CN106198433A - Infrared spectrum method for qualitative analysis based on LM GA algorithm - Google Patents
Infrared spectrum method for qualitative analysis based on LM GA algorithm Download PDFInfo
- Publication number
- CN106198433A CN106198433A CN201610487105.3A CN201610487105A CN106198433A CN 106198433 A CN106198433 A CN 106198433A CN 201610487105 A CN201610487105 A CN 201610487105A CN 106198433 A CN106198433 A CN 106198433A
- Authority
- CN
- China
- Prior art keywords
- genetic entities
- algorithm
- gene
- data
- fitness
- 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 34
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 28
- 238000004451 qualitative analysis Methods 0.000 title claims abstract description 28
- 238000012937 correction Methods 0.000 claims abstract description 6
- 230000002068 genetic effect Effects 0.000 claims description 72
- 108090000623 proteins and genes Proteins 0.000 claims description 47
- 238000001228 spectrum Methods 0.000 claims description 24
- 238000012360 testing method Methods 0.000 claims description 10
- 230000007613 environmental effect Effects 0.000 claims description 9
- 238000004611 spectroscopical analysis Methods 0.000 claims description 9
- 238000000862 absorption spectrum Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000035772 mutation Effects 0.000 claims description 7
- GNFTZDOKVXKIBK-UHFFFAOYSA-N 3-(2-methoxyethoxy)benzohydrazide Chemical compound COCCOC1=CC=CC(C(=O)NN)=C1 GNFTZDOKVXKIBK-UHFFFAOYSA-N 0.000 claims description 3
- 108700005084 Multigene Family Proteins 0.000 claims description 3
- 239000003102 growth factor Substances 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 abstract description 11
- 238000005033 Fourier transform infrared spectroscopy Methods 0.000 abstract description 8
- 230000002745 absorbent Effects 0.000 abstract description 4
- 239000002250 absorbent Substances 0.000 abstract description 4
- 238000013528 artificial neural network Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 230000003595 spectral effect Effects 0.000 description 4
- 235000013405 beer Nutrition 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000002360 preparation method Methods 0.000 description 3
- FGUUSXIOTUKUDN-IBGZPJMESA-N C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 Chemical compound C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 FGUUSXIOTUKUDN-IBGZPJMESA-N 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 2
- 229910002056 binary alloy Inorganic materials 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012628 principal component regression Methods 0.000 description 2
- 238000011410 subtraction method Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 102100031687 Galactose mutarotase Human genes 0.000 description 1
- 206010064571 Gene mutation Diseases 0.000 description 1
- 101001066315 Homo sapiens Galactose mutarotase Proteins 0.000 description 1
- 238000002835 absorbance Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- IYRWEQXVUNLMAY-UHFFFAOYSA-N carbonyl fluoride Chemical compound FC(F)=O IYRWEQXVUNLMAY-UHFFFAOYSA-N 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000000155 melt Substances 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000009466 transformation Effects 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
-
- 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
- G01N2021/3595—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using FTIR
Abstract
The invention discloses a kind of infrared spectrum method for qualitative analysis based on LM GA algorithm, global optimum's search capability of its multi-parameter fitting ability and GA algorithm of being combined with LM algorithm carries out qualitative analysis to Fourier transform infrared spectroscopy, shows, through reality application, the absorbent components that the method can quick and precisely identify in Fourier transform infrared spectroscopy.Compared with prior art, use the absorbent components in method qualitative analysis infrared spectrum disclosed by the invention the most simple to operate, and in advance without loaded down with trivial details data prediction steps such as modeling and baseline corrections, it can automatically analyze spectrogram with real-time online, analysis result is quickly, accurately.
Description
Technical field
The present invention relates to infrared spectrum qualitative analysis field, be specifically related to a kind of infrared spectrum based on LM-GA algorithm fixed
Property analyze method.
