CN104819955A - Method for detecting ligusticum wallichii on basis of particle swarm least square support vector machine algorithm and application - Google Patents
Method for detecting ligusticum wallichii on basis of particle swarm least square support vector machine algorithm and application Download PDFInfo
- Publication number
- CN104819955A CN104819955A CN201510249332.8A CN201510249332A CN104819955A CN 104819955 A CN104819955 A CN 104819955A CN 201510249332 A CN201510249332 A CN 201510249332A CN 104819955 A CN104819955 A CN 104819955A
- Authority
- CN
- China
- Prior art keywords
- chuanxiong hort
- ligusticum
- ligusticum chuanxiong
- vector machine
- moisture
- 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
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention belongs to the technical field of medicine detection, and particularly relates to a method for detecting ligusticum wallichii on basis of a particle swarm least square support vector machine algorithm and an application. The method comprises the following steps: (1) selecting a ligusticum wallichii medicinal material with known content of moisture, extract and fumalic acid, crushing the ligusticum wallichii medicinal material, and sieving the crushed ligusticum wallichii medicinal material for later use; (2) acquiring near-infrared spectrum data of the ligusticum wallichii medicinal material powder, and selecting a characteristic spectrum; (3) establishing a quantitative calibration model of the characteristic spectrum and the content of moisture, extract and fumalic acid of the ligusticum wallichii by adopting a least square support vector machine algorithm of the particle swarm algorithm optimization; (4) scanning an unknown ligusticum wallichii sample by adopting the near-infrared spectrum according to the method of the step (2), selecting the characteristic spectrum, and introducing the characteristic spectrum into the established quantitative calibration model, thereby obtaining the content of moisture, extract or fumalic acid of the unknown ligusticum wallichii sample. By adopting the method, the ligusticum wallichii medicinal material can be rapidly and comprehensively detected; moreover, the method has the advantages of simple operation, high accuracy and high precision.
Description
Technical field
The invention belongs to medicinal material detection technique field, be specifically related to a kind of method and application of the detection Ligusticum wallichii based on population least square method supporting vector machine algorithm.
Background technology
Ligusticum wallichii is the dry rhizome of samphire Ligusticum wallichii Ligusticum chuanxiong Hort., it is the alkaloids of representative that its active component has with ligustrazine, taking forulic acid as the organic acid of representative, take Ligustilide as the phthalide-type of representative, has higher health care and medical value.Blood-activating and qi-promoting is had, the function of wind-expelling pain-stopping according to recording Ligusticum wallichii in China's 2010 editions pharmacopeia, for chest impediment and cardialgia, the shouting pain of the chest side of body, tumbling and swelling, irregular menstruation, through closing dysmenorrhoea, headache, arthralgia due to wind-dampness.Because the chemical composition in Ligusticum chuanxiong Hort is comparatively complicated, and again because the factors such as kind, source, the place of production, growth year and processing mode all can affect the quality of Ligusticum chuanxiong Hort, if using Ligusticum wallichii as a kind of bulk drug in preparation, need the quality evaluation index setting up Ligusticum chuanxiong Hort, to ensure quality and effect of said preparation.But traditional Ligusticum chuanxiong Hort quality determining method is time-consuming, effort, be difficult to be widely used in production practices, the needs of the modernization development of Chinese medicine preparation can not be adapted to, and often adopt single index components to control the quality of medicinal material, the complicacy of traditional Chinese medicine ingredients can not be embodied, therefore in the urgent need to a kind of method that fast, comprehensively can detect Ligusticum chuanxiong Hort, large batch of Ligusticum chuanxiong Hort screened and the requirement of comprehensive control of quality to meet.
Near infrared spectrum (Near Infrared Spectrum Instrument, NIRS) be between visible ray (Vis) and between infrared (MIR), wavelength coverage is the electromagnetic radiation as waves of 780nm-2526nm (12820cm-1-3598cm-1), is first non-visible light district that people find in absorption spectrum.Hydric group (O-H near infrared spectrum district and organic molecule, N-H, C-H) sum of fundamental frequencies vibrated is consistent with the uptake zone of frequency multiplication at different levels, by the near infrared spectrum of scanning samples, the characteristic information of organic molecule hydric group in sample can be obtained, and utilize near-infrared spectrum technique analysis sample to have conveniently, fast, efficiently, accurately and cost lower, do not destroy sample, do not consume chemical reagent, the advantage such as free from environmental pollution, therefore this technology is subject to the favor of more and more people, successively for the assay of effective component, the on-line checkingi of pharmacy procedure and monitoring, the fields such as the place of production discriminating of natural drug discriminating and Chinese crude drug.Therefore, the invention provides a kind of utilize near infrared spectrum to detect Ligusticum wallichii fast method and application, quality near-infrared spectrum technique being applied to Ligusticum chuanxiong Hort detects fast, the source produced from Ligusticum wallichii preparation controls its quality, thus the security of guarantee end product quality, stability and validity, reach the object of quick, efficient quality control.
Quantitative calibration model must be set up when using near infrared to carry out quantitative test.Quantitative Analysis Model relatively more conventional at present mainly comprises partial least-squares regression method (PLSR), artificial neural network method (ANN) and support vector machine method (SVM) etc.SVM is a kind of machine learning algorithm be based upon on Statistical Learning Theory basis, is used widely in near-infrared spectrum analysis.Least square method supporting vector machine (LS-SVM) is structure based risk minimization principle, it is the one improvement of classical SVM, retrain the inequality constrain in replacement standard SVM algorithm in equation, and using the empirical loss of error sum of squares loss function as training set, quadratic programming problem will be solved be converted into and solve system of linear equations problem, improve the speed and convergence precision that solve.Larger impact can be produced on result for choosing of LS-SVM parameter, parameter choose the problem with certain " blindness ", adopt parameter space exhaustive search method to optimize, time-consuming and degree of accuracy is not high.Therefore, the present invention proposes one and detect Ligusticum wallichii method and application based on population least square method supporting vector machine algorithm.
