CN108717497A - Imitative stichopus japonicus place of production discrimination method based on PCA-SVM - Google Patents
Imitative stichopus japonicus place of production discrimination method based on PCA-SVM Download PDFInfo
- Publication number
- CN108717497A CN108717497A CN201810499174.5A CN201810499174A CN108717497A CN 108717497 A CN108717497 A CN 108717497A CN 201810499174 A CN201810499174 A CN 201810499174A CN 108717497 A CN108717497 A CN 108717497A
- Authority
- CN
- China
- Prior art keywords
- stichopus japonicus
- sample
- production
- place
- pca
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
Abstract
The invention discloses the imitative stichopus japonicus place of production discrimination method based on PCA-SVM, the method includes：It acquires different sources and imitates stichopus japonicus sample：Measure the aliphatic acid relative amount data that different sources imitate stichopus japonicus sample：Or measure the fatty acid carbons stable isotope composition data that different sources imitate stichopus japonicus sample：Joint principal component analysis (PCA), support vector machines (SVM) and particle group optimizing (PSO) algorithm establish PCA-SVM conjunctive models, the present invention chooses the aliphatic acid with food source feature as geographical origin mark, different sources are imitated the aliphatic acid relative amount of stichopus japonicus and fatty acid carbons stable isotope composition data reduces input of the data dimension as model by principal component analysis, after extraction sample support vector machine classifier is continued to optimize using cross validation and particle swarm optimization algorithm, form conjunctive model, it is objective to realize, accurately the stichopus japonicus place of production is imitated in identification, final Production area recognition rate reaches 79.49% and 98.33%.
Description
Technical field
The invention belongs to geography symbol product field of authenticity identification, it is related to the imitative stichopus japonicus place of production based on PCA-SVM
Discrimination method utilizes gas chromatography and Stable isotope mass spectrometry to measure aliphatic acid relative amount and fatty acid carbons stabilization together
The plain composition data in position, differentiates the method for imitating the stichopus japonicus place of production.
Background technology
According to the definition of GT/T 17924-2008, geography symbol product refers to being pressed using the raw material for originating from specific region
It is produced in specific region according to traditional handicraft, it is geographical that quality, characteristic or reputation depend on its Local Geographical Indication in itself
Feature, and ratify with the product of Local Geographical Indication name nominating through audit by legal procedure.Imitative stichopus japonicus is that typical geographical sign is protected
Product is protected, there is Dalian Zhangzi Islands China, Dalian Changhai County, Yantai load island, Yantai Laizhou, Yantai Mouping, Pulandian, Wafangdian, prestige
The imitative stichopus japonicus of nearly 20 kinds of geography symbol products such as extra large Rushan.
Currently, having carried out imitative stichopus japonicus Production area recognition identification research both at home and abroad, instrument detection combines chemometrics application side
Method is main Production area recognition method, and instrument detection method mainly has inorganic elements, stable isotope, near infrared spectrum, polycyclic
Aromatic hydrocarbons etc., common metrology method include offset minimum binary, clustering, discriminant analysis, principal component analysis etc..
Liu little Fang etc. determines 7 places of production totally 26 imitative stichopus japonicus using inductivity coupled plasma mass spectrometry (ICP-MS) method
The content of AL, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Mo, Cd, Hg and the Pb of sample totally 15 kinds of inorganic elements, and pass through
Clustering and principal component analysis have carried out preliminary differentiation to the imitative stichopus japonicus of different sources, but there is sample size it is very few and
The unconspicuous problem of effect is distinguished in the part place of production, and does not establish specific place of production discriminating model, can not be actually consumer
It uses.
Zou little Bo etc. determines Dalian, Fujian, Lianyun Harbour, mountain using near-infrared spectrum technique by spectrophotometer method
The collagen content of the imitative stichopus japonicus sample in eastern 4 regional 43, the imitative thorn of different sources is established using Partial Least Squares
The problems such as ginseng differentiates model, but there is the place of production that sample is chosen is very few, and place of production discriminating model does not carry out further parameter optimization.