Background technology
When Fourier transform infrared spectroscopy being carried out measured portions and analyzing, it is necessary first to the problem of solution is clearly to know
Its component absorbed of road, that is need first it to be carried out qualitative analysis.Along with Fourier transform infrared spectroscopy qualitative analysis algorithm
Development, prior art has had many algorithms Fourier transform infrared spectroscopy can be carried out qualitative analysis: such as spectrum subtraction
Method, classical least square regression (CLS), principal component regression (PCR), partial least square method (PLS), artificial neural network algorithm
(ANN), genetic algorithm based on bright rich Beer law (GA) etc., be then based on these algorithms and infrared spectrum carried out qualitative analysis
Time all there is certain defect.Such as, use spectral subtraction method to need one when realizing spectral component identification and have experience
Operator effective zoom factor is set, along with the development of algorithm, although some algorithm can realize the automatic of zoom factor
Arrange, but when spectral component becomes complexity, it is invalid that the result that the mode of zoom factor obtains is set the most automatically；And for example,
Use CLS when carrying out component identification, it is necessary first to select the absorbance data of all spectrum of investigated wave band, but all light
Modal data finally can have influence on the recognition result of this algorithm, and when using PLS modeling to carry out qualitative analysis, going out of new component
Now will reduce the accuracy of identification.
Artificial neural network algorithm quickly grows in recent years, but this algorithm needs complicated training process, and single
Artificial neural network algorithm can only carry out qualitative recognition to one-component.For comparing, heredity based on bright rich Beer law is calculated
Method can be to optimize the neural network structure that the design of LM neural network algorithm is outstanding, and therefore genetic algorithm can be to ir data
Middle multicomponent identifies simultaneously, but in the case of the spectral drift brought in instrument parameter etc. and low resolution, Fu Li
The absorption of leaf transformation infrared spectrum can deviate bright rich Beer law, thus causes the accuracy of identification poor.To this end, for component
For complicated sample, want to realize the accurate qualitative analysis of infrared spectrum, at present for the most difficult.
In prior art, Zhang Changsheng et al. Jilin University's journal of 2008 discloses one entitled " a kind of based on
Genetic algorithm and the hybrid learning scheme of LM algorithm " paper, first method disclosed in this paper is to be carried out by genetic algorithm
Coarse adjustment obtains an overall approximate solution, is then alternately trained by genetic algorithm and LM algorithm as initial value again, until
Till reaching limited precision or maximum alternately step number, although the mixing disclosing genetic algorithm and LM algorithm in this paper is calculated
Method, but in the method, LM algorithm and GA algorithm act as identical function, and both are of equal value, i.e. GA algorithm finds the overall situation
Excellent solution, LM algorithm searching locally optimal solution, specifically, in the method, LM algorithm and GA algorithm are that concurrent operation is for training
The parameter of ANN algorithm, it practice, LM algorithm is the innovatory algorithm of Gauss-Newton iterative algorithm, it is the quickest to initial value
Sense since therefore this paper be based on this point use GA algorithm and then overcome the Gaussian weighting marks algorithm dependency to initial value
If, and during actual operation use be still LM algorithm, then GA algorithm is also footy in its GALM algorithm.
It addition, applicant experiment proves that the qualitative analysis showing that algorithm structure disclosed in this paper is not particularly suited for infrared spectrum.Cause
This, how to realize infrared spectrum real-time, accurately analyze, this need research.
Summary of the invention
It is an object of the invention to provide a kind of infrared spectrum based on LM-GA algorithm simple to operate, with high accuracy
Method for qualitative analysis.
For achieving the above object, the technical solution used in the present invention is: a kind of infrared spectrum based on LM-GA algorithm is qualitative
Analysis method, a kind of infrared spectrum method for qualitative analysis based on LM-GA algorithm, its step is as follows:
1) extract the ir data of testing sample, set up Parameter File, then read Parameter File and to infrared light
The effectiveness of modal data enters and judges, if effectively, then enters step 2), otherwise terminate program output report；
Described Parameter File includes: molecular spectrum data base's number, type and routing information thereof；Pending spectroscopic data
Routing information；Analyze band class information (initial wave band, termination wave band, extension wavenumber information)；Wave number drift initial value information；Apodization
Function correction type and correction initial value information；Baseline fitting exponent number and its initial value information；Environmental information (temperature, pressure
Strong etc.)；GA parameter；LM parameter；
2) by reading the absorption spectra data of all components in molecular spectrum data base, total quantity is M.Stochastic generation N
Individual genetic entities, a length of M of each genetic entities.Gene each represent a kind of component respectively, when this position is 0, represent
There is not this component, when this position be 1 be represent existence this component.Setting up population data storehouse, population data storehouse includes three simultaneously
Individual part, current gene sets, eliminate gene sets, amorph set.Each for the N of stochastic generation genetic entities is put into and works as
Front gene；；
3) LM algorithm is used to calculate the fitness of each genetic entities；
4) if current iteration number of times >=maximum iteration time, then terminating to calculate and output report, otherwise iterations adds 1 also
Enter step 5)；
5) the effective gene individuality in population data storehouse is intersected, mutation operator, analyze select；
6) if the optimal solution of program setting occurs, then terminate program output report, otherwise return to step 4), so-called
Optimal solution that is to say that error of fitting is less than in parameter the minimum error set, and the direct usual definition that is sized of minimum error is
Can, in general, minimum error is all set in 10-6～10-10In the range of.
