CN104596984A - Method for detecting medicated leaven fermentation process quality on line by using near infrared spectrum - Google Patents
Method for detecting medicated leaven fermentation process quality on line by using near infrared spectrum Download PDFInfo
- Publication number
- CN104596984A CN104596984A CN201510002633.0A CN201510002633A CN104596984A CN 104596984 A CN104596984 A CN 104596984A CN 201510002633 A CN201510002633 A CN 201510002633A CN 104596984 A CN104596984 A CN 104596984A
- Authority
- CN
- China
- Prior art keywords
- sample
- near infrared
- infrared spectrum
- medicated leaven
- activity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a method for detecting medicated leaven fermentation process quality indexes on line. The method comprises the following steps: (1) sampling at different medicated leaven fermentation time points; (2) obtaining the near infrared spectrum data of each sample by adopting a diffuse reflectance method, and measuring the protease activity and/or the amylase activity of each sample; (3) establishing the correlation between the near infrared spectrum data of each sample and the protease activity of each sample and/or the correlation between the near infrared spectrum data of each sample and the amylase activity of each sample by adopting a least square support vector machine method, a partial least square method or a neural network, preferentially adopting the least square support vector machine method to obtain an enzyme activity predication model based on a near infrared spectrum; and (4) detecting the medicated leaven fermentation process quality indexes on line by adopting the enzyme activity predication model. The method disclosed by the invention can be used for controlling the production process quality rapidly in real time and has the advantages of wide application range and good applicability.
Description
Technical field
The present invention relates to a kind of Medicated Leaven sweat near infrared online quality determining method.
Background technology
Medicated Leaven is most representational traditional Chinese medicine fermentation processed product simply, is by semen armeniacae amarae, rde bean, fresh sweet wormwood, fresh Siberian cocklebur grass, fresh polygonum flaccidum etc., and the herbal leaven made through fermentation after adding flour (or wheat bran) mixing is with a long history, is widely used.
The application of fermentation technique in Chinese medicine preparation process has a long history, it refers under certain humidity and temperature condition, medicinal material after treated or clean system or medicinal material are mixed and are added auxiliary material etc., utilize the decomposition catalytic action of enzyme and microorganism, make the concocting method of medicine foaming, raw clothing.
At present, rely on subjective experience to the control of the traditional zymotic concocting process of Medicated Leaven, control technology is delayed, lacks objective index and effective method, and the quality stability of Medicated Leaven fermented product is not good, and the mass discrepancy between different batches product is larger more.The Medicated Leaven adopting modern zymotechnique to produce and other herb fermenting processed products, also be only that the technological parameter such as temperature, humidity to sweat controls, and to produce main body---the fermentation materials being in dynamic change does not carry out the detection of correlated quality index, after fermentation ends, just product is carried out to the detection of quality control index, judge that whether product is qualified, and the quality of the processed product that now ferments is formed already, now also cannot exert one's influence even if product is defective.Therefore, the method that there is no in the zymotechnique of current Medicated Leaven changes the quality of fermentation processed product by the control of production run, promote it towards qualified future development.
In addition, the qualification of the Medicated Leaven quality of traditional zymotic explained hereafter there is no perfect index.In recent years, a large amount of research work had been carried out in the quality control for Medicated Leaven, and current research focuses mostly on using diastase and the proteinase activity quality control index as Medicated Leaven.The mensuration of diastase and prolease activity has multiple method, conventional has Folin-phenol method and DNS method, these methods all need to expend the regular hour, and all for concocting the quality control of final products to fermentation, and cannot control effectively to the production run of Medicated Leaven.
In order to understand the dynamic change of sweat intermediate product quality in real time, be conducive to the developing direction of the control and optimize intermediate product quality by technological parameter, thus improve the up-to-standard rate of finished product, need the online test method setting up a kind of Medicated Leaven sweat quality index.
Summary of the invention
The object of the present invention is to provide a kind of online test method of Medicated Leaven sweat quality index, thus be conducive to the dynamic change understanding sweat intermediate product quality in real time, be conducive to the developing direction of the control and optimize intermediate product quality by technological parameter, thus improve the up-to-standard rate of finished product.