Summary of the invention
For this reason, technical matters to be solved by this invention there are provided a kind of based on population least square method supporting vector machine algorithm detection Ligusticum wallichii method and application.
For solving the problems of the technologies described above, the invention provides a kind of based on population least square method supporting vector machine algorithm detection Ligusticum wallichii method, comprising at least one item in following determination of moisture, extract content mensuration and ferulaic acid content determination step:
The mensuration of A, moisture comprises the steps:
(1) choose the Ligusticum chuanxiong Hort of known moisture levels, pulverize and sieve, obtain described Ligusticum chuanxiong Hort powder, for subsequent use;
(2) the Ligusticum chuanxiong Hort powder after above-mentioned process is carried out near infrared spectrum scanning, gather the near infrared spectrum of described Ligusticum chuanxiong Hort, choose 4500-7500cm
-1spectral band as characteristic spectrum;
(3) particle cluster algorithm is adopted to calculate the value of parameter σ in least square method supporting vector machine algorithm and γ, the quantitative calibration model between the moisture setting up described characteristic spectrum and described Ligusticum wallichii with least square method supporting vector machine algorithm;
(4) according to the method for described step (2), near infrared spectrum scanning is carried out to unknown Ligusticum chuanxiong Hort sample, and choose 4500-7500cm
-1spectral information under characteristic wave bands, imports the moisture content value that the quantitative calibration model set up obtains described unknown Ligusticum chuanxiong Hort sample;
The mensuration of B, extract content comprises the steps:
(1) choose the Ligusticum chuanxiong Hort of known extract content, pulverize and sieve, obtain described Ligusticum chuanxiong Hort powder, for subsequent use;
(2) the Ligusticum chuanxiong Hort powder after above-mentioned process is carried out near infrared spectrum scanning, gather the near infrared spectrum of described Ligusticum chuanxiong Hort, choose 4500-7500cm
-1spectral band as characteristic spectrum;
(3) particle cluster algorithm is adopted to calculate the value of parameter σ in least square method supporting vector machine algorithm and γ, the quantitative calibration model between the extract content setting up described characteristic spectrum and described Ligusticum wallichii with least square method supporting vector machine algorithm;
(4) according to the method for described step (2), near infrared spectrum scanning is carried out to unknown Ligusticum chuanxiong Hort sample, and choose 4500-7500cm
-1spectral information under characteristic wave bands, imports the extract content value that the quantitative calibration model set up obtains described unknown Ligusticum chuanxiong Hort sample;
The mensuration of C, ferulaic acid content comprises the steps:
(1) choose the Ligusticum chuanxiong Hort of known ferulaic acid content, pulverize and sieve, obtain described Ligusticum chuanxiong Hort powder, for subsequent use;
(2) the Ligusticum chuanxiong Hort powder after above-mentioned process is carried out near infrared spectrum scanning, gather the near infrared spectrum of described Ligusticum chuanxiong Hort, choose 4500-7500cm
-1spectral band as characteristic spectrum;
(3) particle cluster algorithm is adopted to calculate the value of parameter σ in least square method supporting vector machine algorithm and γ, the quantitative calibration model between the ferulaic acid content setting up described characteristic spectrum and described Ligusticum wallichii with least square method supporting vector machine algorithm;
(4) according to the method for described step (2), near infrared spectrum scanning is carried out to unknown Ligusticum chuanxiong Hort sample, and choose 4500-7500cm
-1spectral information under characteristic wave bands, imports the ferulaic acid content value that the quantitative calibration model set up obtains described unknown Ligusticum chuanxiong Hort sample.
Described detects Ligusticum wallichii method based on population least square method supporting vector machine algorithm, described in the described step (3) of described determination of moisture step, σ is 165.4678, described γ is 2.9044e6, σ described in the described step of described extract content determination step (3) is 167.1067, and described γ is 9.7129e6; And σ is 161.6838 described in the described step (3) of described ferulaic acid content determination step, described γ is 6.0675e.
Described detects Ligusticum wallichii method based on population least square method supporting vector machine algorithm, in the described step (3) of described moisture, extract content and ferulaic acid content determination step, the parameter σ calculated according to following particle cluster algorithm and the value of γ, arranging primary group number is 100, target search space dimensionality is set to 2, using square error as objective function, find individual extreme value p at particle
bestwith global extremum g
besttime, according to following equation adjustment Position And Velocity:
Wherein, v is particle rapidity, and x is particle position, r
1, r
2the random number between 0-1, speedup factor c
1, c
2be set to: c
1=c
2=2.
Described detect Ligusticum wallichii method based on population least square method supporting vector machine algorithm, described determination of moisture, extract content measure and described ferulaic acid content determination step described step (3) in described r
1and r
2value be 0.5.
Described detects Ligusticum wallichii method based on population least square method supporting vector machine algorithm, described determination of moisture, described extract content measure and described ferulaic acid content determination step described step (2) in, diffuse reflection method is adopted to carry out the near infrared spectra collection of described Ligusticum chuanxiong Hort powder, actual conditions is take air as reference, scanning times is 32, and resolution is 8cm
-1, scanning optical spectrum scope is 4000-12000cm
-1.