Use having the beneficial effects that of technique scheme generation: present invention incorporates the multi-parameter fitting ability of LM algorithm
And global optimum's search capability of GA algorithm carries out qualitative analysis to Fourier transform infrared spectroscopy, show this through reality application
Method can quick and precisely identify the absorbent components in Fourier transform infrared spectroscopy, compared with prior art, uses the present invention public
Absorbent components in the method qualitative analysis infrared spectrum opened is the most simple to operate, and in advance without modeling and baseline correction etc.
Loaded down with trivial details data prediction step, it can automatically analyze spectrogram with real-time online, and analysis result is quickly, accurately.It addition, it is long with opening
The disclosed LMGA algorithm of victory et al. is compared, and the LM algorithm related in analysis method of the present invention is for GA algorithm service, LM calculation
Method participates in the calculating of fitness in GA algorithm, and therefore the present invention is the most entirely different with existing above-mentioned algorithm structure.Need
Illustrating, the physical property of testing sample is not limited by method disclosed by the invention, in other words, for solid, melts
For shape and gas testing sample, method disclosed by the invention all can carry out qualitative analysis to its infrared spectrum.
Concrete scheme, step 3) in the calculation procedure of fitness of genetic entities be:
A) extract the gene data of a genetic entities, use binary coding to resolve gene, extract in data base corresponding
Component is analyzing the spectroscopic data of wave band；
B) multiplying factor Rampt of LM algorithm, growth factor Beta and maximum iteration time B are initialized；
C) matching measure spectrum, calculates current root-mean-square error F (k), and wherein k is current iteration number of times；
D) calculate iterative compensation, then calculate root-mean-square error F (k+1)；
E) judge whether to reach end condition: (| whether (F (k)-F (k+1) |)/F (k+1) is less than minimum error, if entering
Entering step f), otherwise enter step k), minimum error is typically set to 10-9Left and right；
If f) F (k+1) < F (k), then enter step g), otherwise enter step i)；
G) Rampt=Rampt/Beta is updated.
H) update iteration result, k=k+1, and return to step d)；
I) Rampt=Rampt*Beta is updated.
J) iteration result keeps constant, returns to step d)；
K) judging whether to be computed last genetic entities, if entering step l), otherwise returning to step a)；
L) fitness of each genetic entities is calculated: the standard deviation of the genotype of genetic entities is defined as:
In formula: StdE (Gi) represent GiThe standard deviation of individual genotype；yc(vj)、ym(vj) represent actual measure spectrum respectively
With calculating spectrum light intensity at jth wave number, n represents the some number in measure spectrum data, and j represents beam location；
The G of genetic entitiesiThe fitness of individual genotype is defined as:
In practice for the calculating of fitness, the present invention does not illustrate its Fitting Calculation method, but this area skill
Art personnel all understand on the basis of the genotype of known measure spectrum data and the genetic entities of foundation and use above-mentioned formula
And the mathematical model of the Fitting Calculation spectrum realizes the calculating of fitness of genetic entities.