The object of the invention is to be achieved through the following technical solutions:
In Medicated Leaven sweat, an online test method for quality index, comprises the steps:
(1) sample in Medicated Leaven sweat different time points;
(2) diffuse reflectance is adopted to obtain the near infrared spectrum data of each sample; Measure prolease activity and/or the amylase activity of each sample respectively;
(3) least square method supporting vector machine method (Least squares-support vectormachine is adopted, LS-SVM), partial least square method (Partial Least Squares method, PLS) or neural network (Neutral network) set up relation between the near infrared spectrum data of sample and proteinase and/or amylase activity, preferred employing least square method supporting vector machine method, obtains the enzyme activity forecast model based near infrared spectrum;
(4) the above-mentioned enzyme activity forecast model based near infrared spectrum is adopted to realize the on-line checkingi of quality index in Medicated Leaven sweat.
Described quality index typically refers to the vigor of enzyme, as proteinase and/or diastatic vigor.
In step (1), described Medicated Leaven sweat can adopt conventional method to carry out, such as, adopt the spontaneous fermentation that traditional multi-cultur es mixes.
In a preferred embodiment of the present invention, described Medicated Leaven fermentation is the pure-blood ferment adopting single bacterial classification.Compared with traditional spontaneous fermentation, can not there are other microbial contaminations in the process of pure-blood ferment, in material, enzyme activity constantly raises, thus avoid the enzyme activity constantly raised to be interfered.The enzyme activity of different time points sample is increased to maximal value gradually from zero, and enzyme activity distribution range is comparatively wide, is conducive to the robustness promoting institute's established model.
Described bacterial classification can be the fermented bacterium that this area routine uses, such as bacillus subtilis (Bacillus subtilis), Pichia pastoris (Pichia Hansen), Sai Shi aspergillus (Aspergillusversicolor) etc.
In step (1), the sampling of described Medicated Leaven sweat, can be sample in same Medicated Leaven sweat, sample in the different Medicated Leaven sweat that also can carry out at the same time or successively, be not specifically limited, as long as the different phase that can obtain residing for enzyme activity and the enough large sample of formation scope.
In step (1), when described Medicated Leaven sweat is more than one, each Medicated Leaven sweat can adopt identical or different fermentation medium, preferably adopts different fermentation mediums.Adopt different nutrient culture media, the distribution range of diastase and prolease activity in sample can be expanded on the one hand further, on the other hand, from the different time points sampling the sweat of different culture media, when gained sample is used for proteinase and amylase assay, its background is different, expands the scope of application of forecast model further.
The nutrient culture media that described fermentation medium can adopt any routine to use, such as flour, wheat bran or the potpourri of different proportion both adopting as required.
In a kind of preferred embodiment of the present invention, the potpourri of flour and wheat bran 1:1 and whole meal is adopted to carry out the sweat of Medicated Leaven respectively, and from the different time points sampling these two different fermentations processes.
In step (1), the sampling time point of described sample is not strict with, as long as the different phase that can obtain residing for enzyme activity and the enough large sample of formation scope.
Preferably, with the 1/50-1/15 of all times that ferments, be preferably the 1/40-1/20 of fermentation all times, the 1/35-1/25 being more preferably fermentation all times is sampling interval time; Each sampling amount is 1-20, is preferably 1-15, is more preferably 1-10, then is preferably 1-8.Such sampling mode can ensure that the stage residing for enzyme activity of got sample is different and formation scope is enough large, makes enzyme activity distribution in sample enough even, is beneficial to set up reliable model.
In step (2), preferably, the wavelength coverage of described near infrared spectrum is selected from 400-2500nm wavelength coverage.
In step (2), amylase activity, prolease activity are quality detecting index the most frequently used in Medicated Leaven sweat.Which kind of index adopted can be selected according to actual needs, such as can using amylase activity, prolease activity one of them as Testing index, also can Simultaneously test diastase and the vigor both proteinase as Testing index.
In step (2), described prolease activity and amylase activity can adopt the existing assay method in this area.Preferably, adopt the prolease activity of Folin-phenol method working sample, adopt the amylase activity of DNS method working sample.The concrete operation method of these two kinds of methods is well known to those skilled in the art, is not repeated herein.
In step (3), described least square method supporting vector machine method, partial least square method and neural network all have implication well known in the art.
Wherein, least square method supporting vector machine method is one of expansion of support vector machine method, it changes the inequality constrain in support vector machine (Support vector machine) into equality constraint, and using the cost metric of error sum of squares loss function as method establishment, be converted into solve system of linear equations problem solving quadratic programming problem, improve the precision of speed and the convergence solved.Its concrete grammar is, first uses nonlinear mapping function ψ (.) that sample is mapped to feature space from luv space
(.), at high-dimensional feature space structure optimization objective function, the conversion of objective function is carried out by Lagrangian method, the last parameter being obtained model by least square method, obtain forecast model, such as document LeastSquares Support Vector Machines (J.A.K.Suykens, T.Van Gestel, J.DeBrabanter, B.De Moor, J.Vandewalle,, World Scientific, Singapore, 2002 (ISBN 981-238-151-1), pp.98-99) described in method:
Described partial least square method, on the basis (least square method) of common multiple regression, combines the thought of principal component analysis (PCA) PCA and canonical correlation analysis CCA, solves the problem of independent variable multicollinearity in regretional analysis.It finds the direction (latent variable) that can maximize and explain Y variation in X space, carries out the prediction of Y based on latent variable.First X is resolved into score T and load p by it, then Y is decomposed into score U and load Q:
X＝TP
T+E
Y＝UQ
T+F
E and F is residual error item, then T and U is done linear regression:
U＝TB；B＝(T
TT)
-1T
TY
Prediction for new samples:
Y
newly=T
newlybQ
Described neural network refers to BP neural network.It comprises three parts: input layer, hidden layer and output layer.Data are inputted by input layer, and through standardization, and weighted transmission is to the second layer (hidden layer).Hidden layer is transferred to the result that output layer provides prediction after the process such as weighting, the process of s type function.The study of neural network parameter adopts Feedback error method (Back propagation, BP), take least square function as the objective function of parameter optimization training.
In step (3), described near infrared spectrum data can adopt analysis software conventional in field to process, and in employing field, conventional method carries out pre-service to spectroscopic data, to make baseline steady, improves the robustness of model.
In one embodiment of the invention, in step (3), near infrared spectrum data process adopts MATLAB R2011b analysis software, expansion multiplicative scatter correction method (ExtendedMultiplicative Signal Correction, EMSC) is adopted to carry out pre-service to spectroscopic data.
In step (3), can adopt 5 folding cross-validation methods, 10 folding cross-validation methods or 20 folding cross-validation methods, preferably adopt 10 folding cross-validation methods, the best latent variable of preference pattern is because of subnumber.Cross validation mean square deviation (Root Mean Square Error of Cross-validation, RMSECV) tends to be steady with the increase because of subnumber, preferably report to the leadship after accomplishing a task checking variance minimum value corresponding to because of subnumber be that optimum latent variable is because of subnumber.
Described cross-validation method is the method being commonly used to testing algorithm accuracy.Such as 10 folding cross-validation methods, its thought is divided into very by data set, in turn will wherein 9 parts as training data, as test data, test for 1 part.Each test all can draw corresponding accuracy (totally 10).The accuracy of 10 results is averaged as the estimation to model accuracy.
In Medicated Leaven sweat, can the near infrared spectrum data of Real-time Obtaining material, by the model of above-mentioned foundation, calculate diastase and/or the prolease activity of this material in real time, thus realize the on-line checkingi of quality index in Medicated Leaven sweat.
After obtaining the enzyme activity forecast model of near infrared spectrum, can by the following method institute's established model be verified and be evaluated.
The sample of sampling is subdivided into calibration set to collect with checking, using the coefficient R c of the chemical measurements of calibration set sample and model predication value with corrects mean square deviation as the fitting effect of metrics evaluation model and inner robustness, to verify the coefficient R p of chemical measurements and the model predication value integrating sample and to predict the estimated performance of mean square deviation as metrics evaluation model.
Preferably, adopt Kennard-Stone (KS) sample partitioning sampling sample to be subdivided into calibration set to collect with checking.
KS method is using all samples as training set candidate samples, therefrom selects part sample successively and enters training set.First, select Euclidean distance two samples farthest in candidate samples to enter training set, then that candidate samples in candidate samples with maximum minimum Eustachian distance is selected in training set, until sample number meets the demands in sample set.The advantage of the method to guarantee that training set and checking concentrate sample to be evenly distributed in sample space.
Euclidean distance is a distance definition usually adopted, and refers to the actual distance between two points in m-dimensional space, or the natural length of vector (namely this point is to the distance of initial point).Euclidean distance in two and three dimensions space is exactly the actual range between 2.
The fitting effect of institute's established model and inner robustness are using the coefficient R c of the chemical measurements of calibration set sample and model predication value and correct mean square deviation (Root Mean Square Error ofCalibration, RMSEC) as evaluation index.Rc is larger, and RMSEC value is less, corresponding model robustness and fitting effect better.
Wherein, C
ifor proteinase or the work of determination of amylase enzyme of i-th sample in calibration set;
for the near-infrared spectrum analysis predicted value of i-th sample in calibration set;
for the mean value that calibration set sample enzyme is lived; N is the number of calibration set sample.
The estimated performance of institute's established model integrates the chemical measurements of sample and the coefficient R p of model predication value with checking and predicts that mean square deviation (Root Mean Square Error of Prediction, RMSEP) is as evaluation index.Rp is larger, and RMSEP value is less, and corresponding model prediction performance is better.
Wherein, C
ifor the proteinase or determination of amylase enzyme of verifying concentrated i-th sample are lived;
for verifying the near-infrared spectrum analysis predicted value of concentrated i-th sample;
for the mean value that checking collection sample enzyme is lived; N is the number of checking collection sample.
Adopt the inventive method can set up the qualified enzyme activity forecast model based near infrared spectrum.After described forecast model is set up, carry out fermenting adopting any-mode Medicated Leaven production run in, can to intermediate material on-line period and its near infrared spectrum data of Real-time Obtaining, based on set up enzyme activity forecast model, calculate the enzyme activity value of fermentation intermediate material fast, thus the fermentation appearance of intermediate material in grasp fermentation tank, according to fermentation appearance, such as temperature can be adjusted in time, humidity, the zymotechnique controling parameters such as throughput, thus the quality of promotion intermediate material (mainly enzyme is lived) is towards qualified future development, ensure up-to-standard rate and the stability of Medicated Leaven final products.
The enzyme activity forecast model adopting the inventive method to set up, has in real time, fast, can be used for production run quality control, the advantages such as applied range, applicability are good.Adopt the online test method based on this forecast model, be conducive to the process control realizing Medicated Leaven sweat.
Accompanying drawing explanation
Fig. 1 shows calibration set and collects the distribution of sample in major component space with checking.
Fig. 2 is the original near infrared spectrum of 67 samples.
Fig. 3 carries out pretreated spectrogram to the near infrared spectrum data of 67 samples.
Fig. 4 shows prolease activity near-infrared model predicted value and the measured value correlogram of the foundation of LS-SVM method.
Fig. 5 shows amylase activity near-infrared model predicted value and the measured value correlogram of the foundation of LS-SVM method.
Embodiment
Be described in detail below by way of the online test method of specific embodiment to Medicated Leaven sweat quality index of the present invention, be beneficial to those skilled in the art's the understanding of the present invention.But described embodiment is not construed as limiting the invention.
Embodiment 1
The near infrared spectrometer adopted in the present embodiment is that FOSS company produces.Near infrared spectrum data process adopts MATLAB R2011b analysis software.
1. materials and methods
1.1 sample preparation
Raw material is taken respectively according to following Medicated Leaven formula:
A. flour 60g, wheat bran 60g, semen armeniacae amarae 4.8g, rde bean 4.8g, sweet wormwood 2.8g, Siberian cocklebur grass 2.8g, polygonum flaccidum 2.8g;
B. flour 120g, semen armeniacae amarae 4.8g, rde bean 4.8g, sweet wormwood 2.8g, Siberian cocklebur grass 2.8g, polygonum flaccidum 2.8g;
Above-mentioned two groups of raw materials are placed in two fermentation tanks respectively, inoculate bacillus subtilis 10% respectively, carry out solid-state pure-blood ferment.Sweat carries out 7 days, every sampling in 4 ~ 6 hours once, collects the Medicated Leaven sample of different fermentations time, collects 35, flour wheat bran ratio 1:1 fermentation group sample altogether, 32, whole meal fermentation group sample.
1.2 near-infrared diffuse reflection spectrum collections
Collect all samples, utilize FOSS near infrared spectrometer to gather spectroscopic data.With built-in background for reference, gather near infrared spectrum with diffuse reflectance, wavelength band is 400-2500nm, resolution 0.5nm.Accumulative scanning times 32 times, each sample fills out sample 3 times, fills out sample multiple scanning 3 times at every turn, calculates averaged spectrum and is used for analyzing.
1.3 near infrared spectrum data process
Expansion multiplicative scatter correction method is adopted to carry out pre-service to spectroscopic data.
1.4 proteinase and amylase activity measure
1.4.1 prepared by enzyme liquid
Above-mentioned sample, after near infrared spectra collection, takes 5g, puts in 100mL triangular flask, adds 50mL distilled water, and in 40 DEG C of thermostat water baths, 1h is extracted in concussion, and extract centrifugal (5000r/min, 15min), gets supernatant, i.e. enzyme liquid.
1.4.2 prolease activity measures
Measure by Folin-phenol colourimetry.Typical curve regression equation is y=0.0109x, R
2=0.999.
1.4.3 amylase activity measures
Measure by DNS colourimetry.Typical curve regression equation is y=0.6832x-0.0555, R
2=0.999.
The foundation of 1.5 models
Partial least square method (PLS) and least square method supporting vector machine method (LS-SVM) is adopted to set up the forecast model of Medicated Leaven prolease activity and amylase activity respectively.
When adopting PLS modeling, adopt the best latent variable of 10 folding cross-validation method preference patterns because of subnumber.
When adopting LS-SVM modeling, same employing 10 folding cross-validation methods, thus make two kinds of modeling methods have comparability.
The fitting effect of model is using the coefficient R c of the chemical measurements of calibration set sample and model predication value and correct mean square deviation RMSEC as evaluation index.Rc is larger, and RMSEC value is less, and corresponding fitting effect is better.
Wherein,
cifor proteinase or the work of determination of amylase enzyme of i-th sample in calibration set;
for the near-infrared spectrum analysis predicted value of i-th sample in calibration set;
for the mean value that calibration set sample enzyme is lived; N is the number of calibration set sample.
The estimated performance of model integrates the chemical measurements of sample and the coefficient R p of model predication value with checking and predicts that mean square deviation RMSEP is as evaluation index.Rp is larger, and the less corresponding model prediction performance of RMSEP value is better.
Wherein,
cifor the proteinase or determination of amylase enzyme of verifying concentrated i-th sample are lived;
for verifying the near-infrared spectrum analysis predicted value of concentrated i-th sample;
for the mean value that checking collection sample enzyme is lived; N is the number of checking collection sample.
2 results
Proteinase and diastatic enzyme activity determination result in 2.1 Medicated Leavens
The present invention adopts Folin-phenol method and DNS method to determine prolease activity and the amylase activity of 67 Medicated Leaven samples respectively.Measurement result is in table 1 and table 2.Having a very wide distribution of proteinase and amylase activity, meets modeling demand substantially.
Prolease activity measurement result in table 1 Medicated Leaven
Table 2 Medicated Leaven amylase activity measurement result
The division of 2.2 sample sets
For eliminating the impact of subjective factor in sample set partition process and evaluating the estimated performance of model under the prerequisite not affecting models fitting ability, under MATLAB R2011b, KS sample partitioning is adopted to carry out the division of sample set, choose wherein 50 samples as calibration set, remaining 17 samples as checking collection.Calibration set collects the distribution of sample in major component space as Fig. 1 with checking.
The near-infrared spectrum analysis of 2.3 Medicated Leavens
2.3.1 Pretreated spectra result
Due to solid granulates uneven sized by Medicated Leaven fermented sample, the spectral information gathered includes at the bottom of many high-frequency random noises, code book, baseline wander, light scattering and the unequal noise information of sample particle size, this will have a strong impact on the extraction of data in spectrum, disturb the relation of indices near infrared absorption intensity and sample, reduce reliability and the accuracy of near infrared regression analysis equation.Adopt EMSC method first to carry out pre-service to spectrum herein, to improve the quality of spectrum, result is as Fig. 2 and Fig. 3, and as can be seen from the figure after pre-service, spectrum line is concentrated, and solves the problem of baseline wander preferably.
2.3.2 near infrared modeling
Adopt LS-SVM and PLS to set up the near-infrared model of Medicated Leaven prolease activity and amylase activity respectively, the results are shown in Table 3.
The results contrast of the different modeling method of table 3
As shown in Table 3, for proteinase and diastase, Rc and Rp of LS-SVM model is all greater than PLS model, and RMSEC and RMSEP is all less than PLS model.Therefore, the model selecting LS-SVM method to set up is as the near-infrared model of Medicated Leaven prolease activity and amylase activity.
The proteinase adopting LS-SVM method to set up and the correlationship between the measured value of amylase activity near-infrared model and predicted value are shown in Fig. 4 and Fig. 5.
The above results shows, enzyme activity forecast model of building has good predictive ability.Wherein, the RMSEP of the near-infrared model of amylase activity is less than the RMSEP of proteinase, illustrates that the estimated performance of the near-infrared model of amylase activity is more excellent.Meanwhile, model can be suitable in flour wheat bran 1:1 fermentation group and whole meal fermentation group, shows that model has good applicability to the prediction of different substrate enzyme vigor.
3. result of practical application
The enzyme activity forecast model of the near infrared spectrum that said method is set up by the present invention is applied in the quality testing of the purebred solid ferment process of Medicated Leaven, and result shows: the proteinase of Medicated Leaven and have good correlativity between amylase activity near-infrared spectral measurement and traditional measuring method analysis result.The model set up has good applicability in flour wheat bran ratio 1:1 fermentation group and whole meal fermentation group.
Claims (10)
1. the online test method of quality index in Medicated Leaven sweat, comprises the steps:
(1) sample in Medicated Leaven sweat different time points;
(2) diffuse reflectance is adopted to obtain the near infrared spectrum data of each sample; Measure prolease activity and/or the amylase activity of each sample respectively;
(3) least square method supporting vector machine method, partial least square method or neural network is adopted, preferred employing least square method supporting vector machine method sets up the near infrared spectrum data of sample and the correlationship between the near infrared spectrum data of proteinase and/or sample and amylase activity, obtains the enzyme activity forecast model based near infrared spectrum;
(4) the above-mentioned enzyme activity forecast model based near infrared spectrum is adopted to realize the on-line checkingi of quality index in Medicated Leaven sweat.
2. detection method according to claim 1, is characterized in that, in step (1), described Medicated Leaven fermentation is the pure-blood ferment adopting single bacterial classification.
3. detection method according to claim 1, it is characterized in that, in step (1), described Medicated Leaven sweat adopts the potpourri of flour and wheat bran 1:1 respectively, whole meal two kinds of nutrient culture media ferment, and from the different time points sampling described two sweats.
4. the detection method according to any one of claim 1-3, it is characterized in that, in step (1), with the 1/50-1/15 of all times that ferments, be preferably the 1/40-1/20 of fermentation all times, the 1/35-1/25 being more preferably fermentation all times is sampling interval time; Each sampling amount is 1-20, is preferably 1-15, is more preferably 1-10, then is preferably 1-8.
5. detection method according to claim 1, is characterized in that, in step (2), the wavelength coverage of described near infrared spectrum is selected from 400-2500nm wavelength coverage.
6. detection method according to claim 5, is characterized in that, in step (2), adopts the prolease activity of Folin-phenol method working sample, adopts the amylase activity of DNS method working sample.
7. detection method according to claim 1, is characterized in that, in step (3), near infrared spectrum data adopts MATLAB R2011b analysis software to process, and adopts expansion multiplicative scatter correction method to carry out pre-service to spectroscopic data.
8. detection method according to claim 7, it is characterized in that, in step (3), the sample of sampling is subdivided into calibration set to collect with checking, using the coefficient R c of the chemical measurements of calibration set sample and model predication value with corrects mean square deviation as the fitting effect of metrics evaluation model and inner robustness, to verify the coefficient R p of chemical measurements and the model predication value integrating sample and to predict the estimated performance of mean square deviation as metrics evaluation model.
9. detection method according to claim 8, is characterized in that, adopts Kennard-Stone sample partitioning sampling sample to be subdivided into calibration set and collects with checking.
10. detection method according to claim 9, it is characterized in that, in step (3), 5 folding cross-validation methods, 10 folding cross-validation methods or 20 folding cross-validation methods can be adopted, preferred employing 10 folding cross-validation method, the best latent variable of preference pattern is because of subnumber.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510002633.0A CN104596984A (en) | 2015-01-05 | 2015-01-05 | Method for detecting medicated leaven fermentation process quality on line by using near infrared spectrum |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510002633.0A CN104596984A (en) | 2015-01-05 | 2015-01-05 | Method for detecting medicated leaven fermentation process quality on line by using near infrared spectrum |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104596984A true CN104596984A (en) | 2015-05-06 |
Family
ID=53122905
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510002633.0A Pending CN104596984A (en) | 2015-01-05 | 2015-01-05 | Method for detecting medicated leaven fermentation process quality on line by using near infrared spectrum |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104596984A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106290240A (en) * | 2016-08-29 | 2017-01-04 | 江苏大学 | A kind of method based on near-infrared spectral analysis technology to Yeast Growth curve determination |
| CN106323891A (en) * | 2016-08-25 | 2017-01-11 | 广西科技大学 | Method of quick analysis of azo dyes synthesis process |
| CN106770005A (en) * | 2016-11-25 | 2017-05-31 | 山东大学 | A kind of division methods of the calibration set for near-infrared spectrum analysis and checking collection |
| CN107515204A (en) * | 2017-10-19 | 2017-12-26 | 西华大学 | Detection method using NIR to bean paste sweet tea valve fermenting-ripening degree |
| CN107907540A (en) * | 2017-12-26 | 2018-04-13 | 安徽省食品药品检验研究院 | The discrimination method of the fermented tcm Medicated Leaven true and false |
| CN109242224A (en) * | 2018-11-29 | 2019-01-18 | 杭州电子科技大学 | A kind of novel chemical procedure quality prediction technique |
| CN110320174A (en) * | 2019-06-14 | 2019-10-11 | 湖北省农业科学院果树茶叶研究所 | Using the method for polynomial net structure artificial neural network quick predict Yuanan yellow tea bored yellow time |
| CN112903627A (en) * | 2021-03-06 | 2021-06-04 | 中国烟草总公司郑州烟草研究院 | Method for online determination of biological enzyme activity in tobacco processing process |
| CN113588590A (en) * | 2021-08-11 | 2021-11-02 | 苏州泽达兴邦医药科技有限公司 | Quality control method for traditional Chinese medicine extraction process based on data mining |
| CN114199817A (en) * | 2021-12-02 | 2022-03-18 | 广东一方制药有限公司 | Construction method of near-infrared identification model of cocklebur fruit medicinal material and near-infrared identification method thereof |
| CN116008441A (en) * | 2023-03-24 | 2023-04-25 | 山东省中医药研究院 | Quality evaluation method and application of fermented medicated leaven |
| CN116798506A (en) * | 2023-03-23 | 2023-09-22 | 江苏大学 | Method for predicting thallus concentration in pichia pastoris fermentation process |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101413885A (en) * | 2008-11-28 | 2009-04-22 | 中国农业科学院蜜蜂研究所 | Near-infrared spectrum method for rapidly quantifying honey quality |
-
2015
- 2015-01-05 CN CN201510002633.0A patent/CN104596984A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101413885A (en) * | 2008-11-28 | 2009-04-22 | 中国农业科学院蜜蜂研究所 | Near-infrared spectrum method for rapidly quantifying honey quality |
Non-Patent Citations (2)
| Title |
|---|
| JOSÉ ALVES-RAUSCHA等: "Real time in-line monitoring of large scale Bacillus fermentations with near-infrared spectroscopy", 《JOURNAL OF BIOTECHNOLOGY》 * |
| 徐云等: "六神曲发酵过程中5种消化酶的动态分析", 《中国酿造》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106323891A (en) * | 2016-08-25 | 2017-01-11 | 广西科技大学 | Method of quick analysis of azo dyes synthesis process |
| CN106323891B (en) * | 2016-08-25 | 2019-05-17 | 广西科技大学 | Azo dyes synthesis process rapid analysis method |
| CN106290240A (en) * | 2016-08-29 | 2017-01-04 | 江苏大学 | A kind of method based on near-infrared spectral analysis technology to Yeast Growth curve determination |
| CN106770005A (en) * | 2016-11-25 | 2017-05-31 | 山东大学 | A kind of division methods of the calibration set for near-infrared spectrum analysis and checking collection |
| CN107515204A (en) * | 2017-10-19 | 2017-12-26 | 西华大学 | Detection method using NIR to bean paste sweet tea valve fermenting-ripening degree |
| CN107907540A (en) * | 2017-12-26 | 2018-04-13 | 安徽省食品药品检验研究院 | The discrimination method of the fermented tcm Medicated Leaven true and false |
| CN109242224A (en) * | 2018-11-29 | 2019-01-18 | 杭州电子科技大学 | A kind of novel chemical procedure quality prediction technique |
| CN110320174A (en) * | 2019-06-14 | 2019-10-11 | 湖北省农业科学院果树茶叶研究所 | Using the method for polynomial net structure artificial neural network quick predict Yuanan yellow tea bored yellow time |
| CN112903627A (en) * | 2021-03-06 | 2021-06-04 | 中国烟草总公司郑州烟草研究院 | Method for online determination of biological enzyme activity in tobacco processing process |
| CN112903627B (en) * | 2021-03-06 | 2023-01-24 | 中国烟草总公司郑州烟草研究院 | Method for online determination of biological enzyme activity in tobacco processing process |
| CN113588590A (en) * | 2021-08-11 | 2021-11-02 | 苏州泽达兴邦医药科技有限公司 | Quality control method for traditional Chinese medicine extraction process based on data mining |
| CN113588590B (en) * | 2021-08-11 | 2024-04-16 | 苏州泽达兴邦医药科技有限公司 | Traditional Chinese medicine extraction process quality control method based on data mining |
| CN114199817A (en) * | 2021-12-02 | 2022-03-18 | 广东一方制药有限公司 | Construction method of near-infrared identification model of cocklebur fruit medicinal material and near-infrared identification method thereof |
| CN114199817B (en) * | 2021-12-02 | 2024-05-14 | 广东一方制药有限公司 | Construction method of near infrared identification model of cocklebur fruit medicinal material and near infrared identification method thereof |
| CN116798506A (en) * | 2023-03-23 | 2023-09-22 | 江苏大学 | Method for predicting thallus concentration in pichia pastoris fermentation process |
| CN116798506B (en) * | 2023-03-23 | 2024-03-22 | 江苏大学 | Method for predicting thallus concentration in pichia pastoris fermentation process |
| CN116008441A (en) * | 2023-03-24 | 2023-04-25 | 山东省中医药研究院 | Quality evaluation method and application of fermented medicated leaven |
| CN116008441B (en) * | 2023-03-24 | 2023-06-20 | 山东省中医药研究院 | Quality evaluation method and application of fermented medicated leaven |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104596984A (en) | Method for detecting medicated leaven fermentation process quality on line by using near infrared spectrum | |
| CN103308463B (en) | Characteristic spectrum area selection method for near infrared spectrum | |
| CN103528990B (en) | A kind of multi-model Modeling Method of near infrared spectrum | |
| CN104949936A (en) | Sample component determination method based on optimizing partial least squares regression model | |
| CN111007040B (en) | Near infrared spectrum rapid evaluation method for rice taste quality | |
| Peng et al. | Monitoring of alcohol strength and titratable acidity of apple wine during fermentation using near-infrared spectroscopy | |
| CN104297203B (en) | A kind of Quick method of congou tea fermented quality based on near-infrared spectral analysis technology | |
| CN101655454A (en) | Rapid determination method for evaluation of storage quality of grain | |
| CN103278473B (en) | The mensuration of pipering and moisture and method for evaluating quality in white pepper | |
| CN102313712B (en) | Correction method of difference between near-infrared spectrums with different light-splitting modes based on fiber material | |
| CN111044516B (en) | Remote sensing estimation method for chlorophyll content of rice | |
| CN103344597B (en) | Anti-flavored-interference near infrared non-destructive testing method for internal components of lotus roots | |
| Arazuri et al. | Rheological parameters determination using Near Infrared technology in whole wheat grain | |
| CN105136737A (en) | Method for fast measuring content of potato flour in steamed buns based on near infrared spectrums | |
| CN110569605B (en) | NSGA 2-ELM-based non-glutinous rice leaf nitrogen content inversion model method | |
| CN101762569A (en) | Non-destructive monitoring method of livestock excrement industrialized composting fermentation process | |
| Wang et al. | Rapid detection of protein content in rice based on Raman and near-infrared spectroscopy fusion strategy combined with characteristic wavelength selection | |
| Jiang et al. | Qualitative and quantitative analysis in solid-state fermentation of protein feed by FT-NIR spectroscopy integrated with multivariate data analysis | |
| CN106290240A (en) | A kind of method based on near-infrared spectral analysis technology to Yeast Growth curve determination | |
| Huan et al. | Variable selection in near-infrared spectra: Application to quantitative non-destructive determination of protein content in wheat | |
| Chen et al. | Hyperspectral detection of sugar content for sugar-sweetened apples based on sample grouping and SPA feature selecting methods | |
| Grassi et al. | Interval ANOVA simultaneous component analysis (i-ASCA) applied to spectroscopic data to study the effect of fundamental fermentation variables in beer fermentation metabolites | |
| CN104155262A (en) | Method for selecting spectrum scope in tobacco water-soluble sugar near infrared quantification model | |
| CN112945881B (en) | Method for monitoring water content of potato leaves based on hyperspectral characteristic parameters | |
| Hongqiang et al. | Near-infrared spectroscopy with a fiber-optic probe for state variables determination in solid-state fermentation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150506 |
|
| RJ01 | Rejection of invention patent application after publication |