CN103245713B - Based on the Chinese herbal medicine place of production discrimination method of support vector machine and ion mobility spectrometry - Google Patents

Abstract

The present invention discloses a kind of Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry, the method with Ion transfer spectral analysis technology for detection means, with support vector machine analytical technology for recognition methods, mode identification technology is combined with spectrum analysis technology, provides one to be applicable to Chinese herbal medicine place of production mirror method for distinguishing.The Chinese herbal medicine place of production provided by the invention discrimination method, has advantage fast and accurately, can solve unreliable, the uncontrollable present Research of quality of China Chinese crude drug safety, to guarantee quality of medicinal material and clinical drug safety significant.

Description

Based on the Chinese herbal medicine place of production discrimination method of support vector machine and ion mobility spectrometry

Technical field

The discrimination method of a kind of Chinese herbal medicine involved in the present invention, be specifically related to a kind of Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry, mode identification technology combines with spectrum analysis technology by one specifically, with Ion transfer spectral analysis technology for detection means, the Chinese herbal medicine Quality Identification new technology being recognition methods with support vector machine analytical technology.

Background technology

Chinese medicine is the rarity of the Chinese nation.Over the past thousands of years, Chinese medicine in the medical history of the Chinese nation in occupation of irreplaceable status, for multiplying and living of the Chinese nation has made indelible contribution.Chinese crude drug mostly is the dry organ of plant, and there is medicine resource 12807 kinds in China, and ethnic drug is kind more than 4000, and Folk medicine is kind more than 7000, manages Chinese crude drug kind more than 1200, medicinal herbs most in use 600 kinds, and cultivation kind more than 200, kind more than 400 purchased by natural crude drugs.Because the physical environment of complexity, social condition and people are for the reason such as obscuring, cause the complicacy of Chinese crude drug in application aspect.

The Chinese crude drug of the same race of separate sources, the activity chemistry component contained by them is variant, and this will inevitably have influence on clinical efficacy and the quality of the Chinese patent drug (compound Chinese medicinal preparation) produced using it as raw material.In addition, along with traditional Chinese medicine prevention, treatment and the concern of health care function make application in TCM be not fully exerted, supply falls short of demand to cause Chinese crude drug, thus cause the appearance of the phenomenons such as Market of Chinese Materia Medica artificially adulterates, the medicinal material of poor quality of obscure species, the blindly bright and new good-looking appearance of pursuit medicinal material.Therefore, how to distinguish true and false quality, ensure drug quality and become the matter of great urgency.The bottleneck of the restriction modernization of Chinese medicine and internationalization process, tracing it to its cause, is exactly the quality problems of Chinese medicine.

Along with progress and the development of modern instrument isolation technics and Computerized Information Processing Tech, set up the important step that Chinese medicine modern mass appraisement system has become modernization of Chinese medicine development.The traditional Chinese medicine fingerprint Quality Control technical research carried out in conjunction with the present Research of International Plant drug assessment method, become field the most popular in traditional Chinese medicine research, control and a kind of effective means of quality assessment as traditional Chinese medicine quality, it can from raw-materially introducing a fine variety, cultivate, the specification of Chinese patent drug production technology and optimization, the formulation to target level of product quality provides omnibearing quality assurance.Chinese medicine fingerprint image adopts chromatogram or spectroscopic methodology to carry out macroscopic view, comprehensive and holistic approach to Chinese crude drug, medicine materical crude slice, semi-manufacture and preparation thereof, measure the characteristic fingerprint pattern or the image that obtain active component component colony, reflect kind and the quantity of chemical composition contained by Chinese medicine comprehensively, and then the quality of Efficient Characterization Chinese medicine.Traditional Chinese medicine fingerprint be a kind of comprehensively, quantifiable chemical identification means.So as to evaluating the authenticity of Chinese crude drug and Chinese medicine preparation semi-manufactured goods quality, Optimality and stability, " globality " and " ambiguity " is its distinguishing feature.Its not only characteristic embodiment but also Qualitive test can be used as use, the also concept of the amount of embodying, the concept of introduction volume, quantitatively and qualitative combining traditional Chinese medicine fingerprint can be made to have larger effect.At present, the traditional Chinese medicine fingerprint of centering evaluation of medical materials' quality mainly concentrates on the quality control centered by chemical composition, namely clear with molecular structure, sets up medicinal materials fingerprint, to detect based on the known active component that structure-activity relationship is clear and definite.The shortcomings such as conventional detection means has the instrument detection methods such as UV, FTIR, HPLC, HPLC-MS, GC, GC-MS, X-ray, but there is reference substance deficiency, and trace routine is loaded down with trivial details, and detection efficiency is not high.

Therefore, necessary on the basis of prior art, research relate to a kind of energy accurately, fast, the new method of comprehensive identification Chinese herbal medicine.

Summary of the invention

Goal of the invention: the object of the invention is to solve the deficiencies in the prior art, provides a kind of reference substance demand few, can accurately, fast, the Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry of the different Chinese herbal medicine of efficient qualification.

The present invention's application ion mobility spectrometry characterizes various different chemical substance with gaseous ion mobility, thus realizes the object to various chemical substance analysis detection.Ion transfer spectral analysis technology has the strong and high sensitivity of separating power, can as the separation qualitative and quantitative detection means of COMPLEX MIXED objects system, IMS can at positive and negative two kinds of MODE of operation simultaneously, different detecting patterns can be selected according to analysis thing, the trace analysis being conducive to complex sample be separated detection.Ion mobility spectrometry is that molion is at the feature transit time of drift tube per sample, detects fast minimum gas under atmospheric pressure or nearly atmospheric pressure.

Technical scheme: in order to reach above object, the technical scheme that the present invention takes is:

Based on the Chinese herbal medicine place of production discrimination method of support vector machine and ion mobility spectrometry, comprise Chinese herbal medicine pre-treatment, the structure of the information acquisition of Chinese herbal medicine Ion transfer spectrogram, spectral data analysis and SVM classifier;

(1) Chinese herbal medicine pre-treatment comprises the following steps:

Add ethanol after getting Chinese herbal medicine pulverizing, drying, adopt ultrasonic extraction or Soakage extraction method, merge extract, concentrated, solution obtains test sample after crossing 0.45 μm of filter membrane;

Preferably, described ultrasonic extraction step is: get 10g Milkvetch Root in 250mL tool plug conical flask, adds the ethanol water ultrasonic twice that 100mL concentration is 70%, each 1 hour, merges extract.Extract is evaporated near dry, residue add methyl alcohol dissolve and constant volume in 5mL measuring bottle, obtain need testing solution.Gained need testing solution is crossed sample introduction analysis after 0.45 μm of filter membrane;

Described Soakage extraction step is: get 10g Milkvetch Root in 250mL tool plug conical flask, add 100mL concentration be 70% ethanol water leave standstill 24 hours, filter Soakage extraction liquid, extract be evaporated near dry, add methyl alcohol dissolve and constant volume in 5mL measuring bottle, obtain need testing solution.Gained need testing solution is crossed sample introduction analysis after 0.45 μm of filter membrane.

(2) information acquisition of Chinese herbal medicine Ion transfer spectrogram comprises the following steps:

The test sample utilizing microsyringe to measure 5 μ L steps (1) to obtain, inject the hyaloid bottle of about 1mL, vial is placed in sample introduction place, collected specimens escaping gas, design parameter is: migration tube temperature is 100 ~ 200 DEG C, carrier gas speed is 0.4 ~ 1.0L/min, and drift gas speed is 0.6 ~ 1.0L/min;

(3) structure of spectral data analysis and SVM classifier:

Adopt the raw data matrix of method to Chinese herbal medicine ion mobility spectrometry figure of PCA to carry out Data Dimensionality Reduction, note ion mobility spectrometry sample training collection is $({\stackrel{\→}{\mathrm{\α}}}_1,l_1),({\stackrel{\→}{\mathrm{\α}}}_2,l_2),...,({\stackrel{\→}{\mathrm{\α}}}_n,l_n),\∂=\[{\stackrel{\→}{\mathrm{\α}}}_1;{\stackrel{\→}{\mathrm{\α}}}_2;...;{\stackrel{\→}{\mathrm{\α}}}_n\]$ Be ion mobility spectrometry spectrogram matrix, concrete symbol definition is as follows:

the ion mobility spectrometry spectrogram vector of i-th and a jth Chinese herbal medicine sample respectively;

L _i, l _jbe respectively with class label, i.e. i-th and a jth grade that Chinese herbal medicine sample is corresponding;

A _i, a _jbe with lagrange factor.

For reducing exceptional sample point to the negative effect of model accuracy, before setting up sorter, need to remove these exceptional sample points.The present invention takes following method to remove these exceptional sample points:

Ask for all kinds of center of a sample, in compute classes, sample is to the Euclidean distance of center of a sample, and asks its mean value, Euclidean distance is greater than to the sample point of 2.5 times of mean values, thinks exceptional sample point and removes.

SVM to solve sorter as follows:

$f\left(\stackrel{\→}{\mathrm{\α}}\right)=\mathrm{sgn}(\underset{i=1}{\overset{n}{\Σ}}{l}_i\·a_i^{*}\·k\left({\stackrel{\→}{\mathrm{\α}}}_i,\stackrel{\→}{\mathrm{\α}}\right)+b^{*})$

Wherein a test sample book, the optimum solution of following optimization problem:

$max(\underset{i=1}{\overset{n}{\Σ}}{a}_i-\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{n}{\Σ}}{a}_i\·a_j\·l_i\·l_j\·k({\stackrel{\→}{\mathrm{\α}}}_i,{\stackrel{\→}{\mathrm{\α}}}_j\left)\right)$

$\begin{array}{cc}s.t.& \underset{i=1}{\overset{n}{\Σ}}{a}_i\·l_i=0\end{array},$

a _i≥0；

corresponding vector be called support vector;

B ^*be amount of bias, choose a support vector can be in the hope of:

$b^{*}=l_i-\underset{j=1}{\overset{n}{\Σ}}{l}_j\·a_j^{*}\·k({\stackrel{\→}{\mathrm{\α}}}_i,{\stackrel{\→}{\mathrm{\α}}}_j)$

The present invention, according to the Ion transfer spectrogram characteristic of different medicinal herb components, selects different kernel functions generate different SVM to identify, as Fig. 1, available kernel function is as follows respectively:

Polynomial kernel function is:

K(x,x')＝(γ(x·x')+1) ^q

Radial basis (RBF) kernel function is:

$K(x,x^{\′})=\exp (-\mathrm{\γ}\frac{{(x-x^{\′})}^2}{{\mathrm{\σ}}^2})$

Sigmoid kernel function is:

K(x,x')＝tanh(υ(x·x')+c)。

Preferably, the above-described Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry, the ethanol described in step (1) is concentration 30 ~ 95%.

Preferably, the above-described Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry, the information acquisition of step (2) Chinese herbal medicine Ion transfer spectrogram comprises the following steps:

The test sample utilizing microsyringe to measure 5 μ L steps (1) to obtain, injects the hyaloid bottle of about 1mL, vial is placed in sample introduction place, sample gas, design parameter is: migration tube temperature is 100 DEG C, and carrier gas speed is 0.4L/min, and drift gas speed is 0.6L/min.

Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry of the present invention, can be widely used in the discriminating of different Chinese herbal medicine, as the Radix Astragali.

Beneficial effect: mode identification technology, combines with spectrum analysis technology with support vector machine analytical technology for recognition methods for detection means with Ion transfer spectral analysis technology by the present invention, provides one to be applicable to Chinese herbal medicine place of production mirror method for distinguishing.Method provided by the invention can solve unreliable, the uncontrollable present Research of quality of China's Chinese crude drug safety, has wide range of applications.

Accompanying drawing explanation

Fig. 1 is the design of graphics of SVM classifier.

Embodiment

Below in conjunction with specific embodiment, illustrate the present invention further, these embodiments should be understood only be not used in for illustration of the present invention and limit the scope of the invention, after having read the present invention, the amendment of those skilled in the art to the various equivalent form of value of the present invention has all fallen within the application's claims limited range.

Based on the Astragalus from different habitats discrimination method of support vector machine and ion mobility spectrometry, it comprises Radix Astragali pre-treatment, the structure of the information acquisition of Radix Astragali Ion transfer spectrogram, spectral data analysis and SVM classifier;

(1) Radix Astragali pre-treatment comprises the following steps:

Choose each 40 parts of the Radix Astragali sample of Gansu Province, Heilongjiang Province, Shaanxi Province and Inner Mongolia Autonomous Region, get the above 4 40 parts of each 10g of Milkvetch Root in area respectively, be placed in 250mL tool plug conical flask respectively, add ultrasonic twice of the ethanol water that 100mL volumetric concentration is 70% respectively, each 1 hour, merge extract, extract is evaporated near dry, residue add methyl alcohol dissolve and constant volume in 5mL measuring bottle, obtain need testing solution.After gained need testing solution being crossed 0.45 μm of filter membrane, obtain Radix Astragali test sample;

(2) information acquisition of Radix Astragali Ion transfer spectrogram comprises the following steps:

The Radix Astragali test sample utilizing microsyringe to measure 5 μ L steps (1) respectively to obtain, inject the hyaloid bottle of about 1mL, vial is placed in sample introduction place, sample gas, design parameter is: migration tube temperature is 100 ~ 200 DEG C, carrier gas speed is 0.4 ~ 1.0L/min, and drift gas speed is 0.6 ~ 1.0L/min;

(3) structure of spectral data analysis and SVM classifier:

Adopt the raw data matrix of method to Chinese herbal medicine ion mobility spectrometry figure of PCA to carry out Data Dimensionality Reduction, note ion mobility spectrometry sample training collection is $({\stackrel{\→}{\mathrm{\α}}}_1,l_1),({\stackrel{\→}{\mathrm{\α}}}_2,l_2),...,({\stackrel{\→}{\mathrm{\α}}}_n,l_n),\∂=\[{\stackrel{\→}{\mathrm{\α}}}_1;{\stackrel{\→}{\mathrm{\α}}}_2;...;{\stackrel{\→}{\mathrm{\α}}}_n\]$ Ion mobility spectrometry spectrogram matrix, wherein the ion mobility spectrometry spectrogram vector of i-th Chinese herbal medicine sample, l _ibe class label, the grade that namely i-th Chinese herbal medicine sample is corresponding.

Adopt the raw data matrix of method to Chinese herbal medicine ion mobility spectrometry figure of PCA to carry out Data Dimensionality Reduction, note ion mobility spectrometry sample training collection is $({\stackrel{\→}{\mathrm{\α}}}_1,l_1),({\stackrel{\→}{\mathrm{\α}}}_2,l_2),...,({\stackrel{\→}{\mathrm{\α}}}_n,l_n),\∂=\[{\stackrel{\→}{\mathrm{\α}}}_1;{\stackrel{\→}{\mathrm{\α}}}_2;...;{\stackrel{\→}{\mathrm{\α}}}_n\]$ Be ion mobility spectrometry spectrogram matrix, concrete symbol definition is as follows:

the ion mobility spectrometry spectrogram vector of i-th and a jth Chinese herbal medicine sample respectively;

L _i, l _jbe respectively with class label, i.e. i-th and a jth grade that Chinese herbal medicine sample is corresponding;

A _i, a _jbe with lagrange factor.

SVM to solve sorter as follows:

$f\left(\stackrel{\→}{\mathrm{\α}}\right)=\mathrm{sgn}(\underset{i=1}{\overset{n}{\Σ}}{l}_i\·a_i^{*}\·k\left({\stackrel{\→}{\mathrm{\α}}}_i,\stackrel{\→}{\mathrm{\α}}\right)+b^{*})$

Wherein a test sample book, the optimum solution of following optimization problem:

$max(\underset{i=1}{\overset{n}{\Σ}}{a}_i-\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{n}{\Σ}}{a}_i\·a_j\·l_i\·l_j\·k({\stackrel{\→}{\mathrm{\α}}}_i,{\stackrel{\→}{\mathrm{\α}}}_j\left)\right)$

$\begin{array}{cc}s.t.& \underset{i=1}{\overset{n}{\Σ}}{a}_i\·l_i=0\end{array},$

a _i≥0；

corresponding vector be called support vector;

B ^*be amount of bias, choose a support vector can be in the hope of:

$b^{*}=l_i-\underset{j=1}{\overset{n}{\Σ}}{l}_j\·a_j^{*}\·k({\stackrel{\→}{\mathrm{\α}}}_i,{\stackrel{\→}{\mathrm{\α}}}_j)$

According to the Ion transfer spectrogram characteristic of astragalus root components, select radial basis (RBF) kernel function:

$K(x,x^{\′})=\exp (-\mathrm{\γ}\frac{{(x-x^{\′})}^2}{{\mathrm{\σ}}^2})$

Generate the Radix Astragali sample each 10 part of SVM to Gansu Province, Heilongjiang Province, Shaanxi Province and Inner Mongolia Autonomous Region to identify.The result of qualification is as shown in table 1:

The Radix Astragali sample qualification result of table 1 Gansu Province, Heilongjiang Province, Shaanxi Province and Inner Mongolia Autonomous Region

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (4)

1. based on the Chinese herbal medicine place of production discrimination method of support vector machine and ion mobility spectrometry, it is characterized in that comprising Chinese herbal medicine pre-treatment, the structure of the information acquisition of Chinese herbal medicine Ion transfer spectrogram, spectral data analysis and SVM classifier;

(1) Chinese herbal medicine pre-treatment comprises the following steps:

Add ethanol after getting Chinese herbal medicine pulverizing, drying, adopt ultrasonic extraction or Soakage extraction method, merge extract, concentrated, solution obtains test sample after crossing 0.45 μm of filter membrane;

(2) information acquisition of Chinese herbal medicine Ion transfer spectrogram comprises the following steps:

The test sample utilizing microsyringe to measure 5 μ L steps (1) to obtain, inject the hyaloid bottle of about 1mL, vial is placed in sample introduction place, the escaping gas of collected specimens, design parameter is: migration tube temperature is 100 ~ 200 DEG C, carrier gas speed is 0.4 ~ 1.0L/min, and drift gas speed is 0.6 ~ 1.0L/min;

(3) structure of spectral data analysis and SVM classifier:

Adopt the raw data matrix of method to Chinese herbal medicine ion mobility spectrometry figure of PCA to carry out Data Dimensionality Reduction, note ion mobility spectrometry sample training collection is be ion mobility spectrometry spectrogram matrix, concrete symbol definition is as follows:

the ion mobility spectrometry spectrogram vector of i-th and a jth Chinese herbal medicine sample respectively;

L _i, l _jbe respectively with class label, i.e. i-th and a jth grade that Chinese herbal medicine sample is corresponding;

A _i, a _jbe with lagrange factor;

SVM to solve sorter as follows:

$f\left(\stackrel{\→}{\mathrm{\α}}\right)=\mathrm{sgn}(\underset{i=1}{\overset{n}{\Σ}}{l}_i\·a_i^{*}\·k({\stackrel{\→}{\mathrm{\α}}}_i,\stackrel{\→}{\mathrm{\α}})+b^{*})$

Wherein a test sample book, the optimum solution of following optimization problem:

$max(\underset{i=1}{\overset{n}{\Σ}}{a}_i-\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{n}{\Σ}}{a}_i\·a_j\·l_i\·l_j\·k({\stackrel{\→}{\mathrm{\α}}}_i,{\stackrel{\→}{\mathrm{\α}}}_j\left)\right)$

$\begin{array}{cc}s.t.& \underset{i=1}{\overset{n}{\Σ}}{a}_i\·l_i=0\end{array},$

a _i≥0；

corresponding vector be called support vector;

B ^*be amount of bias, choose a support vector can be in the hope of:

$b^{*}=l_i-\underset{j=1}{\overset{n}{\Σ}}{l}_j\·a_j^{*}\·k\left({\stackrel{\→}{\mathrm{\α}}}_i,{\stackrel{\→}{\mathrm{\α}}}_j\right)$

According to the Ion transfer spectrogram characteristic of different medicinal herb components, select different kernel functions generate different SVM to identify, available kernel function is as follows respectively:

Polynomial kernel function is:

K(x,x')＝(γ(x·x')+1) ^q

Radial basis (RBF) kernel function is:

$K(x,x^{\′})=\exp (-\mathrm{\γ}\frac{{(x-x^{\′})}^2}{{\mathrm{\σ}}^2})$

Sigmoid kernel function is:

K(x,x')＝tanh(υ(x·x')+c)。

2. the Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry according to claim 1, it is characterized in that, the ethanol described in step (1) is concentration 30 ~ 95%.

3. the Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry according to claim 1, it is characterized in that, the information acquisition of step (2) Chinese herbal medicine Ion transfer spectrogram comprises the following steps:

The test sample utilizing microsyringe to measure 5 μ L steps (1) to obtain, injects the hyaloid bottle of about 1mL, vial is placed in sample introduction place, sample gas, design parameter is: migration tube temperature is 100 DEG C, and carrier gas speed is 0.4L/min, and drift gas speed is 0.6L/min.

4. the Chinese herbal medicine place of production discrimination method based on support vector machine and ion mobility spectrometry according to claim 1, described Chinese herbal medicine is the Radix Astragali.