Invention content
To solve the above problem of the existing technology, the present invention is proposed according to aliphatic acid relative amount and fatty acid carbons
The new method in the stichopus japonicus place of production is imitated in stable isotopic composition of authigenic retrospect, can be transmitted in food chain according to the selection of food source characteristic and long-pending
Geographical sign of the tired fatty acid data as imitative stichopus japonicus, this method is based on principal component analysis (PCA), the support in machine learning
Vector machine (SVM) and particle group optimizing (PSO) algorithm establish Conjoint Analysis model, and cross validation is constantly utilized after extracting sample
Optimized model parameter reaches objective, accurate identification and imitates the stichopus japonicus place of production.
In order to achieve the above object, technical scheme is as follows：
Imitative stichopus japonicus place of production discrimination method based on PCA-SVM, the method includes：
(A) acquisition different sources imitate stichopus japonicus sample：
The annual output of the imitative stichopus japonicus of China in 2016 reaches 204359 tons, and culture zone is mainly distributed on Liaoning, Shandong, Fujian etc.
The coastal region in east China, 9 places of production for selecting yield most, are Changhai County, Zhangzi Islands China, Yantai load island, Yantai Laizhou, cigarette respectively
Platform Mouping, Pulandian Pi Kou, Wafangdian, Weihai Rushan and Fujian Xiapu sample, the sample number in each place of production is no less than 6.
(B) the aliphatic acid relative amount data that different sources imitate stichopus japonicus sample are measured：
The relative amount of variety classes aliphatic acid, each sample at least replicate analysis 3 times or more are measured, conduct is averaged
Final result；
Or, measuring the fatty acid carbons stable isotope composition data that different sources imitate stichopus japonicus sample：
The carbon stable isotope composition of variety classes aliphatic acid is measured, each sample at least replicate analysis 3 times or more is made even
Mean value is as final result.
(C) the imitative stichopus japonicus place of production discrimination method based on PCA-SVM, the method have been combined principal component analysis (PCA), have been supported
Vector machine (SVM) and particle group optimizing (PSO) algorithm, the discrimination method are as follows：
(1) significance test is carried out to aliphatic acid relative amount data or fatty acid carbons stable isotope composition data, adopted
It with one-way analysis of variance method, carries out single overall pattern base (Tukey) that confidence level is 95% and examines, and according to experimental result,
Each place of production is established respectively using SQL Server 2017 imitates stichopus japonicus fatty acid data library.
(2) use principal component analysis the PCA imitative stichopus japonicus aliphatic acid relative amount database in each place of production or each place of production from (1) imitative
Stichopus japonicus fatty acid carbons stable isotope composition data library extracts the fatty acid species for characteristic of most tracing to the source, principal component analysis
(PCA) it is most common a kind of dimension reduction method in statistics.Aliphatic acid there are many containing in imitative stichopus japonicus body, by Principal Component Analysis
The fatty acid species for characteristic of most tracing to the source are extracted, and utilize its dimensionality reduction characteristic, the sampling density for improving sample polymerize the spy that traces to the source
Property, the interference of noise and experimental error is eliminated, grader precision is improved.
(3) support vector machines (SVM) is a kind of classification and identification algorithm the most frequently used in machine learning field and best effect.
Relative to sorting algorithms such as neural networks, the optimization aim of support vector machines is structuring least risk, rather than empiric risk
Minimum makes it have outstanding generalization ability and robust property, avoids over-fitting, neural network structure selection and part most
Excellent problem.
The classification performance of support vector machines depends on the selection of kernel function, using the method for cross validation, tests linear kernel
The classification performance of function, Polynomial kernel function and gaussian radial basis function.Due to by the data tool after principal component analysis
There are apparent Clustering features, therefore the accuracy rate of each Kernel function classifier is not much different, but gaussian radial basis function passes through mould
Type is corrected with after parameter optimization, has smaller Hamming distances and extensive error.Therefore, select gaussian radial basis function as support to
The kernel function of amount machine grader.
It traces to the source model with the optimal place of production is established based on the improved particle swarm optimization algorithm of the genetic cross factor (GPSO).With
Machine be arranged gaussian radial basis function parameter σ and complexity parameter C initial value, carry out initial population scale be 50, heredity into
It is 100 to change algebraically, and autognosis factor c1 is 1.49618, and social recognition factor c2 is 1.49618, and weight factor w is 0.7529
Classifier parameters optimizing.
In genetic evolution each time, the Average Accuracy and work of each 100 cross validations of particle different K values are first calculated
For its fitness, then Population Regeneration is optimal and individual is optimal, finally carries out descending sort to particle according to fitness size, compared with
Outstanding the first half particle is directly entered evolves next time, poor later half particle successively at random with the particle in the first half into
Row crisscross inheritance can both improve the problem of convergent speed of model also avoids being absorbed in local optimum in this way.
Cross validation also referred to as recycles estimation, is most common a kind of evaluation method of prediction model performance in statistics.
Effective information as much as possible in finite data can be excavated by cross validation evaluation, and according to verification result constantly to mould
Type optimizes, and can reduce the generation with prophylaxis model over-fitting to a certain extent, improve the precision of model, stability with it is general
Change performance.
But completely different verification result will produce for different data partitioning models.Therefore, using Monte Carlo with
The method of machine sampling, the cross validation evaluation for model progress each 100 random divisions of different K values of tracing to the source the place of production, and calculate
The mean value and standard deviation of different K values model accuracy rate.
(4) blind sample detection：Each class aliphatic acid relative amount data or fatty acid carbons stable isotope data random screening
Go out 1/3 data as forecast sample, is left 2/3 data as training sample, is known using the imitative stichopus japonicus place of production for establishing best performance
Other model is predicted.
Compared with prior art, the beneficial effects of the invention are as follows：
The present invention is based on PCA-SVM conjunctive models, aliphatic acid of the selection with food source feature, will as geographical origin mark
Different sources are imitated the aliphatic acid relative amount of stichopus japonicus and fatty acid carbons stable isotope composition data and are reduced by principal component analysis
Data dimension is extracted and continues to optimize supporting vector using cross validation and particle swarm optimization algorithm after sample as the input of model
Machine grader forms conjunctive model, realizes that the stichopus japonicus place of production is imitated in objective, accurate identification, final Production area recognition rate reaches 79.49%
With 98.33%.
Description of the drawings
Fig. 1 is aliphatic acid relative amount Model Identification result in embodiment 1；
Fig. 2 is fatty acid carbons stable isotope composition model recognition result in embodiment 2.
Specific implementation mode
With reference to specific embodiment, the present invention is described further, but protection scope of the present invention is not limited in
This：
Embodiment 1 (aliphatic acid relative amount model)：
A, acquisition different sources imitate stichopus japonicus sample
54, stichopus japonicus aliphatic acid relative amount sample is imitated in acquisition, and sample includes 6, Changhai County sample；6, Zhangzi Islands China sample；
6, Fujian Xiapu sample；Pulandian skin mouth sample is respectively 6；6, Wafangdian sample；6, Weihai Rushan sample；It carries on a shoulder pole in Yantai
6, sub-island sample；6, Yantai Laizhou sample；6, Yantai Mouping sample, wherein Changhai County, Zhangzi Islands China and 3, Yantai are regional
Sample is bottom sowing culture；The sample of Pulandian Pi Kou, Wafangdian and Weihai Rushan cultivate for stable breeding；The sample in Fujian Xiapu is raft
Formula cultivates.
Except Fujian Xiapu sample is the outer at ginseng of current year, other samples be join 2 years ages at ginseng, body 15~19cm of length,
100~130g of weight.Sample is cleaned with water immediately after fishing, gutting, sandstone and calcareous ring, the ultrapure washing of body wall
Only, it is freeze-dried 48h, is milled into powdered, 80 mesh mesh screens and drying excessively.Total fat extraction is carried out to imitative stichopus japonicus with reference to Folch methods,
Then the 1% total fat of sulfuric acid-methanol solution water-bath esterification 20mg of 1mL is used, is vibrated after cooling, takes supernatant to preserve after standing.
B, the aliphatic acid relative amount data that different sources imitate stichopus japonicus sample are measured
Take the imitative 1 μ L of stichopus japonicus esterification adipic acid solution extracted, by Trace GC Ultra types gas chromatographs into
Promoting the circulation of qi phase chromatographic isolation, then GC-MS experiments are carried out by ISQ type gas chromatograph-mass spectrometers and are compared and analyzed with standard mass spectrogram,
Determine the type and its relative amount data of aliphatic acid.Each sample at least replicate analysis 3 times or more, is averaged as final
As a result, the aliphatic acid relative amount tables of data of stichopus japonicus sample is imitated in the following table 1 selected parts part.
Imitate the aliphatic acid relative amount tables of data of stichopus japonicus sample in 1 part of table：
C, the discrimination method has combined principal component analysis PCA, support vector machines and particle group optimizing PSO algorithms, tool
Body is as follows：
(1) data detection
Significance test is carried out to aliphatic acid relative amount data, using one-way analysis of variance method, carries out confidence level
It is examined for 95% single overall pattern base (Tukey), filters out 17 kinds of significant differences<0.001 fatty acid data, and according to reality
It tests as a result, establishing each place of production using SQL Server 2017 imitates stichopus japonicus aliphatic acid relative amount database.The following table 2 selected parts part
Imitative stichopus japonicus aliphatic acid relative amount database.
Imitate stichopus japonicus aliphatic acid relative amount database table in 2 part of table：
(2) data are extracted in principal component analysis (PCA)
After principal component analysis, different sources imitate stichopus japonicus aliphatic acid relative amount data and show apparent cluster
Feature.It is more than 1 principal component to contribution rate, under the premise of every class on ensureing training set all includes at least a data, successively
The model calculation that 100 random parameters are carried out to top n principal component is calculated each 100 stochastical samplings intersection of different K values and tests
The Average Accuracy of card.Therefore, input variable of preceding 6 principal components as aliphatic acid relative amount model is selected.The following table 3 is
The Average Accuracy of each 100 stochastical sampling cross validations of imitative stichopus japonicus aliphatic acid relative amount model top n principal component different K values
Table.
3 aliphatic acid relative amount model top n principal component Average Accuracy table of table：
| Principal component number | K=3/% | K=6/% | K=9/% | K=18/% | K=54/% |
| N=2 | 43.624 | 47.080 | 48.222 | 49.603 | 51.400 |
| N=3 | 65.308 | 70.907 | 72.852 | 75.367 | 76.690 |
| N=4 | 63.891 | 67.611 | 68.465 | 69.200 | 70.020 |
| N=5 | 68.276 | 72.517 | 73.518 | 74.640 | 74.930 |
| N=6 | 67.880 | 72.953 | 74.497 | 76.450 | 77.630 |
(3) support vector machines (SVM) Model Identification rate
Support vector machine classifier is designed using the machine learning procedure set under Accord.NET frames, and with based on something lost
It passes and intersects the improved particle swarm optimization algorithm of the factor (GPSO), with the Average Accuracy of each 100 cross validations of particle different K values
And as fitness, the parameter of continuous iteration optimization supporting vector machine model establishes the imitative stichopus japonicus Production area recognition mould of best performance
Type.
The Average Accuracy that the optimized parameter that is finally calculated imitates stichopus japonicus aliphatic acid relative amount model is
79.49%.The following table 4 be imitative stichopus japonicus aliphatic acid relative amount model 100 random cross validations of different K values Average Accuracy with
Standard deviation table.
4 aliphatic acid relative amount model cross validation results table of table：
| K values | Average Accuracy/% | Standard deviation |
| 3 | 70.44 | 0.098407 |
| 6 | 74.89 | 0.132049 |
| 9 | 76.50 | 0.178099 |
| 18 | 80.00 | 0.210819 |
| 54 | 86.00 | 0.346987 |
(4) blind sample detection
Go out 1/3 data per one kind random screening as forecast sample, is left 2/3 data as training sample, uses foundation
The optimal imitative stichopus japonicus aliphatic acid relative amount model gone out is predicted that recognition result difference is as shown in Figure 1.Ordinate is 0 in Fig. 1
When, it represents grader and does not provide final classification result.
Embodiment 2 (fatty acid carbons stable isotope composition model)：
A, acquisition different sources imitate stichopus japonicus sample
Acquisition imitates stichopus japonicus fatty acid carbons stable isotope and forms 55, sample, and sample includes 6, Changhai County sample；Zhangzi Islands China
6, sample；6, Fujian Xiapu sample；Pulandian skin mouth sample is respectively 7；6, Wafangdian sample；Weihai Rushan sample 6
It is a；6, Yantai load island sample；6, Yantai Laizhou sample；6, Yantai Mouping sample, wherein Changhai County, Zhangzi Islands China and Yantai 3
The sample in a area is bottom sowing culture；The sample of Pulandian Pi Kou, Wafangdian and Weihai Rushan cultivate for stable breeding；Fujian Xiapu
Sample is raft culture.
Except Fujian Xiapu sample is the outer at ginseng of current year, other samples be join 2 years ages at ginseng, body 15~19cm of length,
100~130g of weight.Sample is cleaned with water immediately after fishing, gutting, sandstone and calcareous ring, the ultrapure washing of body wall
Only, it is freeze-dried 48h, is milled into powdered, 80 mesh mesh screens and drying excessively.Total fat extraction is carried out to imitative stichopus japonicus with reference to Folch methods,
Then the 1% total fat of sulfuric acid-methanol solution water-bath esterification 20mg of 1mL is used, is vibrated after cooling, takes supernatant to preserve after standing.
B, the fatty acid carbons stable isotope composition data that different sources imitate stichopus japonicus sample is measured：
The imitative 1 μ L of stichopus japonicus esterification adipic acid solution extracted are taken, pass through Trace GC Ultra type gas phase colors by 1/10
Spectrometer carries out gas-chromatography separation, then carries out GC-MS experiments by ISQ type gas chromatograph-mass spectrometers and compared with standard mass spectrogram
Analysis, determines the type of aliphatic acid；Simultaneously by remaining 9/10 esterification aliphatic acid, stablized by Delta VAdvantage types
Isotope ratio mass spectrometer measures the carbon stable isotope composition data of corresponding type aliphatic acid.Each sample at least replicate analysis 3
More than secondary, it is averaged as final result, the fatty acid carbons stable isotope that stichopus japonicus sample is imitated in the following table 5 selected parts part forms number
According to table.
Imitate the fatty acid carbons stable isotope composition data table of stichopus japonicus sample in 5 part of table：
C, the discrimination method has combined principal component analysis PCA, support vector machines and particle group optimizing PSO algorithms, tool
Body is as follows：
(1) data detection
Significance test is carried out to fatty acid carbons stable isotope composition data, using one-way analysis of variance method, is carried out
Single overall pattern base (Tukey) that confidence level is 95% is examined, and 17 kinds of significant differences are filtered out<0.001 fatty acid data,
And according to experimental result, establishes each place of production using SQL Server 2017 and imitate stichopus japonicus fatty acid carbons stable isotope composition data
Library.Imitate stichopus japonicus fatty acid carbons stable isotope composition data library in the following table 6 selected parts part.
Imitate stichopus japonicus fatty acid carbons stable isotope composition data library table in 6 part of table：
| id | The place of production | 14_1_5 | 14_0 | 15_1_5 | 15_0 | 16_1_7 |
| 1 | Changhai County | -26.324 | -25.817 | -26.043 | -27.214 | -26.141 |
| 7 | Zhangzi Islands China | -22.447 | -21.659 | -22.938 | -24.688 | -25.411 |
| 13 | Fujian Xiapu | -22.893 | -20.716 | -21.427 | -26.417 | -25.706 |
| 19 | Pulandian | -20.312 | -19.916 | -16.395 | -21.872 | -25.041 |
| 26 | Wafangdian | -20.184 | -20.049 | -17.915 | -20.140 | -22.273 |
| 32 | Weihai Rushan | -21.511 | -22.623 | -21.780 | -21.403 | -23.580 |
(2) principal component analysis (PCA) extraction extraction data
After principal component analysis, the imitative stichopus japonicus fatty acid carbons stable isotope composition data of different sources shows bright
Aobvious cluster feature.It is more than 1 principal component to contribution rate, at least all includes the premise of a data per class on ensureing training set
Under, the model calculation of 100 random parameters is carried out to top n principal component successively, different K values are calculated and adopt at random for each 100 times
The Average Accuracy of sample cross validation.Therefore, preceding 6 principal components are selected as fatty acid carbons stable isotope composition model
Input variable.The following table 7 is imitative stichopus japonicus fatty acid carbons stable isotope composition model top n principal component different K values each 100 times at random
Sample the Average Accuracy table of cross validation.
7 fatty acid carbons stable isotope composition model top n principal component Average Accuracy table of table：
| Principal component number | K=3/% | K=6/% | K=9/% | K=18/% | K=55/% |
| N=2 | 62.987 | 68.391 | 70.127 | 72.810 | 74.790 |
| N=3 | 82.066 | 85.731 | 86.080 | 86.233 | 86.450 |
| N=4 | 91.168 | 95.069 | 95.685 | 96.657 | 96.810 |
| N=5 | 90.960 | 93.662 | 94.212 | 94.250 | 94.440 |
| N=6 | 92.912 | 96.124 | 96.833 | 97.863 | 98.220 |
| N=7 | 81.440 | 84.572 | 85.032 | 85.777 | 86.110 |
| N=8 | 83.199 | 87.514 | 88.625 | 89.930 | 90.240 |
(3) support vector machines (SVM) Model Identification rate
Support vector machine classifier is designed using the machine learning procedure set under Accord.NET frames, and with based on something lost
It passes and intersects the improved particle swarm optimization algorithm of the factor (GPSO), with the Average Accuracy of each 100 cross validations of particle different K values
And as fitness, the parameter of continuous iteration optimization supporting vector machine model establishes the imitative stichopus japonicus Production area recognition mould of best performance
Type.
The optimized parameter being finally calculated imitates the Average Accuracy of stichopus japonicus fatty acid carbons stable isotope composition model
It is 79.49%.The following table 8 is the flat of imitative stichopus japonicus fatty acid carbons stable isotope composition model 100 random cross validations of different K values
Equal accuracy rate and standard deviation table.
8 fatty acid carbons stable isotope composition model cross validation results table of table：
| K values | Average Accuracy/% | Standard deviation |
| 3 | 94.44 | 0.072436 |
| 6 | 97.78 | 0.058794 |
| 9 | 99.00 | 0.039581 |
| 18 | 100.00 | 0 |
| 55 | 100.00 | 0 |
(4) blind sample detection
Go out 1/3 data per one kind random screening as forecast sample, is left 2/3 data as training sample, uses foundation
The optimal imitative stichopus japonicus fatty acid carbons stable isotope composition model gone out is predicted that recognition result difference is as shown in Figure 2.In Fig. 2
When ordinate is 0, represents grader and do not provide final classification result.
Claims (4)
1. the imitative stichopus japonicus place of production discrimination method based on PCA-SVM, includes the following steps：
(A) acquisition different sources imitate stichopus japonicus sample；
The sample number in each place of production is no less than 6；
(B) the aliphatic acid relative amount data or fatty acid carbons for measuring the imitative stichopus japonicus sample of the different sources obtained in (A) are stablized together
The plain composition data in position：
(C) discrimination method has combined principal component analysis PCA, support vector machines and particle group optimizing PSO algorithms, specifically
It is as follows：
(1) significance test is carried out to aliphatic acid relative amount data or fatty acid carbons stable isotope composition data, using list
Analysis of variance method carries out single overall pattern base Tukey that confidence level is 95% and examines, and according to experimental result, utilizes SQL
Server 2017 establishes each place of production and imitates the imitative stichopus japonicus fatty acid carbons stabilization of stichopus japonicus aliphatic acid relative amount database or each place of production respectively
Isotopics database；
(2) each places of production from (1) principal component analysis PCA is used to imitate the imitative stichopus japonicus of stichopus japonicus aliphatic acid relative amount database or each place of production
The fatty acid species for characteristic of most tracing to the source are extracted in fatty acid carbons stable isotope composition data library；
(3) support vector machines Model Identification rate selects kernel function of the gaussian radial basis function as support vector machine classifier, profit
Support vector machine classifier is designed with the machine learning procedure set under Accord.NET frames, and with based on the genetic cross factor
Improved Particle Swarm Optimization GPSO, using the Average Accuracy of each 100 cross validations of particle different K values as its fitness,
The parameter of continuous iteration optimization supporting vector machine model, establishes the imitative stichopus japonicus Production area recognition model of best performance；
(4) blind sample detection：Each class aliphatic acid relative amount data or fatty acid carbons stable isotope data random screening go out 1/3
Data are left 2/3 data as training sample, use the imitative stichopus japonicus Production area recognition model for establishing best performance as forecast sample
It is predicted.
2. the imitative stichopus japonicus place of production discrimination method based on PCA-SVM according to claim 1, it is characterised in that：(B) it in, measures
The relative amount of variety classes aliphatic acid, each sample at least replicate analysis 3 times or more, is averaged as final result.
3. the imitative stichopus japonicus place of production discrimination method based on PCA-SVM according to claim 1, it is characterised in that：(B) it in, measures
The carbon stable isotope of variety classes aliphatic acid forms, and each sample at least replicate analysis 3 times or more is averaged as final
As a result.
4. the imitative stichopus japonicus place of production discrimination method based on PCA-SVM according to claim 1, it is characterised in that：In step (3),
It is randomly provided the initial value of gaussian radial basis function parameter σ and complexity parameter C, it is 50 to carry out initial population scale, heredity
Evolutionary generation is 100, and autognosis factor c1 is 1.49618, and social recognition factor c2 is 1.49618, and weight factor w is
0.7529 classifier parameters optimizing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810499174.5A CN108717497A (en) | 2018-05-23 | 2018-05-23 | Imitative stichopus japonicus place of production discrimination method based on PCA-SVM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810499174.5A CN108717497A (en) | 2018-05-23 | 2018-05-23 | Imitative stichopus japonicus place of production discrimination method based on PCA-SVM |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108717497A true CN108717497A (en) | 2018-10-30 |
Family
ID=63900370
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810499174.5A Pending CN108717497A (en) | 2018-05-23 | 2018-05-23 | Imitative stichopus japonicus place of production discrimination method based on PCA-SVM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108717497A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109841229A (en) * | 2019-02-24 | 2019-06-04 | 复旦大学 | A kind of Neonate Cry recognition methods based on dynamic time warping |
| CN110879263A (en) * | 2019-12-06 | 2020-03-13 | 中国水产科学研究院淡水渔业研究中心 | Analysis method for identifying different geographical groups of eriocheir sinensis |
| CN110929804A (en) * | 2019-12-03 | 2020-03-27 | 无限极（中国）有限公司 | Method, device, equipment and medium for identifying production area of cultivation product |
| CN112180006A (en) * | 2020-09-15 | 2021-01-05 | 大连海洋大学 | Method for constructing tracing of apostichopus japonicus origin |
| CN112598079A (en) * | 2020-12-31 | 2021-04-02 | 上海海洋大学 | Method for identifying cephalopod population and species |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106596453A (en) * | 2016-12-16 | 2017-04-26 | 中国水产科学研究院黄海水产研究所 | Method for discriminating wild and cultured sea bass based on near infrared spectroscopy technology |
-
2018
- 2018-05-23 CN CN201810499174.5A patent/CN108717497A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106596453A (en) * | 2016-12-16 | 2017-04-26 | 中国水产科学研究院黄海水产研究所 | Method for discriminating wild and cultured sea bass based on near infrared spectroscopy technology |
Non-Patent Citations (5)
| Title |
|---|
| XINDA ZHAO ET AL.: "Authentication of the sea cucumber (Apostichopus japonicus) using amino acids carbon stable isotope fingerprinting", 《FOOD CONTROL》 * |
| XUFENG ZHANG ET AL.: "Identification of the geographical origins of sea cucumber (Apostichopus japonicus) in northern China by using stable isotope ratios and fatty acid profiles", 《FOOD CHEMISTRY》 * |
| 刘瑀: "黄渤海刺参稳定同位素组成特征的初步研究", 《海洋环境科学》 * |
| 张旭峰: "海洋底栖动物稳定同位素和脂肪酸组成的地域性差异研究", 《中国博士学位论文全文数据库 基础科学辑》 * |
| 张进 等: "改进的基于粒子群优化的支持向量机特征选择和参数联合优化算法", 《计算机应用》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109841229A (en) * | 2019-02-24 | 2019-06-04 | 复旦大学 | A kind of Neonate Cry recognition methods based on dynamic time warping |
| CN110929804A (en) * | 2019-12-03 | 2020-03-27 | 无限极（中国）有限公司 | Method, device, equipment and medium for identifying production area of cultivation product |
| CN110929804B (en) * | 2019-12-03 | 2024-04-09 | 无限极（中国）有限公司 | Method, device, equipment and medium for identifying production area of cultivated product |
| CN110879263A (en) * | 2019-12-06 | 2020-03-13 | 中国水产科学研究院淡水渔业研究中心 | Analysis method for identifying different geographical groups of eriocheir sinensis |
| CN112180006A (en) * | 2020-09-15 | 2021-01-05 | 大连海洋大学 | Method for constructing tracing of apostichopus japonicus origin |
| CN112598079A (en) * | 2020-12-31 | 2021-04-02 | 上海海洋大学 | Method for identifying cephalopod population and species |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108717497A (en) | Imitative stichopus japonicus place of production discrimination method based on PCA-SVM | |
| Du et al. | Mapping global bamboo forest distribution using multisource remote sensing data | |
| Deng et al. | Predictive geographical authentication of green tea with protected designation of origin using a random forest model | |
| Liu et al. | Convolutional neural network for hyperspectral data analysis and effective wavelengths selection | |
| CN111402197B (en) | Detection method for colored fabric cut-parts defect area | |
| CN102023137B (en) | Method for identifying white spirits | |
| CN106529574B (en) | Image classification method based on sparse autocoder and support vector machines | |
| CN108346154A (en) | The method for building up of Lung neoplasm segmenting device based on Mask-RCNN neural networks | |
| Pérez-Magariño et al. | Comparative study of artificial neural network and multivariate methods to classify Spanish DO rose wines | |
| CN109242848A (en) | Based on OTSU and GA-BP neural network wallpaper defects detection and recognition methods | |
| CN110298450A (en) | A kind of virtual sample generation method based on production confrontation network | |
| Li et al. | SPM-IS: An auto-algorithm to acquire a mature soybean phenotype based on instance segmentation | |
| CN102183467B (en) | Modeling method for grading quality of Xinjiang red dates in near infrared range | |
| CN109657616A (en) | A kind of remote sensing image land cover pattern automatic classification method | |
| CN110132879A (en) | A kind of flue-cured tobacco note discrimination method based near infrared spectrum | |
| CN109242341A (en) | Method based on apple feedstock specifications prediction fruit juice integrated quality | |
| Ma et al. | Color analysis of leaf images of deficiencies and excess nitrogen content in soybean leaves | |
| CN105447844A (en) | New method for characteristic selection of complex multivariable data | |
| Li et al. | Excitation-emission matrix fluorescence spectroscopy combined with chemometrics methods for rapid identification and quantification of adulteration in Atractylodes macrocephala Koidz | |
| CN116740474A (en) | Remote sensing image classification method based on anchoring stripe attention mechanism | |
| Ma et al. | Spatiotemporal patterns and quantitative analysis of influencing factors of PM2. 5 and O3 pollution in the North China Plain | |
| CN113310943A (en) | Lotus root starch adulteration identification method based on machine learning | |
| CN117372710A (en) | Forest gap extraction method based on Sentinel-2MSI remote sensing image | |
| CN110956388B (en) | Method for generating yield increasing scheme of offshore oil and gas reservoir | |
| CN117391724A (en) | Apple origin tracing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181030 |