Described detects Ligusticum wallichii method based on population least square method supporting vector machine algorithm, described determination of moisture, in the described step (3) of described extract content mensuration and described ferulaic acid content determination step, also comprise the step that the estimated performance of the described quantitative calibration model set up is evaluated, described evaluation index comprises coefficient R, correct checking root mean square RMSEC, checking collection root mean square RMSEP and prediction relative deviation RSEP, if R value is close to 1, and RSEP value is when being less than 10%, described quantitative calibration model is applicable to the detection to described Ligusticum chuanxiong Hort, otherwise, then inapplicable.
Described detects Ligusticum wallichii method based on population least square method supporting vector machine algorithm, and in the step (1) of described determination of moisture, employing hypobaric drying method measures the moisture in described Ligusticum chuanxiong Hort; In the step (1) that described extract content measures, employing hot dipping measures the extract content in described Ligusticum chuanxiong Hort; In the step (1) that described ferulaic acid content measures, adopt the ferulaic acid content in Ligusticum chuanxiong Hort described in high effective liquid chromatography for measuring.
Described detect Ligusticum wallichii method based on population least square method supporting vector machine algorithm, the qualified index of described Ligusticum chuanxiong Hort is: moisture≤12.0%, extract content >=12.0% and ferulaic acid content >=0.10%.
The invention provides a kind of above-mentioned purposes of detection Ligusticum wallichii method in Ligusticum chuanxiong Hort quality testing and control field.
Technique scheme of the present invention has the following advantages compared to existing technology:
(1) method of the detection Ligusticum wallichii based on population least square method supporting vector machine algorithm of the present invention, NIR technology and particle cluster algorithm are incorporated in the quality testing of Ligusticum chuanxiong Hort, can not only to each quality control index (moisture of Ligusticum chuanxiong Hort, extract, forulic acid) carry out Fast Measurement, achieve to Ligusticum chuanxiong Hort fast, comprehensive detection, also optimize quantitative calibration model performance, improve the precision of prediction of this model, substantially increase the accuracy of detection, degree of accuracy, can judge that whether quality of medicinal material is qualified fast, determine whether medicinal material can enter subsequent production process procedure, meet in producing quick, efficient requirement, there is the application prospect of the screening of on-the-spot medicinal material and quality thoroughly evaluating, and in Chinese medicine is produced, control the raw-material quality of Ligusticum wallichii from source, shorten detection time, save production cost, enhance productivity and economic benefit, ensure that the safety of Ligusticum wallichii finished dosage form quality, effectively,
(2) method of the detection Ligusticum wallichii based on population least square method supporting vector machine algorithm of the present invention, by selecting the spectral band of each quality control index in the near infrared spectrum of Ligusticum chuanxiong Hort, extract effective characteristic spectrum wave band, this characteristic spectrum wave band has good correlativity with each quality control index measured according to existing conventional method, effectively can monitor the moisture of Ligusticum chuanxiong Hort, extract and ferulaic acid content.
Accompanying drawing explanation
In order to make content of the present invention be more likely to be clearly understood, below according to a particular embodiment of the invention and by reference to the accompanying drawings, the present invention is further detailed explanation, wherein
Fig. 1 is Ligusticum chuanxiong Hort powder near infrared original absorbance spectrogram described in embodiment 1;
Fig. 2 is the correlogram of Ligusticum chuanxiong Hort aqueous powder content measured value and near infrared predicted value described in embodiment 1;
Fig. 3 is the comparison diagram of Ligusticum chuanxiong Hort aqueous powder measured value and near infrared predicted value described in embodiment 1;
Fig. 4 is the correlogram of Ligusticum chuanxiong Hort powder extract content measured value and near infrared predicted value described in embodiment 2;
Fig. 5 is the comparison diagram of Ligusticum chuanxiong Hort powder extract measured value and near infrared predicted value described in embodiment 2;
Fig. 6 is the correlogram of Ligusticum chuanxiong Hort powder ferulaic acid content measured value and near infrared predicted value described in embodiment 3;
Fig. 7 is the comparison diagram of Ligusticum chuanxiong Hort powder ferulaic acid content measured value and near infrared predicted value described in embodiment 3.
Embodiment
Major equipment used in the present invention is as follows:
The model of near infrared spectrometer is MATRIX-Fibre-based Emission, manufacturer is Bruker optik GmbH.
The present invention following example provides a kind of based on population least square method supporting vector machine algorithm detection Ligusticum wallichii method, comprises at least one item in following determination of moisture, extract content mensuration and ferulaic acid content determination step:
Embodiment 1
The mensuration of A, moisture comprises the steps:
(1) Ligusticum chuanxiong Hort deriving from Different sources choosing known moisture levels totally 116 parts pulverize after, cross 80 mesh sieves, obtain the more uniform Ligusticum chuanxiong Hort powder of granularity, for subsequent use; Wherein said Ligusticum chuanxiong Hort adopts hypobaric drying method to measure the moisture of described Ligusticum chuanxiong Hort, and concrete steps are as follows:
Get Ligusticum chuanxiong Hort powder 1g, be placed in the flat bottle (X being dried to constant weight (double difference of weighing is less than 5mg) with test sample Ligusticum chuanxiong Hort powder under the same conditions
0) in, then precise weighing (X
1), open bottle cap, by described flat bottle (X
1) be placed in vacuum drying apparatus (bottom is covered with the thick phosphorus pentoxide of about 2cm), be decompressed to 2.67kPa (20mmHg) and continue 30min below, room temperature places 24 hours.Connect anhydrous calcium chloride drying tube in vacuum drying apparatus outlet, open piston and treat that external and internal pressure is consistent, open exsiccator, cover the rapid micrometric measurement weight of bottle cap (X
2), according to the weight of less loss, calculate water cut (%)=(X in test sample
1-X
2+ X
0)/X
1× 100, calculate water cut (%) in described Ligusticum chuanxiong Hort powder;
(2) precision takes described Ligusticum chuanxiong Hort powder 1g and is placed in measuring cup, keeps powder surface smooth, and adopt diffuse reflection method to gather near infrared spectrum, spectra collection condition is take air as reference, and sweep limit is 4000-12000cm
-1, scanning times is 32 times, and resolution is 8cm
-1, every batch sample scanning repetition 3 times, be averaged spectrum, described Ligusticum chuanxiong Hort powder near infrared original absorbance spectrogram is shown in accompanying drawing 1, chooses 4500-7500cm
-1spectral band as characteristic spectrum;
(3) the parameter σ in employing particle cluster algorithm calculating least square method supporting vector machine algorithm and the value of γ, arranging primary group number is 100, and target search space dimensionality is set to 2, using square error as objective function, finds individual extreme value p at particle
bestwith global extremum g
besttime, according to following equation adjustment Position And Velocity:
Wherein, v is particle rapidity, and x is particle position, r
1, r
2the random number between 0-1, described r
1and r
2value be 0.5, speedup factor c
1, c
2be set to: c
1=c
2=2;
According to the value obtaining parameter σ and γ after the described particle cluster algorithm of operation, described σ is 165.4678, and described γ is 2.9044e6, and employing least square method supporting vector machine algorithm sets up the quantitative calibration model between the moisture of described characteristic spectrum and described Ligusticum wallichii;
After rejecting abnormalities sample, Stochastic choice 92 samples are as calibration set, and 24 samples are as checking collection (for prediction).Described quantitative calibration model adopts related coefficient (R), calibration set root mean square (RMSEC) investigates model performance, adopt checking collection root mean square (RMSEP) and prediction relative deviation (RSEP) to carry out the predictive ability of evaluation model to unknown sample simultaneously, when R value is close to 1, it is less that calibration set root mean square (RMSEC) and checking collect root mean square (RMSEP), and it is more close, when RSEP value is less than 10%, evaluation model has good predictive ability, the requirement that Ligusticum wallichii detects fast can be met, described quantitative calibration model is applicable to the detection of described Ligusticum chuanxiong Hort.As following table 1 be as described in the modeling result of near-infrared model of moisture of Ligusticum chuanxiong Hort compare, near-infrared model is linear good as can be seen from Table 1, coefficient R is all more than 0.95, and the RMSEC value of water model is less, illustrates that set up near infrared quantitative calibration modelling effect is better.Correlogram between the measured value that the moisture of described Ligusticum chuanxiong Hort conventionally measures and the predicted value utilizing described quantitative calibration model to calculate is shown in accompanying drawing 2.
Table 1 Ligusticum chuanxiong Hort moisture model parameter
| Model | R | RMSEC |
| Moisture model | 0.9834 | 0.4322 |
Correlogram between the measured value that the moisture of 24 described Ligusticum chuanxiong Hort samples of separately getting checking collection conventionally measures and the predicted value utilizing described quantitative calibration model to calculate is shown in accompanying drawing 3, can find out moisture measured value and near infrared predicted value close.The parameter being near infrared quantitative calibration model prediction result as following table 2 gathers, the model prediction related coefficient of moisture is all more than 0.8 as can be seen from Table 2, RMSEP value is less, and RSEP, within 10%, illustrates that set up quantitative calibration model has good predictive ability and stability.
The model prediction result of table 2 Ligusticum chuanxiong Hort moisture
| Model | RMSEP | R | RSEP(％) |
| Moisture model | 0.8507 | 0.8433 | 7.93 |
(4) unknown Ligusticum chuanxiong Hort sample is carried out near infrared spectrum scanning according to step (2), choose 4500-7500cm
-1characteristic wave bands under spectral information import set up quantitative calibration model in, then the content of the moisture of described Ligusticum chuanxiong Hort to be measured is calculated, described unknown Ligusticum chuanxiong Hort sample moisture≤12.0% adopting the described method utilizing near infrared spectrum to detect Ligusticum chuanxiong Hort fast to calculate.
Embodiment 2
The mensuration of B, extract content comprises the steps:
(1) Ligusticum chuanxiong Hort deriving from Different sources choosing known extract content totally 116 parts pulverize after, cross 80 mesh sieves, obtain the more uniform Ligusticum chuanxiong Hort powder of granularity, for subsequent use; Wherein said Ligusticum chuanxiong Hort measures the content of described extract according to hot dipping, and concrete steps are as follows:
Get Ligusticum chuanxiong Hort powder 1g, accurately weighed (X
1), be then placed in the conical flask of 50ml, precision adds ethanol 25ml, close plug, weighed weight, in 85 DEG C of refluxing extraction 1h after standing 1h, after letting cool, weighed weight, supplies the weight of less loss, shakes up with ethanol, be placed in the centrifugal 30min of 15ml centrifuge tube, rotating speed is 3800r/min, and precision measures supernatant 10ml, is placed in the flat bottle (X being dried to constant weight
0), after volatilizing solvent subsequently in water-bath, in 105 DEG C of dryings 3 hours, put in exsiccator and cool 30min, rapid accurately weighed weight (X
2), content (%)=(X of extract with the formula
2-X
0) × 2.5/X
1the content (%) of extract in the described Ligusticum chuanxiong Hort powder of × 100 calculating;
(2) precision takes described Ligusticum chuanxiong Hort powder 1g and puts in flat bottle, keeps powder surface smooth, and adopt diffuse reflection method to gather near infrared spectrum, spectra collection condition is take air as reference, and sweep limit is 4000-12000cm
-1, scanning times is 32 times, and resolution is 8cm
-1, every batch sample scanning repetition 3 times, be averaged spectrum, described Ligusticum chuanxiong Hort powder near infrared original absorbance spectrogram is shown in accompanying drawing 1, chooses 4500-7500cm
-1spectral band as characteristic spectrum;
(3) the parameter σ in employing particle cluster algorithm calculating least square method supporting vector machine algorithm and the value of γ, arranging primary group number is 100, and target search space dimensionality is set to 2, using square error as objective function, finds individual extreme value p at particle
bestwith global extremum g
besttime, according to following equation adjustment Position And Velocity:
Wherein, v is particle rapidity, and x is particle position, r
1, r
2the random number between 0-1, described r
1and r
2value be 0.5, speedup factor c
1, c
2be set to: c
1=c
2=2;
According to the value obtaining parameter σ and γ after the described particle cluster algorithm of operation, described σ is 167.1067, described γ is 9.7129e6, and employing least square method supporting vector machine algorithm sets up the quantitative calibration model between the extract content of described characteristic spectrum and described Ligusticum wallichii;
After rejecting abnormalities sample, Stochastic choice 92 samples are as calibration set, and 24 samples are as checking collection (for prediction).Described quantitative calibration model adopts related coefficient (R), calibration set root mean square (RMSEC) investigates model performance, adopt checking collection root mean square (RMSEP) and prediction relative deviation (RSEP) to carry out the predictive ability of evaluation model to unknown sample simultaneously, when R value is close to 1, it is less that calibration set root mean square (RMSEC) and checking collect root mean square (RMSEP), and it is more close, when RSEP value is less than 10%, evaluation model has good predictive ability, the requirement that Ligusticum wallichii detects fast can be met, described quantitative calibration model is applicable to the detection of described Ligusticum chuanxiong Hort.As following table 3 be as described in the modeling result of near-infrared model of Ligusticum chuanxiong Hort extract compare, near-infrared model is linear good as can be seen from Table 3, and coefficient R, all more than 0.95, illustrates that set up near infrared quantitative calibration modelling effect is better.Correlogram between the measured value conventionally measured of the extract content of described Ligusticum chuanxiong Hort and the predicted value utilizing described quantitative calibration model to calculate is shown in accompanying drawing 4.
Table 3 Ligusticum chuanxiong Hort extract content model parameter
| Model | R | RMSEC |
| Extract content model | 0.9962 | 0.3945 |
Separately get checking collection 24 described Ligusticum chuanxiong Hort samples extract content according to hot dipping measure measured value and the predicted value utilizing described quantitative calibration model to calculate between correlogram see accompanying drawing 5, can find out extract content measured value and near infrared predicted value close.The parameter being near infrared quantitative calibration model prediction result as following table 4 gathers, the model prediction related coefficient of extract content is all more than 0.8 as can be seen from Table 4, RSEP, within 10%, illustrates that set up quantitative calibration model has good predictive ability and stability.
The model prediction result of table 4 Ligusticum chuanxiong Hort extract content
| Model | RMSEP | R | RSEP(％) |
| Extract content model | 1.4109 | 0.7909 | 5.01 |
(4) by unknown Ligusticum chuanxiong Hort sample according to step (2) near infrared spectrum scanning, and choose 4500-7500cm
-1spectral information under spectral band imports in the quantitative calibration model set up, then calculate the content of the extract of described Ligusticum chuanxiong Hort to be measured, adopt the described method utilizing near infrared spectrum to detect Ligusticum chuanxiong Hort fast calculate described in treat unknown Ligusticum chuanxiong Hort sample extract content>=12.0%.
Embodiment 3
The mensuration of C, ferulaic acid content comprises the steps:
(1) Ligusticum chuanxiong Hort deriving from Different sources choosing known ferulaic acid content totally 116 parts pulverize after, cross 80 mesh sieves, obtain the more uniform Ligusticum chuanxiong Hort powder of granularity, for subsequent use; Wherein said Ligusticum chuanxiong Hort is according to the content of high effective liquid chromatography for measuring forulic acid, and concrete steps are as follows:
A. preprocess method is: get Ligusticum chuanxiong Hort powder (crossing 80 mesh sieves) about 1g, accurately weighed, precision adds volumetric concentration 70% methyl alcohol 50ml, 90 DEG C of refluxing extraction 0.5h, let cool, supply weightlessness with the methyl alcohol of 70%, get appropriate amount of sample solution in 1.5ml centrifuge tube under rotating speed 13000r/min condition centrifugal 10min, get supernatant, obtain need testing solution;
B. liquid phase chromatogram condition: chromatographic column: GraceSmart RP 18 analytical column (specification is 4.6 × 250mm, 5 μm); Mobile phase: methyl alcohol-0.1% phosphoric acid solution volume ratio is 25:75; Determined wavelength 321nm, flow velocity is 1mL/min, and sample size is 10 μ L.Theoretical cam curve is calculated should be not less than 4000 by Quercetin.
(2) precision takes described Ligusticum chuanxiong Hort powder 1g and puts in flat bottle, keeps powder surface smooth, and adopt diffuse reflection method to gather near infrared spectrum, spectra collection condition is take air as reference, and sweep limit is 4000-12000cm
-1, scanning times is 32 times, and resolution is 8cm
-1, every batch sample scanning repetition 3 times, be averaged spectrum, described Ligusticum chuanxiong Hort powder near infrared original absorbance spectrogram is shown in accompanying drawing 1, chooses 4500-7500cm
-1spectral band as characteristic spectrum;
(3) the parameter σ in employing particle cluster algorithm calculating least square method supporting vector machine algorithm and the value of γ, arranging primary group number is 100, and target search space dimensionality is set to 2, using square error as objective function, finds individual extreme value p at particle
bestwith global extremum g
besttime, according to following equation adjustment Position And Velocity:
Wherein, v is particle rapidity, and x is particle position, r
1, r
2the random number between 0-1, described r
1and r
2value be 0.5, speedup factor c
1, c
2be set to: c
1=c
2=2;
According to the value obtaining parameter σ and γ after the described particle cluster algorithm of operation, described σ is 161.6838, described γ is 6.0675e, and employing least square method supporting vector machine algorithm sets up the quantitative calibration model between the ferulaic acid content of described characteristic spectrum and described Ligusticum wallichii;
After rejecting abnormalities sample, Stochastic choice 92 samples are as calibration set, and 24 samples are as checking collection (for prediction).Described quantitative calibration model adopts related coefficient (R), calibration set root mean square (RMSEC) investigates model performance, adopt checking collection root mean square (RMSEP) and prediction relative deviation (RSEP) to carry out the predictive ability of evaluation model to unknown sample simultaneously, when R value is close to 1, it is less that calibration set root mean square (RMSEC) and checking collect root mean square (RMSEP), and it is more close, when RSEP value is less than 10%, evaluation model has good predictive ability, the requirement that Ligusticum wallichii detects fast can be met, described quantitative calibration model is applicable to the detection of described Ligusticum chuanxiong Hort.As following table 5 be as described in the modeling result of near-infrared model of content of forulic acid of Ligusticum chuanxiong Hort compare, as can be seen from Table 5, the R value of forulic acid model is more than 0.95, but RMSECV value is less.Correlogram between the measured value according to high effective liquid chromatography for measuring of the ferulaic acid content of described Ligusticum chuanxiong Hort and the predicted value utilizing described quantitative calibration model to calculate is shown in accompanying drawing 6.
Table 5 Ligusticum chuanxiong Hort ferulaic acid content model parameter
| Model | R | RMSEC |
| Ferulaic acid content model | 0.9916 | 0.0052 |
Separately get and verify that 24 the described Ligusticum chuanxiong Horts integrated are as sample, correlogram between the measured value according to high effective liquid chromatography for measuring of the ferulaic acid content of described Ligusticum chuanxiong Hort and the predicted value utilizing described quantitative calibration model to calculate is shown in accompanying drawing 7, can find out above-mentioned ferulaic acid content measured value and near infrared predicted value close.The parameter being near infrared quantitative calibration model prediction result as following table 6 gathers, the model prediction related coefficient of ferulaic acid content as can be seen from Table 6, although forulic acid is 0.71, RMSEP value is less, RSEP is 15.21%, predictive ability is not good enough but still can accept, and illustrates that set up quantitative calibration model has good predictive ability and stability.
The model prediction result of table 6 Ligusticum chuanxiong Hort ferulaic acid content
| Model | RMSEP | R | RSEP(％) |
| Ferulaic acid content model | 0.0266 | 0.7181 | 15.21 |
(4) unknown Ligusticum chuanxiong Hort sample is carried out near infrared spectrum scanning according to step (2), and choose 4500-7500cm
-1spectral information under characteristic wave bands imports in the quantitative calibration model set up, then the content of the forulic acid of described Ligusticum chuanxiong Hort to be measured is calculated, ferulaic acid content>=0.10% of the described unknown Ligusticum chuanxiong Hort sample adopting the described method utilizing near infrared spectrum to detect Ligusticum chuanxiong Hort fast to calculate.
Obviously, above-described embodiment is only for clearly example being described, and the restriction not to embodiment.For those of ordinary skill in the field, can also make other changes in different forms on the basis of the above description.Here exhaustive without the need to also giving all embodiments.And thus the apparent change of extending out or variation be still among the protection domain of the invention.
Claims (9)
1. detect a Ligusticum wallichii method based on population least square method supporting vector machine algorithm, it is characterized in that, comprise at least one item in following determination of moisture, extract content mensuration and ferulaic acid content determination step:
The mensuration of A, moisture comprises the steps:
(1) choose the Ligusticum chuanxiong Hort of known moisture levels, pulverize and sieve, obtain described Ligusticum chuanxiong Hort powder, for subsequent use;
(2) the Ligusticum chuanxiong Hort powder after above-mentioned process is carried out near infrared spectrum scanning, gather the near infrared spectrum of described Ligusticum chuanxiong Hort, choose 4500-7500cm
-1spectral band as characteristic spectrum;
(3) particle cluster algorithm is adopted to calculate the value of parameter σ in least square method supporting vector machine algorithm and γ, the quantitative calibration model between the moisture then setting up described characteristic spectrum and described Ligusticum wallichii with least square method supporting vector machine algorithm;
(4) according to the method for described step (2), near infrared spectrum scanning is carried out to unknown Ligusticum chuanxiong Hort sample, and choose 4500-7500cm
-1characteristic spectrum, imports the moisture content value that the quantitative calibration model set up obtains described unknown Ligusticum chuanxiong Hort sample;
The mensuration of B, extract content comprises the steps:
(1) choose the Ligusticum chuanxiong Hort of known extract content, pulverize and sieve, obtain described Ligusticum chuanxiong Hort powder, for subsequent use;
(2) the Ligusticum chuanxiong Hort powder after above-mentioned process is carried out near infrared spectrum scanning, gather the near infrared spectrum of described Ligusticum chuanxiong Hort, choose 4500-7500cm
-1spectral band as characteristic spectrum;
(3) particle cluster algorithm is adopted to calculate the value of parameter σ in least square method supporting vector machine algorithm and γ, the quantitative calibration model between the extract content then setting up described characteristic spectrum and described Ligusticum wallichii with least square method supporting vector machine algorithm;
(4) according to the method for described step (2), near infrared spectrum scanning is carried out to unknown Ligusticum chuanxiong Hort sample, and choose 4500-7500cm
-1characteristic spectrum, imports the extract content value that the quantitative calibration model set up obtains described unknown Ligusticum chuanxiong Hort sample;
The mensuration of C, ferulaic acid content comprises the steps:
(1) choose the Ligusticum chuanxiong Hort of known ferulaic acid content, pulverize and sieve, obtain described Ligusticum chuanxiong Hort powder, for subsequent use;
(2) the Ligusticum chuanxiong Hort powder after above-mentioned process is carried out near infrared spectrum scanning, gather the near infrared spectrum of described Ligusticum chuanxiong Hort, choose 4500-7500cm
-1spectral band as characteristic spectrum;
(3) particle cluster algorithm is adopted to calculate the value of parameter σ in least square method supporting vector machine algorithm and γ, the quantitative calibration model between the ferulaic acid content then setting up described characteristic spectrum and described Ligusticum wallichii with least square method supporting vector machine algorithm;
(4) according to the method for described step (2), near infrared spectrum scanning is carried out to unknown Ligusticum chuanxiong Hort sample, and choose 4500-7500cm
-1characteristic spectrum, imports the ferulaic acid content value that the quantitative calibration model set up obtains described unknown Ligusticum chuanxiong Hort sample.
2. according to claim 1 based on population least square method supporting vector machine algorithm detection Ligusticum wallichii method, it is characterized in that, described in the described step (3) of described determination of moisture step, σ is 165.4678, and described γ is 2.9044e6; Described in the described step (3) of described extract content determination step, σ is 167.1067, and described γ is 9.7129e6; And σ is 161.6838 described in the described step (3) of described ferulaic acid content determination step, described γ is 6.0675e.
3. according to claim 1 and 2 based on population least square method supporting vector machine algorithm detection Ligusticum wallichii method, it is characterized in that, in the described step (3) of described moisture, extract content and ferulaic acid content determination step, the parameter σ calculated according to following particle cluster algorithm and the value of γ, arranging primary group number is 100, target search space dimensionality is set to 2, using square error as objective function, finds individual extreme value p at particle
bestwith global extremum g
besttime, according to following equation adjustment Position And Velocity:
Wherein, v is particle rapidity, and x is particle position, r
1, r
2the random number between 0-1, speedup factor c
1, c
2be set to: c
1=c
2=2.
4. described detect Ligusticum wallichii method based on population least square method supporting vector machine algorithm according to claim 1-3 is arbitrary, it is characterized in that, described determination of moisture, extract content measure and described ferulaic acid content determination step described step (3) in described r
1and r
2value be 0.5.
5. described detect Ligusticum wallichii method based on population least square method supporting vector machine algorithm according to claim 1-4 is arbitrary, it is characterized in that, described determination of moisture, described extract content measure and described ferulaic acid content determination step described step (2) in, diffuse reflection method is adopted to carry out the near infrared spectra collection of described Ligusticum chuanxiong Hort powder, actual conditions is take air as reference, scanning times is 32, and resolution is 8cm
-1, scanning optical spectrum scope is 4000-12000cm
-1.
6. described detect Ligusticum wallichii method based on population least square method supporting vector machine algorithm according to claim 1-5 is arbitrary, it is characterized in that, described determination of moisture, in the described step (3) of described extract content mensuration and described ferulaic acid content determination step, also comprise the step that the estimated performance of the described quantitative calibration model set up is evaluated, described evaluation index comprises coefficient R, correct checking root mean square RMSEC, checking collection root mean square RMSEP and prediction relative deviation RSEP, if R value is close to 1, and RSEP value is when being less than 10%, described quantitative calibration model is applicable to the detection to described Ligusticum chuanxiong Hort, otherwise, then inapplicable.
7. described detect Ligusticum wallichii method based on population least square method supporting vector machine algorithm according to claim 1-6 is arbitrary, it is characterized in that, in the step (1) of described determination of moisture, employing hypobaric drying method measures the moisture in described Ligusticum chuanxiong Hort; In the step (1) that described extract content measures, employing hot dipping measures the extract content in described Ligusticum chuanxiong Hort; In the step (1) that described ferulaic acid content measures, adopt the ferulaic acid content in Ligusticum chuanxiong Hort described in high effective liquid chromatography for measuring.
8. described detect Ligusticum wallichii method based on population least square method supporting vector machine algorithm according to claim 1-7 is arbitrary, it is characterized in that, the qualified index of described Ligusticum chuanxiong Hort is: moisture≤12.0%, extract content >=12.0% and ferulaic acid content >=0.10%.
9. the arbitrary described purposes of detection Ligusticum wallichii method in Ligusticum chuanxiong Hort quality testing and control field of claim 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510249332.8A CN104819955B (en) | 2015-05-15 | 2015-05-15 | Ligusticum wallichii method and application are detected based on population least square method supporting vector machine algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510249332.8A CN104819955B (en) | 2015-05-15 | 2015-05-15 | Ligusticum wallichii method and application are detected based on population least square method supporting vector machine algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104819955A true CN104819955A (en) | 2015-08-05 |
| CN104819955B CN104819955B (en) | 2017-06-16 |
Family
ID=53730311
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510249332.8A Active CN104819955B (en) | 2015-05-15 | 2015-05-15 | Ligusticum wallichii method and application are detected based on population least square method supporting vector machine algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104819955B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106770007A (en) * | 2016-11-29 | 2017-05-31 | 中国石油大学(华东) | A kind of characteristic wavelength of near-infrared spectrum system of selection for least square method supporting vector machine model |
| CN108398515A (en) * | 2017-02-04 | 2018-08-14 | 北京蓝标成科技有限公司 | A kind of quality determining method of Dendrobium Chrysotoxum Lindl |
| CN113074663A (en) * | 2021-03-29 | 2021-07-06 | 大连理工大学 | Light beam vector deviation planning compensation method for spectrum confocal on-line measurement |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102313714A (en) * | 2011-09-21 | 2012-01-11 | 浙江大学 | Determination method of carthamus tinctorius extract |
| CN104597153A (en) * | 2014-12-30 | 2015-05-06 | 上海现代中医药股份有限公司 | Assay method of ligusticum wallichii fingerprint |
-
2015
- 2015-05-15 CN CN201510249332.8A patent/CN104819955B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102313714A (en) * | 2011-09-21 | 2012-01-11 | 浙江大学 | Determination method of carthamus tinctorius extract |
| CN104597153A (en) * | 2014-12-30 | 2015-05-06 | 上海现代中医药股份有限公司 | Assay method of ligusticum wallichii fingerprint |
Non-Patent Citations (4)
| Title |
|---|
| 张海珠等: "云连药材水分、灰分和浸出物的含量测定", 《大理学院学报》 * |
| 毕艳兰等: "粒子群最小二乘支持向量机结合偏最小二乘法用于芝麻油质量的鉴别", 《分析 化学研究报告》 * |
| 王小梅等: "近红外漫反射光谱法定量分析川芎中的阿魏酸含量", 《药物分析杂志》 * |
| 金叶等: "基于粒子群算法的最小二乘支持向量机在红花提取液近红外定量分析中的应用", 《分析化学研究报告》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106770007A (en) * | 2016-11-29 | 2017-05-31 | 中国石油大学(华东) | A kind of characteristic wavelength of near-infrared spectrum system of selection for least square method supporting vector machine model |
| CN108398515A (en) * | 2017-02-04 | 2018-08-14 | 北京蓝标成科技有限公司 | A kind of quality determining method of Dendrobium Chrysotoxum Lindl |
| CN108398515B (en) * | 2017-02-04 | 2020-09-29 | 蓝标一成(海南)生物科技有限公司 | Quality detection method of dendrobium chrysotoxum |
| CN113074663A (en) * | 2021-03-29 | 2021-07-06 | 大连理工大学 | Light beam vector deviation planning compensation method for spectrum confocal on-line measurement |
| CN113074663B (en) * | 2021-03-29 | 2022-01-18 | 大连理工大学 | Light beam vector deviation planning compensation method for spectrum confocal on-line measurement |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104819955B (en) | 2017-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104359853B (en) | Method and application using near infrared spectroscopy quick detection Uncaria | |
| CN103439288B (en) | A kind of real-time release detection method for ginkgo leaf medicinal material | |
| CN104062257B (en) | A kind of based on the method for general flavone content near infrared ray solution | |
| CN104849233B (en) | A kind of method and device of detection cereal new-old degree | |
| CN102288572A (en) | Method for quickly detecting content of index ingredient of traditional Chinese medicinal material by utilizing near infrared spectrum technique | |
| CN103364362A (en) | Method for identifying Chinese herbal medicines by using THz-TDS combined with chemometrics | |
| CN102243170A (en) | Method for identifying production place of dwarf lilyturf root by using near infrared spectrum technology | |
| CN101231274B (en) | Method for rapid measuring allantoin content in yam using near infrared spectrum | |
| CN104833654A (en) | Method and application for rapidly detecting mother chrysanthemum medicinal materials by utilizing near infrared spectroscopy | |
| CN103969211B (en) | A kind of method using near infrared spectrum detection FUFANG DANSHEN PIAN moisture | |
| CN106138131B (en) | A kind of bupleurum particles and its Chinese medicine preparation | |
| CN104568822A (en) | Method for rapidly and simultaneously detecting multiple indexes of fructus forsythia medicinal material | |
| CN105548026A (en) | Quick detection method for quality control of radix curcumae medicinal material | |
| CN104792652A (en) | Multi-index rapid detection method for radix astragali | |
| CN102914596B (en) | Method for rapidly determining PienTzeHuang quality by using near-infrared spectroscopy | |
| CN104568813A (en) | Multi-index fast detection method of dogwood medicinal materials | |
| CN104316490A (en) | Method and application for rapidly detecting caulis spatholobi medicine material by using near infrared spectroscopy method | |
| CN105784635A (en) | Folium apocyni veneti total flavonoid near infrared super rapid detection method | |
| CN104819955B (en) | Ligusticum wallichii method and application are detected based on population least square method supporting vector machine algorithm | |
| CN108051396A (en) | A kind of rapid detection method of Xin Ke Shu ' tablet for treating coronary heart disease active constituent content | |
| CN102147360A (en) | Detection method for drug combination | |
| CN103792205A (en) | High-flux near-infrared sensitive fast non-destructive analysis for impurities and tensile strength of tablets | |
| CN105784951A (en) | Multiple indicator rapid detection method for raw medicinal powder of condensed pill of six drugs with rehmannia | |
| CN104535524A (en) | Method for near-infrared online monitoring content change of total flavonol glycosides in folium ginkgo extraction process | |
| CN102323236B (en) | Method for detecting contents of a plurality of components during sophora flavescens extracting process through near infrared spectrum |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20221207 Address after: No.1 Guanqing Road, Guanlan hi tech park, Guanhu street, Longhua District, Shenzhen, Guangdong Province Patentee after: CHINA RESOURCES SANJIU MEDICAL & PHARMACEUTICAL Co.,Ltd. Patentee after: China Resources Sanjiu modern traditional Chinese Medicine Pharmaceutical Co.,Ltd. Address before: No.1 Guanqing Road, Guanlan high tech park, Longhua New District, Shenzhen, Guangdong 518110 Patentee before: CHINA RESOURCES SANJIU MEDICAL & PHARMACEUTICAL Co.,Ltd. |
|
| TR01 | Transfer of patent right |