Specifically, step 1) the Effective judgement step of mid-infrared light modal data is:
1., extract testing sample ir data, reject invalid environmental information, if exist effective environment letter
2. breath, then enter step, otherwise terminate program output report；
2., the effective band class information of assembly, if its band class information is effective, then enter step 3., otherwise terminate program defeated
Go out report, say, that for the infrared spectrum of the testing sample extracted, first its corresponding collection of illustrative plates is carried out preliminary sentencing
Surely determine if it is effective；
3. the absorption spectra data, extracted in data base is also changed, if it exists effective molecular absorption spectrum data letter
Breath, then enter step 2), otherwise terminate program output report.It is to say, the infrared spectrum of testing sample is carried out fixed
Property need to do the preparation of a little parameter when analyzing, temperature when the most first judging its infrared detection, atmospheric pressure,
Environmental information and the band class information etc. such as light path information, whether in effective range, if errorless, then carry out ensuing analysis work
Make, the most directly quit a program, specifically can as shown in Figure 1, infrared spectrum is actually carried out fixed by these parameter preparations
Property analyze before initial stage preparation, only meeting is effective ir data, next it is carried out qualitative analysis
The most meaningful.
Concrete, step 4) in maximum iteration time be 0.5-1 times of genetic entities length.
As further preferred version, step 5) analyze select step be:
The first step, the progeny population gene that the effective gene individuality intersection in population data storehouse, mutation operator are generated
Body and the comparison of population data storehouse, reject the reiterated genes in progeny population genetic entities individual；Second step, calculates filial generation gene
The fitness of body, then by fitness individual for the effective gene in filial generation genetic entities and population data storehouse in the lump by from greatly to
Little order arrangement, only retains the big top n genetic entities of fitness as the currently active gene in population data storehouse.So
Optimal solution is i.e. can get by being iteratively repeated operational analysis.
Additionally, it is preferred that, the first step is calculating during genetic entities fitness, the singular solution that will appear from or discrete solution
(such as some component spectra data there will be singular solution analysis wave band is linear correlation when) classifies as amorph
Body, the genetic entities that in second step, the fitness of rejecting is little classifies as the superseded genetic entities of non-candidate genetic entities, described
Amorph is individual, eliminate genetic entities is included in amorph set and superseded gene sets respectively, programme-control gene
Individual intersection, the filial generation genetic entities of mutation operator generation and the genetic entities phase in amorph set and superseded gene sets
Different.So for a new generation's effective gene in population data storehouse, can be prevented effectively from it in computing, produce amorph
Or superseded gene, and then improve the efficiency of computing.
Accompanying drawing explanation
Fig. 1 is the basic flow sheet of the methods analyst of the present invention；
Fig. 2 is fitness calculation flow chart based on the genetic entities in LM algorithm pattern 1；
Fig. 3 is the infrared spectrogram of mixed gas.
Detailed description of the invention
Below by way of 1 embodiment, technical scheme disclosed by the invention is further described:
Embodiment 1: the qualitative analysis of infrared spectrum
1) using Fourier transform infrared spectrometer to gather a mixing gas component spectrogram, spectrogram is as shown in Figure 3；
2) Parameter File is set up: determine effective analysis wave band 1900cm-1-2145cm-1, extension wave number is 10cm-1, intend
Closing wave number, drift about and arrange initial value be 0, and fitting function employing promise is paused Bill's apodizing function, and matching baseline initial value is set to 1,
0,0,0, environmental information temperature is set, pressure is currently practical environmental information；The initial population number N arranged in GA parameter is
10, maximum iteration time is 10, and gene mutation probability is 0.9, and arranging maximum iteration time B in LM parameter is 50；
3) program reads spectrum, by the initial wave band of testing sample spectrum, terminates wave band, and intensity signal reads in.
4) environmental information of checking input, errorless；
5) band class information in certificate parameter, errorless；
6) (extraction of this embodiment is that the wave band in HITRAN high-resolution spectroscopy data base is new to analytical database information
Newly), and extract in data base corresponding wave band 1900cm-1～2145cm-1Spectroscopic data, in data base, this wave band is comprised
Spectroscopic data component totally 19 kinds, as shown in table 1；
7) 10 genetic entities of stochastic generation, set up initial population data base, and the number gene in genetic entities is data
The component number of corresponding wave band in storehouse, uses binary coding by 19 kinds of performances of genetic entities in population data storehouse by i.e. 19
Type is decoded into binary system genotype, and as shown in table 2, wherein 1 expression in binary system genotype comprises this component, and 0 expression is not wrapped
Containing this component, table 1 respectively shows component (by H2O, CO2, O3..., HBr, HI, OCS ..., NO+Order) with table 2-
Genotype in 11 one_to_one corresponding the most in order, the currently active genetic entities in the initial population data base being built such that is
10, additionally the amorph set in initial population data base and the individual amount in superseded gene sets are 0；
8) fitness of each genetic entities inside the currently active gene is calculated:
A) gene data of a genetic entities inside extraction effective gene, corresponding in its binary coding and data base
The analysis of the spectroscopic data of wave band selects in step 6), 7) in complete；
B) initializing LM algorithm parameter, multiplying factor Rampt=0.01, growth factor Beta=10, due to the volume of program
Row, some Parameter Conditions arrange order can the adjustment of appropriateness, but do not affect specifically applying of the analysiss method of the present invention,
Maximum iteration time B of such as LM algorithm is in step 2) in degree, be directly set as 50, therefore Fig. 1, the meter that is given in 2
Calculating step only to facilitate understand, it does not conflicts with claimed step；
C) matching measure spectrum, calculates current root-mean-square error F (k), and wherein k is current iteration number of times；
D) calculating iterative compensation, then calculate root-mean-square error F (k+1), calculating of iterative compensation uses in LM algorithm
Conventional Calculation Method；
E) judge whether to reach end condition, i.e. (| whether (F (k)-F (k+1) |)/F (k+1) is less than minimum error, if
Entering step f), otherwise enter step k), wherein setting those skilled in the art of minimum error are easy to really based on experience value
Fixed, in actual gene to be seen individuality and individuality, the number of genotype, sets minimum error as 10 in this embodiment-6；
If f) F (k+1) < F (k), then enter step g), otherwise enter step i)；
G) Rampt=Rampt/Beta is updated.
H) update iteration result, k=k+1, and return to step d)；
I) Rampt=Rampt*Beta is updated.
J) iteration result keeps constant, returns to step d)；
K) judging whether to be computed last genetic entities, if entering step l), otherwise returning to step a)；
L) fitness of each genetic entities is calculated: the standard deviation of the genotype of genetic entities is defined as:
In formula: StdE (Gi) represent GiThe standard deviation of individual genotype；yc(vj)、ym(vj) represent actual measure spectrum respectively
With calculating spectrum light intensity at jth wave number, n represents the some number in the measure spectrum data of analysing mixed wave band, j
Represent beam location, n=2197 in this embodiment；I is 1,2 ..., 19；
The G of genetic entitiesiThe fitness of individual genotype is defined as:
9) if current iteration number of times >=10, then terminating to calculate and output report, otherwise iterations adds 1 and enters step
10)；
10) the effective gene individuality in population data storehouse is intersected, progeny population gene that mutation operator generates
Body, with the comparison of population data storehouse, rejects the reiterated genes in progeny population genetic entities individual；
According to step 8) method calculate filial generation genetic entities fitness, then by filial generation genetic entities and population data
The fitness of the effective gene individuality in storehouse by order arrangement from big to small, only retains front 10 genes that fitness is big in the lump
Individual as the currently active gene in population data storehouse, the singular solution or the discrete solution that occur during calculating are included in simultaneously
Amorph set, correspondingly includes population data storehouse or superseded gene sets in by genetic entities little for the fitness of rejecting；
11) this embodiment sets the minimum error of genetic algorithm as 10-10, its iterative computation 10 times, until GA algorithm knot
Bundle, in extraction population data storehouse, the optimal solution inside the currently active gene is as recognition result.
As shown in table 2-11 be GA algorithm iterative process in the current genetic entities of every generation and each base
Because of individual standard deviation, it can be seen that optimal solution is gene numbering 1 from the table 11 of iteration the 10th time, its genotype is:
1100100101000111001, it is decoded understanding to the genotype synopsis 1 of this genetic entities, uses said method analysis
Component out has 9 kinds of (1 representative in genotype comprises component corresponding in table 1) H2O、CO2、CO、NH3、HBr、HCN、C2H2、
PH3,NO+, and may determine that the component comprised in mixed gas is H in testing2O,CO2, CO, NH3, and other 5 kinds of components are all
The less usual ingredients of content in air, and at this, analysis wave band does not have strong INFRARED ABSORPTION, simply makees by contrast
Occur for interfering component, thus negligible, thus prove to use the infrared spectrum accuracy of said method qualitative analysis mixed gas
High.
Table 1 spectra database medium wave band 1900cm-1～2145cm-1The spectroscopic data component comprised
| H2O | CO2 | O3 | N2O | CO | CH4 | NO | NH3 | OH | HBr |
| HI | OCS | N2 | HCN | C2H2 | PH3 | COF2 | H2S | NO+ |
Table 2 iterations 1
Table 3 iterations 2
Table 4 iterations 3
Table 5 iterations 4
Table 6 iterations 5
Table 7 iterations 6
Table 8 iterations 7
Table 9 iterations 8
Table 10 iterations 9
Table 11 iterations 10
Claims (6)
1. an infrared spectrum method for qualitative analysis based on LM-GA algorithm, its step is as follows:
1) extract the ir data of testing sample, set up Parameter File, then read Parameter File and to infrared spectrum number
According to effectiveness enter judge, if effectively, then enter step 2), otherwise terminate program output report；
Described Parameter File includes: molecular spectrum data base's number, type and routing information thereof；Pending spectroscopic data path
Information；Analyze band class information；Wave number drift initial value information；Apodizing function correction type and correction initial value information；Baseline
Matching exponent number and its initial value information；Environmental information；GA parameter；LM parameter；
2) reading the absorption spectra data of all components in molecular spectrum data base, in molecular absorption spectrum data base, component is the most individual
Number for M, the then N number of genetic entities of stochastic generation, a length of M of each genetic entities, gene each represent a kind of group respectively
Point, when this position is 0, represent there is not this component, when this position be 1 be represent existence this component, set up population data storehouse simultaneously,
Population data storehouse includes three parts, current gene sets, eliminates gene sets, amorph set, N number of by stochastic generation
Genetic entities is put in current gene sets；
3) LM algorithm is used to calculate the fitness of current each genetic entities in population data storehouse；
4) if current iteration number of times >=maximum iteration time, then terminating to calculate and output report, otherwise iterations adds 1 and enters
Step 5)；
5) the effective gene individuality in population data storehouse is intersected, mutation operator, analyze select, by calculate, will eliminate
Gene put into the superseded gene sets in population data storehouse, optimum N number of gene is updated to current gene sets, by incorrect
Gene put into amorph set；
6) if the optimal solution of program setting occurs, then terminate program output report, otherwise return to step 4).
A kind of infrared spectrum method for qualitative analysis based on LM-GA algorithm the most according to claim 1, it is characterised in that:
Step 3) in the calculation procedure of fitness of genetic entities be:
A) extract the gene data of a genetic entities, use binary coding to resolve gene, extract corresponding component in data base
At the spectroscopic data analyzing wave band；
B) multiplying factor Rampt and growth factor Beta of LM algorithm are initialized；
C) matching measure spectrum, calculates current root-mean-square error F (k), and wherein k is current iteration number of times；
D) calculate iterative compensation, then calculate root-mean-square error F (k+1)；
E) judge whether to reach end condition: | ((F (k)-F (k+1))/F (k+1) | whether less than minimum error, if entering step
F), step k) is otherwise entered；
If f) F (k+1) < F (k), then enter step g), otherwise enter step i)；
G) Rampt=Rampt/Beta is updated.
H) update iteration result, k=k+1, and return to step d)；
I) Rampt=Rampt*Beta is updated.
J) iteration result keeps constant, returns to step d)；
K) judging whether to be computed last genetic entities, if entering step l), otherwise returning to step a)；
L) fitness of each genetic entities is calculated: the standard deviation of the genotype of genetic entities is defined as:
In formula: StdE (Gi) represent GiThe standard deviation of individual genotype；yc(vj)、ym(vj) represent actual measure spectrum and meter respectively
Calculating spectrum light intensity at jth wave number, n represents the some number in measure spectrum data, and j represents beam location；
The G of genetic entitiesiThe fitness of individual genotype is defined as:
A kind of infrared spectrum method for qualitative analysis based on LM-GA algorithm the most according to claim 2, it is characterised in that:
Step 1) the Effective judgement step of mid-infrared light modal data is:
1., extract testing sample ir data, reject invalid environmental information, if there is effective environmental information, then
Enter step 2., otherwise terminate program output report；
2., the effective band class information of assembly, if its band class information is effective, then enters step 3., otherwise terminate program and export report
Accuse；
3., extract the absorption spectra data in data base and change, if it exists effective molecular absorption spectrum data message, then
Enter step 2), otherwise terminate program output report.
