CN110163276A - A kind of screening technique of near infrared spectrum modeling sample - Google Patents
A kind of screening technique of near infrared spectrum modeling sample Download PDFInfo
- Publication number
- CN110163276A CN110163276A CN201910407282.XA CN201910407282A CN110163276A CN 110163276 A CN110163276 A CN 110163276A CN 201910407282 A CN201910407282 A CN 201910407282A CN 110163276 A CN110163276 A CN 110163276A
- Authority
- CN
- China
- Prior art keywords
- sample
- modeling
- near infrared
- infrared spectrum
- screening technique
- 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
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000012216 screening Methods 0.000 title claims abstract description 30
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 22
- 238000001228 spectrum Methods 0.000 claims abstract description 17
- 241000208125 Nicotiana Species 0.000 claims description 8
- 235000002637 Nicotiana tabacum Nutrition 0.000 claims description 8
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000013145 classification model Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 4
- 230000033310 detection of chemical stimulus Effects 0.000 claims description 2
- 238000002474 experimental method Methods 0.000 claims description 2
- 239000000463 material Substances 0.000 abstract description 8
- 238000005259 measurement Methods 0.000 abstract description 8
- 238000001514 detection method Methods 0.000 abstract description 3
- 239000000523 sample Substances 0.000 description 108
- 239000000126 substance Substances 0.000 description 8
- 238000004566 IR spectroscopy Methods 0.000 description 3
- 238000012937 correction Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000012497 inhomogeneous sample Substances 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000007630 basic procedure Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000001311 chemical methods and process Methods 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000007334 copolymerization reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 238000004817 gas chromatography Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000007789 sealing Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 235000019505 tobacco product Nutrition 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/245—Classification techniques relating to the decision surface
- G06F18/2451—Classification techniques relating to the decision surface linear, e.g. hyperplane
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Analytical Chemistry (AREA)
- Probability & Statistics with Applications (AREA)
- Pathology (AREA)
- Immunology (AREA)
- General Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
It is an object of the invention to provide a kind of modeling sample screening techniques based near infrared spectrum modeling, in the case where not given other information, sample are only divided into several classes by spectrum, the sample in every one kind is closer to.Wherein, close definition is to be modeled under the classification assigned at random, and close sample has biggish probability to be classified as one kind;Specifically, by the near infrared spectrum of modeling sample, measurement and modeling that several samples carry out content of material are picked out automatically.In the case where not can be carried out the detection of high-volume content of material under due to factors such as time costs, this method reduces the use of modeling sample, and not will lead to declining to a great extent for modeling accuracy.
Description
Technical field
The invention belongs to data analyses and mathematical modeling field, and in particular to a kind of modeling sample of near infrared spectrum modeling
Screening technique.
Background technique
Infrared spectroscopy is widely used in the industries such as chemical industry, food, pharmacy because its is quick, accurate and lossless.Spectrum is more
First alignment technique can be efficiently used for material component content detection and online process monitoring.With Flow Analyzer, gas-chromatography
It is different etc. traditional analytical chemistry methods, the content letter of such as substances of interest can not be directly obtained from atlas of near infrared spectra
Breath.Qualitative and quantitative analysis is carried out using near infrared spectrum and generally uses two modelings mode, i.e., is established based on one group of known sample
Calibration model.This group of known sample is known as calibration set sample or training set sample, passes through the spectrum and its correspondence of this group of sample
Data (such as in sample substances of interest content, measured by other analytical chemistry means), utilize Multivariate Correction or mode to know
Other method establishes correction or identification model.For sample to be tested, its infrared spectroscopy only need to be measured, it can be fast according to the model of foundation
Fast quantitative or qualitative results.
The basic procedure of near infrared spectrum modeling analysis includes following several steps: (1) collection of modeling sample, (2) are red
The measurement of external spectrum and reference data, (3) data processing and modeling, (4) new samples infrared spectrum measurement, (5) new samples content
Prediction.After obtaining the infrared spectroscopy of test sample, can by established calibration model to its substances of interest content into
Row prediction.
In practical applications, needing how many sample that could successfully model is a critical issue.Modeling data is concentrated every
One sample all needs acquisition spectrum and measures the content of substance to be modeled.And it uses needed for traditional chemical routes measurement content of material
Artificial, instrument, consumptive material etc. is more demanding, and enterprise is subjected to very high time and economic cost pressure.And it not yet finds at present
The effective method for reducing modeling sample quantity, covering material content range increase the empirical instructions such as differences between samples
It is only capable of providing reference, does not have actual operability.
Summary of the invention
It is an object of the invention to provide a kind of modeling sample screening techniques based near infrared spectrum modeling, are not giving other
In the case where information, sample is only divided by several classes by spectrum, the sample in every one kind is closer to.Wherein, it approaches
Definition be to be modeled under the classification assigned at random, close sample has biggish probability to be classified as one kind;Specifically, lead to
The near infrared spectrum of modeling sample is crossed, picks out measurement and modeling that several samples carry out content of material automatically.Due to the time,
In the case where not can be carried out the detection of high-volume content of material under the factors such as cost, this method reduces the use of modeling sample, and not
It will lead to declining to a great extent for modeling accuracy.
In order to achieve the above-mentioned object of the invention, the invention adopts the following technical scheme:
A kind of modeling sample screening technique based near infrared spectrum modeling, includes the following steps:
Step 1) acquires its sample spectra to K sample of offer near infrared spectrometer;
Step 2) gives its class label to K sample at random;
Step 3) repeats step 2) N1 times, in each time, carries out classification with given label to sample spectrum and builds
Mould saves the label and modeling accuracy of this time;
Step 4) remembers that accuracy is highest once secondary for pth in N1 modeling, finds out the corresponding class label of pth time;
Step 5) carries out classification model construction using the correct sample of modeling in pth time, incorrect sample as forecast sample,
Original random tags are replaced with prediction label;
Step 6) constructs the matrix of a M (K row K column), and the result of step 5) step is recorded;Wherein, if sample i
Belong to same class with sample j, M (i, j)=1, otherwise, M (i, j)=0;
Step 7) executes step 2 to step 6N2 times, adds up to the M generated each time, generates Ms;
Element in step 8) given threshold value ths1, Ms less than ths1 is equal to 0;Since the 1st sample, no to M (1 :)
Sample corresponding to element for 0 is added in the set at 1 place of sample one by one, needs to meet: snp/ (snn+snp) > ths2,
Wherein, snp is the number that it is 0 that all samples compare M (i, j) not two-by-two in set s;Snn is all two two-phases of sample in set s
Than the number that M (i, j) is 0, ths2 is set as 0.7;
Step 9) repeats this process, until each sample belongs in a certain subset;
Step 10) selects a certain proportion of sample, composition screening collection in each subset;
Step 11) carries out the detection of chemical substance to the sample that screening is concentrated, and models.
In some preferred embodiments of the present invention, the quantitative range of classification is 3-5 class in step 2);In the present invention
In specific embodiment, the quantitative range of classification is 3 classes in the step 2), and the label of each sample is one in 1,2,3.
In some preferred embodiments of the present invention, the frequency in sampling in the step 3) is not less than 100 times, this experiment
Middle N1 is set as 200 times.
In some preferred embodiments of the present invention, the modeling method in step 3) is linear discriminant analysis (LDA);From change
Amount is spectrum, and target is given label.
In some preferred embodiments of the present invention, the modeling method in the step 5) is LDA.
In some preferred embodiments of the present invention, the number of iterations N2 time in step 7) is not less than 100 times.
In some preferred embodiments of the present invention, the ths1 in the step 8) is set as 0.6*N2.
In some preferred embodiments of the present invention, the subset number in the step 9) is less than sample number;Preferably,
The subset number is about the 1/5 of sample number.
In some preferred embodiments of the present invention, the ratio in step 10) is set as the 1/3 to 1/4 of each subset;If
Sample in subset is less than 3, then all selections.
In some specific embodiments of the present invention, the sample is tobacco leaf.
Compared with prior art, the invention has the following advantages:
The present invention can content of chemical substances measurement before Screening Samples, to reduce modeling sample amount;The present invention is logical
The mode of oversampling statistics, clusters sample from the angle of spectrum;Modeling sample screening technique modeling provided by the invention
The decline degree of more all sample precision of precision is acceptable, and is much higher than and randomly selects the essence that the number sample is modeled
Degree.
Detailed description of the invention
Fig. 1 is method and step 3 provided by the invention) 200 secondary label assignment and LDA modeling result;
Label is (intuitively to show, on new label value after Fig. 2 provides in classification model construction sample original label and update for the present invention
It floats 0.2);
Fig. 3 Ms: whether sample belongs to same class (brighter pixel indicates the sample that the ranks where the pixel represent two-by-two
Similitude with higher);
Fig. 4 is this modeling result of bulk sample;
Fig. 5 is the method Screening Samples modeling result that the present invention mentions;
Fig. 6 is using random device Screening Samples modeling result.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, the technical solution below in the present invention carries out clear
Chu is fully described by, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Embodiment 1
The present embodiment carries out the mark of random classification to sample first, and (under the conditions of such as three classes, each sample random labelling is
1,2,3 one kind);Modeled after mark using spectrum and label, repeatedly after, find out the best mark of modeling effect,
The sample of same mark is closer to;It repeats the above process several times, to arbitrary sample and other samples whether close to can be with
Statistics;Finally, then by close sample one kind is merged into.
The present invention is in modeling, and as far as possible with inhomogeneous sample, in of a sort sample, sampling is chosen a small amount of
Modeling;This method can be in the case where unknown other information;Representativeness sample is analyzed, of a sort sample can
By one of those or several sample representations;Modeling should preferentially select inhomogeneous sample, to enhance the representativeness of model.
Detailed technical solution includes the following steps:
Step 1) chooses certain 94, business inventory tobacco leaf sample.By sample according to tobacco business standard " YC/T after sampling
The preparation and determination of moisture Oven Method of 31-1996 tobacco and tobacco product sample " it is prepared into Powder samples and (tobacco leaf is placed in baking oven
In, dry 4h at 40 DEG C, with milled 40 mesh of Cyclone mill (FOSS)), spectral measurement is carried out after sealing and balancing 1d;
Step 2) is to 94 tobacco leaf samples of the acquisition in step 1), random given to the spectrum of 94 tobacco leaf samples
Its class label, label are selected from 1 or 2 or 3;Then carry out linear discriminant analysis (Linear Discriminate Analysis,
LDA it) models, independent variable is spectrum, and target is given label.
Linear discriminant analysis is a kind of linear learning method, also known as Fisher linear discriminant.The thought of linear discriminant: given
Training examples collection tries to project to sample on straight line, so that the subpoint of similar sample is as close possible to abnormal sample
Subpoint be away as far as possible;When classifying to new samples, it is projected on same straight line, further according to subpoint
Position determines the classifications of new samples.
Step 3) repeats step 2 200 times, in each time, carries out LDA with given label to sample spectrum and builds
Mould, as shown in Figure 1, saving the label and modeling accuracy of this time;
Modeling accuracy herein refers to sample number/total sample number that model prediction label is consistent with setting label;
Step 4) remembers that accuracy is highest primary in 200 modelings, in the present embodiment, selects accuracy the highest 56th
It is secondary, find out the 56th time corresponding to class label;It should be noted that repeatedly random sample modeling, highest accuracy should be higher than that
55%;
Step 5) carries out classification model construction using correct sample is modeled in the 56th time, and incorrect sample is as pre- test sample
This replaces original random tags with prediction label, and Fig. 2 is shown in variation before and after label value;The purpose of this step be to hypothesis with
Machine label is modified, so that sample is voluntarily attributed to several classes that can divide, so that sample is closer in class;
Step 6) constructs the matrix of a M (K row K column), and the result of step 5) is recorded.Wherein, if sample i with
Sample j belongs to same class, M (i, j)=1, otherwise, M (i, j)=0;It is can reflect out using M (K row K column) different in sample set
Whether approached between sample;
Step 7) executes step 2 to step 6 200 times, adds up to the M generated each time, generates Ms, that is, repetition is held
The M generated every time is summed up (matrix that M is 01 compositions) to step 6 by row step 2;Using Ms characterization two-by-two sample whether
Belong to same class (see Fig. 3);
Step 8) clusters sample using Ms, and concrete operations process is as follows:
Element in given threshold value ths1, Ms less than ths1 is equal to 0;It is not 0 to Ms (1 :) since the 1st sample
Sample corresponding to element is added to one by one in the set where sample 1, snp be in set s all samples compare two-by-two M (i,
J) be 0 number;Snn is that all samples compare the number that M (i, j) is 0 two-by-two in set s.It needs to meet:
snp/(snn+snp)>ths2;Ths2 is set as 0.7;
If two sample a, b are closer to, then there is greater probability M (a, b)=1;Therefore Ms (a, b) value represents a, b
The degree of closeness of two samples;Threshold value ths1 is that value lesser in Ms is set to 0, that is, thinks these samples to keeping off;
Threshold value ths1 is related to the frequency in sampling of step 7), is set as 0.6*200=120 in commission.
Class where sample 1 is referred to as set S;According to the value of Ms, gradually will be added with sample similar in S;Its criterion
For the Similar numbers of sample are greater than certain threshold value (threshold value ths2) in the sample being newly added and former set.
Step 9) repeats this process, until each sample belongs in a certain subset.Set S is rejected in former data
Contained sample deletes corresponding ranks in Ms.By the Ms (1 :) after deletion as new enlightenment sample, expand by step 8
For set S2.Finally, so that all samples are all summed up in the point that in some set;In the present embodiment;94 sample copolymerization are 26
A subset;
Step 10) selects a certain proportion of sample, composition screening collection in each subset, and the ratio is set as each
The 1/3 to 1/4 of subset;If the sample in subset less than 3, all selects.
Step 11) is modeled using the sample selected, and modeling method is offset minimum binary;Target is total reducing sugar;Fig. 5 is this
Invent the method Screening Samples modeling result mentioned.
Embodiment 2
The present embodiment is the modeling knot of this modeling method of bulk sample using this modeling method of the bulk sample of prior art offer, Fig. 4
Fruit.
Embodiment 3
The present embodiment uses random device Screening Samples modeling result, and Fig. 6 is to model to tie using random device Screening Samples
Fruit.
Comparative example
To the full sample of embodiment 2, in the modeling of 1 Screening Samples of embodiment and 3 random screening of embodiment modeling 3 method into
Row compares;Wherein, the statistical result that stochastic modeling is 30 times;Forecast sample is 97 samples independently of modeling sample, total
Sugared value is measured by Flow Analyzer.
Table 1 lists the average forecasting error of the modeling method of embodiment 1,2 and 3.
Method | Average forecasting error |
Embodiment 2 | 0.74 |
Embodiment 1 | 0.80 |
Embodiment 3 | 0.92(std:0.17) |
From table 1 it follows that full sample effect is best, average forecasting error 0.74;If in modeling sample
In select sample at random by a certain percentage and modeled, the mean errors of 500 modelings are 0.92;Using this method, modeling
In the case that sample is only original sample 1/3, average forecasting error 0.80.I.e. under conditions of meeting error requirements, substantially
Modeling sample quantity is reduced, is saved as time cost needed for measurement modeling sample chemical content and artificial, instrument, reagent
Cost.
Claims (10)
1. a kind of modeling sample screening technique based near infrared spectrum modeling, characterized by the following steps:
Step 1) acquires its sample spectra to K sample of offer near infrared spectrometer;
Step 2) gives its class label to K sample at random;
Step 3) repeats step 2) N1 times, in each time, carries out classification model construction to sample spectrum and given label, protects
Deposit the label and modeling accuracy of this time;
Step 4) remembers that accuracy is highest once secondary for pth in N1 modeling, finds out the corresponding class label of pth time;
Step 5) carries out classification model construction using the correct sample of modeling in pth time, and incorrect sample is as forecast sample, with pre-
Mark label replace original random tags;
Step 6) constructs the matrix of a M (K row K column), and the result of step 5) step is recorded;Wherein, if sample i and sample
This j belongs to same class, M (i, j)=1, otherwise, M (i, j)=0;
Step 7) executes step 2 to step 6 N2 times, adds up to the M generated each time, generates Ms;
Element in step 8) given threshold value ths1, Ms less than ths1 is equal to 0;It is not 0 to M (1 :) since the 1st sample
Element corresponding to sample, be added in the set where sample 1 one by one, need to meet: snp/ (snn+snp) > ths2,
In, snp is the number that it is 0 that all samples compare M (i, j) not two-by-two in set s;Snn is that all samples are compared two-by-two in set s
The number that M (i, j) is 0, ths2 are set as 0.7;
Step 9) repeats this process, until each sample belongs in a certain subset;
Step 10) selects a certain proportion of sample, composition screening collection in each subset;
Step 11) carries out the detection of chemical substance to the sample that screening is concentrated, and models.
2. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: step
2) quantitative range of classification is 3-5 class in;Preferably, the quantitative range of classification is 3 classes, the mark of each sample in the step 2)
Label are one in 1,2,3.
3. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: described
Step 3) in frequency in sampling be not less than 100 times, N1 is set as 200 times in this experiment.
4. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: step
3) modeling method in is linear discriminant analysis (LDA);Independent variable is spectrum, and target is given label.
5. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: described
Step 5) in modeling method be LDA.
6. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: step
7) it is not less than 100 times the number of iterations N2 time in.
7. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: described
Step 8) in ths1 be set as 0.6*N2.
8. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: described
Step 9) in subset number be less than sample number;Preferably, the subset number is about the 1/5 of sample number.
9. the modeling sample screening technique according to claim 1 based near infrared spectrum modeling, it is characterised in that: step
10) ratio in is set as the 1/3 to 1/4 of each subset;If the sample in subset less than 3, all selects.
10. the modeling sample screening technique as claimed in any one of claims 1 to 9 wherein based near infrared spectrum modeling, special
Sign is: the sample is tobacco leaf.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910407282.XA CN110163276B (en) | 2019-05-15 | 2019-05-15 | Screening method of near infrared spectrum modeling sample |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910407282.XA CN110163276B (en) | 2019-05-15 | 2019-05-15 | Screening method of near infrared spectrum modeling sample |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110163276A true CN110163276A (en) | 2019-08-23 |
CN110163276B CN110163276B (en) | 2021-03-26 |
Family
ID=67634736
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910407282.XA Active CN110163276B (en) | 2019-05-15 | 2019-05-15 | Screening method of near infrared spectrum modeling sample |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110163276B (en) |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102798607A (en) * | 2012-08-13 | 2012-11-28 | 浙江大学 | Method for estimating soil organic carbon content by using mid-infrared spectrum technology |
CN103177180A (en) * | 2011-12-20 | 2013-06-26 | 郑芳田 | Method for screening modeling samples of prediction model |
CN104062262A (en) * | 2014-07-09 | 2014-09-24 | 中国科学院半导体研究所 | Crop seed variety authenticity identification method based on near infrared spectrum |
CN104376325A (en) * | 2014-10-30 | 2015-02-25 | 中国科学院半导体研究所 | Method for building near-infrared qualitative analysis model |
CN104990895A (en) * | 2015-07-27 | 2015-10-21 | 浙江中烟工业有限责任公司 | Near infrared spectral signal standard normal correction method based on local area |
CN105138834A (en) * | 2015-08-18 | 2015-12-09 | 浙江中烟工业有限责任公司 | Tobacco chemical value quantifying method based on near-infrared spectrum wave number K-means clustering |
CN105548079A (en) * | 2015-12-28 | 2016-05-04 | 浙江中烟工业有限责任公司 | Method for determining cut tobacco composition based on near infrared spectrum |
CN105787518A (en) * | 2016-03-17 | 2016-07-20 | 浙江中烟工业有限责任公司 | Near infrared spectrum pretreatment method based on null-space projection |
CN108181263A (en) * | 2017-12-29 | 2018-06-19 | 浙江中烟工业有限责任公司 | The extraction of tobacco leaf genius loci and method of discrimination based near infrared spectrum |
CN108760647A (en) * | 2018-03-20 | 2018-11-06 | 南京财经大学 | A kind of wheat content of molds line detecting method based on Vis/NIR technology |
CN108830253A (en) * | 2018-06-27 | 2018-11-16 | 中石化西南石油工程有限公司地质录井分公司 | Screening model method for building up, spectrum screening plant and method |
CN109374565A (en) * | 2018-09-30 | 2019-02-22 | 华东交通大学 | A kind of methanol gasoline ethanol petrol differentiates and content assaying method |
-
2019
- 2019-05-15 CN CN201910407282.XA patent/CN110163276B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103177180A (en) * | 2011-12-20 | 2013-06-26 | 郑芳田 | Method for screening modeling samples of prediction model |
CN102798607A (en) * | 2012-08-13 | 2012-11-28 | 浙江大学 | Method for estimating soil organic carbon content by using mid-infrared spectrum technology |
CN104062262A (en) * | 2014-07-09 | 2014-09-24 | 中国科学院半导体研究所 | Crop seed variety authenticity identification method based on near infrared spectrum |
CN104376325A (en) * | 2014-10-30 | 2015-02-25 | 中国科学院半导体研究所 | Method for building near-infrared qualitative analysis model |
CN104990895A (en) * | 2015-07-27 | 2015-10-21 | 浙江中烟工业有限责任公司 | Near infrared spectral signal standard normal correction method based on local area |
CN105138834A (en) * | 2015-08-18 | 2015-12-09 | 浙江中烟工业有限责任公司 | Tobacco chemical value quantifying method based on near-infrared spectrum wave number K-means clustering |
CN105548079A (en) * | 2015-12-28 | 2016-05-04 | 浙江中烟工业有限责任公司 | Method for determining cut tobacco composition based on near infrared spectrum |
CN105787518A (en) * | 2016-03-17 | 2016-07-20 | 浙江中烟工业有限责任公司 | Near infrared spectrum pretreatment method based on null-space projection |
CN108181263A (en) * | 2017-12-29 | 2018-06-19 | 浙江中烟工业有限责任公司 | The extraction of tobacco leaf genius loci and method of discrimination based near infrared spectrum |
CN108760647A (en) * | 2018-03-20 | 2018-11-06 | 南京财经大学 | A kind of wheat content of molds line detecting method based on Vis/NIR technology |
CN108830253A (en) * | 2018-06-27 | 2018-11-16 | 中石化西南石油工程有限公司地质录井分公司 | Screening model method for building up, spectrum screening plant and method |
CN109374565A (en) * | 2018-09-30 | 2019-02-22 | 华东交通大学 | A kind of methanol gasoline ethanol petrol differentiates and content assaying method |
Non-Patent Citations (4)
Title |
---|
C.FANTI等: "ISO observations of a sample of Compact Steep Spectrum and GHz Peaked Spectrum Radio Galaxies", 《ARXIV:ASTRO-PH/0005035V1》 * |
郭仡等: "面向运动想象脑-机接口 的样本筛选在线分类器更新策略研究", 《航天医学与医学工程》 * |
郭佳等: "利用均匀设计原理进行近红外建模样本的筛选", 《当代中国近红外光谱技术--全国第一届近红外光谱学术会》 * |
陈达等: "近红外光谱波长与样本筛选新方法研究", 《中国化学会第二十五届学术年会论文摘要集(下册)》 * |
Also Published As
Publication number | Publication date |
---|---|
CN110163276B (en) | 2021-03-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110687072B (en) | Calibration set and verification set selection and modeling method based on spectral similarity | |
CN109597968B (en) | SMT big data-based solder paste printing performance influence factor analysis method | |
CN101936895B (en) | Near infrared spectroscopy analysis rapid detection method of rice freshness | |
TWI759577B (en) | Device and method for avoiding false positive identification during spectroscopic quantification and relevant non-transitory computer-readable medium | |
CN106845579A (en) | Improve product quality checks the information control method and control system of correctness | |
Febbi et al. | Automated determination of poplar chip size distribution based on combined image and multivariate analyses | |
Kalogiouri et al. | Liquid chromatographic methods coupled to chemometrics: A short review to present the key workflow for the investigation of wine phenolic composition as it is affected by environmental factors | |
CN105203498A (en) | Near infrared spectrum variable selection method based on LASSO | |
CN114417954B (en) | Information processing method and system for improving food detection effect | |
CN104807777A (en) | Rapid detection method for areca-nut water content based on near infrared spectrum analysis technology | |
WO2020248961A1 (en) | Method for selecting spectral wavenumber without reference value | |
CN103278467A (en) | Rapid nondestructive high-accuracy method with for identifying abundance degree of nitrogen element in plant leaf | |
CN105223140A (en) | The method for quickly identifying of homology material | |
CN108663334B (en) | Method for searching spectral characteristic wavelength of soil nutrient based on multi-classifier fusion | |
CN115034303A (en) | Directional detection method and system for harmful substances in food | |
Syazwan et al. | Categorization of watermelon maturity level based on rind features | |
WO2023207453A1 (en) | Traditional chinese medicine ingredient analysis method and system based on spectral clustering | |
CN110163276A (en) | A kind of screening technique of near infrared spectrum modeling sample | |
CN108446213A (en) | A kind of static code mass analysis method and device | |
Doerflinger et al. | Objective assessment of dried sultana grape quality using digital image analysis | |
CN104181125A (en) | Method for rapidly determining Kol-bach value of beer malt | |
CN105954206B (en) | The measurement method and system of purple maize leaf anthocyanin content | |
CN109829513B (en) | Sequential wavelength dispersion X-ray fluorescence spectrum intelligent analysis method | |
CN114139643A (en) | Monoglyceride quality detection method and system based on machine vision | |
CN109724637B (en) | Sensor element quality evaluation method based on test data |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |