluted soil enterprise pollution source discrimination based on source remittance space variable reasoning |
| CN109145296A (en) * | 2018-08-09 | 2019-01-04 | 新华智云科技有限公司 | A kind of general word recognition method and device based on monitor model |
Non-Patent Citations (1)
| Title |
|---|
| 郭斯羽: "神经网络设计方法与实例分析", 北京邮电大学出版社, pages: 158 - 159 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111739588A (en) * | 2020-06-19 | 2020-10-02 | 中科三清科技有限公司 | Method and device for analyzing atmospheric pollutant source, storage medium and terminal |
| CN111739588B (en) * | 2020-06-19 | 2021-01-01 | 中科三清科技有限公司 | Method and device for analyzing atmospheric pollutant source, storage medium and terminal |
| CN112614553A (en) * | 2020-12-29 | 2021-04-06 | 南开大学 | VOCs source analysis method based on chemical kinetics-receptor model fusion technology |
| CN112614553B (en) * | 2020-12-29 | 2022-07-26 | 南开大学 | VOCs source analysis method based on chemical kinetics-receptor model fusion technology |
| CN113311081A (en) * | 2021-05-17 | 2021-08-27 | 清华大学 | Pollution source identification method and device based on three-dimensional liquid chromatography fingerprint |
| CN113311081B (en) * | 2021-05-17 | 2023-08-11 | 清华大学 | Pollution source identification method and device based on three-dimensional liquid chromatography fingerprint |
| CN116662853A (en) * | 2023-05-29 | 2023-08-29 | 新禾数字科技(无锡)有限公司 | Method and system for automatically identifying analysis result of pollution source |
| CN116662853B (en) * | 2023-05-29 | 2024-04-30 | 新禾数字科技(无锡)有限公司 | Method and system for automatically identifying analysis result of pollution source |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109785912A (en) | A kind of factor method for quickly identifying and device for target contaminant source resolution | |
| Lepš et al. | Multivariate analysis of ecological data using CANOCO | |
| Walker et al. | SIMPLE An inductive modelling and mapping tool for spatially-oriented data | |
| Pal | Mining educational data to reduce dropout rates of engineering students | |
| Lambert et al. | The use of statistics in phytosociology | |
| Wille | Conceptual landscapes of knowledge: a pragmatic paradigm for knowledge processing | |
| CN109492549A (en) | A kind of processing of training sample set, model training method and system | |
| CN105631203A (en) | Method for recognizing heavy metal pollution source in soil | |
| Mentis | Hypothetico-deductive and inductive approaches in ecology | |
| CN106650020A (en) | Analysis method of complex receptor model pollution source | |
| CN110378206A (en) | A kind of intelligence Audit System and method | |
| CN104331361B (en) | A kind of test device and method for white-box testing coverage rate calculation visualization | |
| Anthycamurty et al. | Analysis of problem solving in terms of cognitive style | |
| CN106503050B (en) | Method and system for recommending reading articles based on big data | |
| Bishop et al. | Another look at the statistician's role in experimental planning and design | |
| Richter et al. | Are Neural Bug Detectors Comparable to Software Developers on Variable Misuse Bugs? | |
| Mahiny et al. | Towards improvement of environmental impact assessment methods-a case study in Golestan Province, Iran | |
| Peltola | IT-time for mechanized forest work study | |
| Pirnay | Data-Driven Strategy Maps: A Hybrid Approach to Strategic and Performance Management Combining Hard Data and Experts' Knowledge. | |
| Ajitha et al. | Prediction of air quality based on supervised learning | |
| CN107944003A (en) | A kind of classic poetry is collected and data analysing method | |
| Raco et al. | GEOBASI: The geochemical Database of Tuscany Region (Italy) | |
| Rassbach et al. | Providing decision support for cosmogenic isotope dating | |
| Lyman | Diversity Metrics Are Convenient, but Their Archaeological Meanings Are Still Obscure | |
| Comber et al. | Lessons from spatial transcriptomics and computational geography in mapping the transcriptome |
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 |