A kind of infrared spectrum method for qualitative analysis based on LM-GA algorithm the most according to claim 3, it is characterised in that:
Step 4) in maximum iteration time be 0.5-1 times of genetic entities length.
A kind of infrared spectrum method for qualitative analysis based on LM-GA algorithm the most according to claim 3, it is characterised in that:
Step 5) analyze select step be:
The first step, by population data storehouse effective gene individuality intersect, mutation operator generate progeny population genetic entities with
The comparison of population data storehouse, rejects the reiterated genes in progeny population genetic entities individual；
Second step, calculates the fitness of filial generation genetic entities, then by the effective base in filial generation genetic entities and population data storehouse
Because of individual fitness in the lump by order arrangement from big to small, only retain the big top n genetic entities of fitness as population
The currently active gene in data base.
A kind of infrared spectrum method for qualitative analysis based on LM-GA algorithm the most according to claim 5, it is characterised in that:
The first step is during calculating genetic entities fitness, and it is individual that the singular solution that will appear from or discrete solution classify as amorph,
The genetic entities that in second step, the fitness of rejecting is little classifies as the superseded genetic entities of non-candidate genetic entities, and described is invalid
Genetic entities, superseded genetic entities are included in amorph set and superseded gene sets respectively, programme-control genetic entities
The filial generation genetic entities that intersection, mutation operator generate is different with the genetic entities in amorph set and superseded gene sets.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610487105.3A CN106198433B (en) | 2016-06-27 | 2016-06-27 | Infrared spectroscopy method for qualitative analysis based on LM-GA algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610487105.3A CN106198433B (en) | 2016-06-27 | 2016-06-27 | Infrared spectroscopy method for qualitative analysis based on LM-GA algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106198433A true CN106198433A (en) | 2016-12-07 |
| CN106198433B CN106198433B (en) | 2019-04-12 |
Family
ID=57462152
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610487105.3A Active CN106198433B (en) | 2016-06-27 | 2016-06-27 | Infrared spectroscopy method for qualitative analysis based on LM-GA algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106198433B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107064042A (en) * | 2017-03-02 | 2017-08-18 | 中国科学院合肥物质科学研究院 | The method for qualitative analysis of infrared spectrum |
| CN107862036A (en) * | 2017-11-03 | 2018-03-30 | 上海海事大学 | A kind of gantry crane loaded-up condition detection method based on genetic algorithm |
| CN110161181A (en) * | 2018-02-13 | 2019-08-23 | 中国石油化工股份有限公司 | The concentration of component recognition methods of mixed gas and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101832909A (en) * | 2010-03-12 | 2010-09-15 | 江苏大学 | Selection method for subintervals of near infrared spectral characteristics based on simulated annealing-genetic algorithm |
| CN102305772A (en) * | 2011-07-29 | 2012-01-04 | 江苏大学 | Method for screening characteristic wavelength of near infrared spectrum features based on heredity kernel partial least square method |
| CN102650982A (en) * | 2012-03-28 | 2012-08-29 | 中国航空工业集团公司北京长城计量测试技术研究所 | LM (Levenberg-Marquard) algorithm realizing method based on FPGA (Field Programmable Gate Array) |
-
2016
- 2016-06-27 CN CN201610487105.3A patent/CN106198433B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101832909A (en) * | 2010-03-12 | 2010-09-15 | 江苏大学 | Selection method for subintervals of near infrared spectral characteristics based on simulated annealing-genetic algorithm |
| CN102305772A (en) * | 2011-07-29 | 2012-01-04 | 江苏大学 | Method for screening characteristic wavelength of near infrared spectrum features based on heredity kernel partial least square method |
| CN102650982A (en) * | 2012-03-28 | 2012-08-29 | 中国航空工业集团公司北京长城计量测试技术研究所 | LM (Levenberg-Marquard) algorithm realizing method based on FPGA (Field Programmable Gate Array) |
Non-Patent Citations (2)
| Title |
|---|
| HONGTU XIE ET AL.: "Soil type recognition as improved by genetic algorithm-based variable selection using near infrared spectroscopy and partial least squares discriminant analysis", 《SCIENTIFIC REPORTS》 * |
| 张长胜 等: "一种基于遗传算法和LM算法的混合学习算法", 《吉林大学学报（理学版）》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107064042A (en) * | 2017-03-02 | 2017-08-18 | 中国科学院合肥物质科学研究院 | The method for qualitative analysis of infrared spectrum |
| CN107064042B (en) * | 2017-03-02 | 2020-01-10 | 中国科学院合肥物质科学研究院 | Qualitative analysis method of infrared spectrum |
| CN107862036A (en) * | 2017-11-03 | 2018-03-30 | 上海海事大学 | A kind of gantry crane loaded-up condition detection method based on genetic algorithm |
| CN107862036B (en) * | 2017-11-03 | 2021-03-26 | 上海海事大学 | Shore bridge load state detection method based on genetic algorithm |
| CN110161181A (en) * | 2018-02-13 | 2019-08-23 | 中国石油化工股份有限公司 | The concentration of component recognition methods of mixed gas and system |
| CN110161181B (en) * | 2018-02-13 | 2022-03-04 | 中国石油化工股份有限公司 | Component concentration identification method and system for mixed gas |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106198433B (en) | 2019-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103528990B (en) | A kind of multi-model Modeling Method of near infrared spectrum | |
| CN104677875B (en) | A kind of three-dimensional fluorescence spectrum combines the method that parallel factor differentiates different brands Chinese liquor | |
| CN110174359B (en) | Aviation hyperspectral image soil heavy metal concentration assessment method based on Gaussian process regression | |
| Ventura et al. | Artificial neural networks for estimation of kinetic analytical parameters | |
| CN109187392B (en) | Zinc liquid trace metal ion concentration prediction method based on partition modeling | |
| de Santana et al. | Comparison of PLS and SVM models for soil organic matter and particle size using vis-NIR spectral libraries | |
| CN105486655A (en) | Rapid detection method for organic matters in soil based on infrared spectroscopic intelligent identification model | |
| CN104778298A (en) | Gaussian process regression soft measurement modeling method based on EGMM (Error Gaussian Mixture Model) | |
| CN103018177A (en) | Spectrogram abnormal sample point detection method based on random sampling agree set | |
| CN106198433B (en) | Infrared spectroscopy method for qualitative analysis based on LM-GA algorithm | |
| CN108956583A (en) | Characteristic spectral line automatic selecting method for laser induced breakdown spectroscopy analysis | |
| CN104730042A (en) | Method for improving free calibration analysis precision by combining genetic algorithm with laser induced breakdown spectroscopy | |
| Tang et al. | On-line multi-component alkane mixture quantitative analysis using Fourier transform infrared spectrometer | |
| Yu et al. | Prediction of soil properties based on characteristic wavelengths with optimal spectral resolution by using Vis-NIR spectroscopy | |
| CN116818687B (en) | Soil organic carbon spectrum prediction method and device based on spectrum guide integrated learning | |
| CN107064042B (en) | Qualitative analysis method of infrared spectrum | |
| CN108827909A (en) | Soil rapid classification method based on visible and near infrared spectrum and multiple targets fusion | |
| CN107300535A (en) | The method of near-infrared quick detection organic fertilizer active constituent content | |
| CN111141809A (en) | Soil nutrient ion content detection method based on non-contact type conductivity signal | |
| CN115728290A (en) | Method, system, equipment and storage medium for detecting chromium element in soil | |
| CN113406037B (en) | Infrared spectrum online rapid identification analysis method based on sequence forward selection | |
| CN112326574B (en) | Spectrum wavelength selection method based on Bayesian classification | |
| CN112630180B (en) | Ultraviolet/visible light absorption spectrum model for detecting concentration of organophosphorus pesticide in water body | |
| CN114970698A (en) | Metering equipment operation performance prediction method based on improved LWPLS | |
| CN106950193A (en) | Based on the near infrared spectrum Variable Selection that cluster analysis is combined from weight variable |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |