CN114201853A - Method for analyzing relevance of corn starch process parameters and starch milk DE values - Google Patents

Abstract

The embodiment of the invention provides a method for analyzing the relevance of corn starch process parameters and a starch milk DE value, which belongs to the field of process modeling of big data and comprises the following steps: performing feature extraction processing and dimension reduction processing on the initial input data by using a principal component analysis method to obtain a feature vector of the initial input data and principal component input data of the initial input data; training the principal component input data by using a convolutional neural network model, and calculating a total weight matrix of the trained convolutional neural network model; sequentially combining the characteristic vectors of the initial input data into a matrix as a weight matrix of the initial input data; and taking a matrix obtained by multiplying the total weight matrix of the convolutional neural network model by the weight matrix of the initial input data as an association matrix corresponding to the DE value of the corn starch process site and the starch milk product. According to the method, a process adjustment scheme can be quickly formed for different raw materials according to the correlation condition of the corn starch process site and the DE value of the starch milk product, and the DE value of the starch milk is improved.

Description

Correlation analysis method of corn starch process parameters and starch milk DE value

technical field

The invention relates to the field of process modeling of big data, in particular to a method for analyzing the correlation degree between corn starch process parameters and starch milk DE value.

Background technique

The corn starch industry is the basic industry of carbohydrate derivatives, and the corn starch in its products is an important raw material for food processing. In terms of national life, economy and agricultural industrialization development, the starch industry in food processing is an important part of my country's food industry and one of the most important links in corn deep processing.

Different sources of corn raw materials have different appropriate processing parameters during processing. With the refinement of management, the data collected by corn deep processing enterprises in the production process is becoming more and more complex, and the amount of data is increasing explosively, and the data are all time series data. Faced with massive data records, it is difficult for production managers to accurately analyze the correlation between data only by relying on production experience. In the corn starch process, an important indicator that reflects the quality of starch milk products is the DE value, which represents the percentage of reducing sugar in the dry matter of the syrup; how to use the historical process data to dig out the process parameters that have the greatest impact on the DE value is an urgent issue for us. technical issues to be resolved.

SUMMARY OF THE INVENTION

The purpose of the embodiment of the present invention is to provide a method for analyzing the correlation degree of corn starch process parameters and starch milk DE value, mainly to solve the correlation degree between corn starch process parameters and starch milk DE value, so as to realize the improvement of the DE value of starch milk. Process adjustment program.

In order to achieve the above purpose, a method for analyzing the correlation between corn starch process parameters and starch milk DE value, comprising:

Obtaining corn feedstock data associated with sites in the corn starch process and production monitoring raw data in the corn starch process together form initial input data;

Using the principal component analysis method to perform feature extraction processing on the initial input data to obtain the feature vector of the initial input data; and using the principal component analysis method to reduce the dimension of the initial input data to obtain the principal component input data of the initial input data;

The principal component input data is trained using the convolutional neural network model, and the total weight matrix of the convolutional neural network model after training is calculated; and the matrix that the eigenvectors of the initial input data are sequentially combined into is used as the initial input The weight matrix of the data;

Multiply the total weight matrix of the convolutional neural network model with the weight matrix of the initial input data, and use the obtained matrix as the correlation between the corn starch process site and the starch milk product DE value in the corn starch process. degree matrix.

Preferably, the acquisition of the corn raw material data associated with the site in the corn starch process and the production monitoring raw data in the corn starch process forms initial input data, including:

Denoising the corn raw material data associated with the site in the corn starch process and the production monitoring raw data in the corn starch process to obtain denoised data;

Feature selection is performed on the denoised data using Lasso regression analysis, and the data after feature selection is used as the initial input data.

Preferably, the principal component analysis method is used to perform feature extraction processing on the initial input data to obtain the feature vector of the initial input data, including:

Standardize the initial input data to obtain a standardized matrix Z, and use the standardized matrix Z to obtain a correlation coefficient matrix R between the standardized matrix Z and the transposed matrix of the standardized matrix Z;

作为初始输入数据的特征向量。Use the eigen equation of the correlation coefficient matrix R to get the unit eigenvector

Feature vector as initial input data.

Preferably, the principal component analysis method is used to perform dimensionality reduction processing on the initial input data to obtain the principal component input data of the initial input data, including:

运用关联公式得到的矩阵作为初始输入数据的主成分输入数据；所述关联公式为：The normalized matrix Z and the unit eigenvector

The matrix obtained by using the correlation formula is used as the principal component input data of the initial input data; the correlation formula is:

j＝1,2,...,m；m是主成分个数。

j=1,2,...,m; m is the number of principal components.

Preferably, the matrix formed by sequentially combining the eigenvectors of the initial input data is used as the weight matrix of the initial input data, including:

依次作为列向量，得到的矩阵作为初始输入数据的权值矩阵W_p：the unit eigenvector

As a column vector in turn, the resulting matrix is used as the weight matrix W _p of the initial input data:

m是主成分个数；p是特征位点的个数。

m is the number of principal components; p is the number of feature sites.

Preferably, the convolutional neural network model includes a plurality of one-dimensional convolution kernels and a fully connected neural network; the principal component input data includes a training set and a validation set;

The principal component input data is trained using a convolutional neural network model, including:

Using multiple one-dimensional convolution kernels in the training set to perform two convolution operations on the principal component input data, the data is input to the fully connected neural network for iterative training until the mean square errors of the training set and the validation set converge, and the training is completed.

Preferably, the calculation of the total weight matrix of the trained convolutional neural network model includes:

1) Add the weights of the same feature in the convolutional layer at different times to obtain the weights of the convolutional layer w _jc :

Among them, s is the number of convolution kernels; p is the size of the convolution kernel; n is the size of the convolution kernel c after the convolution operation; wh is the weight coefficient in the convolution kernel; wq is the fully connected layer. The weight coefficient of : wq=wq ₁ wq ₂ ,...,wq _a ; where a is the number of layers of the fully connected layer;

2) A matrix formed by combining the weights w _jc of the convolutional layer is used as the total weight matrix W _c of the convolutional neural network model: W _c =(w _1c , . . . , w _jc ).

Preferably, the correlation matrix W corresponding to the corn starch process site and the DE value of the starch milk product is:

W _{= WpWc} .

Preferably, the correlation analysis method further includes:

Sort the absolute values of the elements in the correlation matrix corresponding to the corn starch process site and the DE value of the starch milk product to obtain a sorting sequence; determine the difference between the corn starch process site and the DE value according to the sequence of the sorting sequence. influence level.

Preferably, the degree of influence includes positive correlation and negative correlation;

When the element in the correlation matrix corresponding to the corn starch processing site and the starch milk product DE value is negative, it is determined that the influence degree of the corn starch processing site corresponding to the element on the starch milk DE value is negative correlation;

When the element in the correlation degree matrix corresponding to the corn starch processing site and the starch milk product DE value is positive, it is determined that the influence degree of the corn starch processing site corresponding to the element on the starch milk DE value is positive correlation.

Through the above technical solutions, the principal component analysis method PCA reduces the dimensionality of the characteristic sites of the corn starch process data, which solves the problem of the dimensional disaster of the data to a certain extent; the convolutional neural network CNN model also captures the temporal features and solves the problem of the data. The problem of time series inconsistency makes the big data modeling more accurate and reliable. Finally, the weight connection method WCM can intuitively and easily understand the relationship between the feature sites that affect the DE value of starch milk products. The raw materials can quickly form a process adjustment plan to improve the DE value of starch milk products.

Other features and advantages of embodiments of the present invention will be described in detail in the detailed description section that follows.

Description of drawings

The accompanying drawings are used to provide a further understanding of the embodiments of the present invention, and constitute a part of the specification, and are used to explain the embodiments of the present invention together with the following specific embodiments, but do not constitute limitations to the embodiments of the present invention. In the attached image:

Fig. 1 is the correlation analysis method schematic diagram of a kind of corn starch technological parameter and starch milk DE value provided in the embodiment;

2 is a schematic diagram of the contribution rate and cumulative contribution rate of the principal components provided in the embodiment;

Fig. 3 is a kind of convolutional neural network model schematic diagram provided in the embodiment;

Fig. 4 is the training result schematic diagram of the convolutional neural network model provided in the embodiment;

Fig. 5 is the DE value detection diagram of starch milk products before and after adjusting the parameters corresponding to the site determined by the present invention provided in the Examples.

Detailed ways

The specific implementations of the embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific implementation manners described herein are only used to illustrate and explain the embodiments of the present invention, and are not used to limit the embodiments of the present invention.

Example 1

As shown in Figure 1, the correlation analysis method of a kind of corn starch technological parameter and starch milk DE value that the present embodiment provides, comprises:

Obtaining corn feedstock data associated with sites in the corn starch process and production monitoring raw data in the corn starch process together form initial input data;

Use the principal component analysis method to perform feature extraction processing on the initial input data to obtain the feature vector of the initial input data; and use the principal component analysis method to reduce the dimension of the initial input data to obtain the principal component input data of the initial input data; Perform feature extraction and dimensionality reduction on the initial input data once;

The principal component input data is trained using the convolutional neural network model, and the total weight matrix of the convolutional neural network model after training is calculated; and the matrix that the eigenvectors of the initial input data are sequentially combined into is used as the initial input The weight matrix of the data; the machine learning through the convolutional neural network model mainly uses the convolution kernel for secondary feature extraction and then performs machine learning;

The total weight matrix of the convolutional neural network is multiplied by the weight matrix of the initial input data, and the obtained matrix is used as the correlation degree matrix corresponding to the corn starch process site and the DE value of the starch milk product. According to the correlation degree matrix, the degree of influence can be determined, so that the parameters of the characteristic sites with high degree of influence can be used to optimize the production process parameters and improve the DE value of starch milk products. Among them, the DE value reflects the quality of starch milk products in the corn starch process.

Preferably, the acquisition of the corn raw material data associated with the site in the corn starch process and the production monitoring raw data in the corn starch process forms initial data; the initial data is matrix data, and the matrix data is listed as one of the corn raw material data. The physical and chemical data and the processing technology data in the raw data of production monitoring in the corn starch process do not require an order; behavior time (according to the time order, the interval is 10min), and then the feature selection of the initial data is carried out. Specifically, the corn raw material data associated with the site in the corn starch process and the production monitoring raw data in the corn starch process are denoised to obtain denoised data;

Feature selection is performed on the denoised data using Lasso regression analysis, and the data after feature selection is used as the initial input data.

Specifically, the specific method of denoising processing may adopt one or more data processing methods among adjusting residence time, removing outliers, supplementing missing values, removing noise and reducing feature dimension. The specific methods of various data processing methods are shown in the steps 1.1)-1.4) as follows:

1.1) Considering that the same batch of materials will have a long residence time in different processes, simply intercepting the process time recording point at the same time will cause the material batch to be dislocated, and the "current" input parameters are used to correlate and be affected. The "Past" input affects the target parameter. Therefore, based on the original corn data, the process data of the later stage is correlated according to the residence time: the starch process data and starch detection data time is 48h (corn soaking), the saccharification section time is 50h (retention in the liquefaction column), and the OCC data time of fructose For 98h (saccharification time). In this way, it is ensured that each column of data in the time row of the input data matrix corresponds to the same batch of corn raw materials, and the relative consistency of data information is ensured.

1.2) Due to the stability of the sensor of the process equipment, the recorded process data may have abnormal values. Identify outliers in the process and delete all site data at the time corresponding to the data. For example, if the data in the j column (W site) of the data matrix is abnormal at the ith (T time), then delete the i row in the data matrix. The judgment of outliers is preferably to use the Chauville method to identify the outliers of the data; the discriminant formula of the Chauville method is:

ω _n =1+0.4ln(n) (2)

为数据的均值，S_x为数据的标准差，ω_n为肖维勒系数，n为数据点个数；运用(1)式，可以确定数据中的异常值。in,

is the mean value of the data, S _x is the standard deviation of the data, ω _n is the Chauville coefficient, and n is the number of data points; using formula (1), the outliers in the data can be determined.

1.3) Since the time interval of the corn raw material recording data is greater than the set value, it is different from the data of the subsequent corn starch process operation, and the corn raw material data needs to be filled to ensure the stability of the data. The vacant data is filled by interpolation method, as shown in formula (3-4):

x _{t_i} = x _{t_n} +k*(in) (4)

Among them, x _{t_n} and x _{t_m} are the known data before and after the missing value; K is the newly added vacancy data.

1.4) Due to the limitations of process operation stability and detection technology, there will be a lot of noise in the collected data, which greatly affects the model training results. Therefore, it is necessary to perform noise reduction processing on the original process data. After processing the outliers of the data matrix, the process does not need to judge the site data, and denoises all the data in the data matrix by column. The "(2n+1 point) simple moving average" method is used to smooth the data for noise reduction, as shown in formula (5):

Among them, x' _i is the data after noise reduction, and x _i+1 is the data of n before and after _xi as the center.

The above-mentioned feature selection of the denoised data by using the Lasso regression analysis method is to select meaningful features for the next step of model training. The definition function of Lasso regression analysis method is shown in formula (6):

Among them, λ is a non-negative regular parameter, which controls the complexity of the model; β is the regression coefficient vector; n is the number of features. The larger the λ, the greater the penalty for the linear model with more features, so as to obtain a model with fewer features. This model selects λ=0.01.

The data after feature selection is standardized and transformed to obtain a standardized matrix Z, and the standardized matrix Z is used to obtain a correlation coefficient matrix R between the standardized matrix Z and the transposed matrix of the standardized matrix Z; the formula of the specific standardized transformation (7 ):

为特征位点数据的平均值，s为特征位点数据的方差。Z_ij表示标准化矩阵Z中的元素，i和j为对应的行和列索引。Among them, n is the number of time dimension data, p is the number of feature sites,

is the mean of the feature site data, and s is the variance of the feature site data. Z _ij represents the elements in the normalized matrix Z, and i and j are the corresponding row and column indices.

作为初始输入数据的特征向量。具体的，相关系数矩阵的特征方程为：Use the eigen equation of the correlation coefficient matrix R to get the unit eigenvector

Feature vector as initial input data. Specifically, the characteristic equation of the correlation coefficient matrix is:

|R-λI _p |=0 (8)

Solve to get p characteristic roots λ;

Use formula (9-10) to solve to get the unit eigenvector

Rb=λj b, _j =1,2,...,m (10)

Among them, m is the determined number of principal components, and the contribution rate K _j and cumulative contribution rate S _m of the principal components are shown in Fig. 2 .

Preferably, the principal component analysis method is used to perform dimensionality reduction processing on the initial input data to obtain the principal component input data of the initial input data, including:

运用关联公式得到的矩阵作为初始输入数据的主成分输入数据；所述关联公式为：The normalized matrix Z and the unit eigenvector

The matrix obtained by using the correlation formula is used as the principal component input data of the initial input data; the correlation formula is:

j＝1,2,...,m；m是主成分个数；Z_i是标准化矩阵Z的行向量。

j=1,2,...,m; m is the number of principal components; Z _i is the row vector of the normalized matrix Z.

Preferably, the matrix formed by sequentially combining the eigenvectors of the initial input data is used as the weight matrix of the initial input data, including:

依次作为列向量，得到的矩阵作为初始输入数据的权值矩阵W_p：the unit eigenvector

As a column vector in turn, the resulting matrix is used as the weight matrix W _p of the initial input data:

m是主成分个数；p是特征位点的个数。

m is the number of principal components; p is the number of feature sites.

Preferably, as shown in FIG. 3 , the convolutional neural network model includes a plurality of one-dimensional convolution kernels and a fully connected neural network; the principal component input data includes a training set and a validation set;

The principal component input data is trained using a convolutional neural network model, including:

First, multiple one-dimensional convolution kernels are used to perform convolution operations on the principal component input data, and then input to the fully connected layer neural network. The iterative training of convolutional and neural networks is performed for this process until the mean squared errors of the training set and the validation set converge, and the training is completed. The calculation of the mean square error is shown in formula (12). As shown in Figure 4, the mean square errors of both the training set and the validation set are consistently close to zero, so the model is trained well without overfitting.

Among them, y _i is the real value, and y′ _i is the predicted value.

The reason for using a one-dimensional convolution kernel is to capture the impact of time series on the corn starch process, and the relative position characteristics between the site data are not helpful for the knowledge learning of this process data, otherwise the machine learning model will mine the characteristics of the data too much. Performing the training of the fully connected neural network model and mapping the learned "distributed feature representation" to the sample label space is equivalent to linking the features extracted by the convolution operation with the target, which is obtained through continuous iterative training and learning. patterns in the data.

Preferably, the calculation of the total weight matrix of the trained convolutional neural network model includes:

1) Add the weights of the same feature in the convolutional layer at different times to obtain the weights of the convolutional layer w _jc :

Among them, s is the number of convolution kernels; p is the size of the convolution kernel; n is the size of the convolution kernel c after the convolution operation; wh is the weight coefficient in the convolution kernel; wq is the fully connected layer. The weight coefficient of : wq=wq ₁ wq ₂ ,...,wq _a ; where a is the number of layers of the fully connected layer;

2) A matrix formed by combining the weights w _jc of the convolutional layer is used as the total weight matrix W _c of the convolutional neural network model: W _c =(w _1c , . . . , w _jc ).

Preferably, the correlation matrix W corresponding to the corn starch process site and the DE value of the starch milk product is: W=W _p W _c .

Preferably, the correlation analysis method further includes:

Sort the absolute values of the elements in the correlation matrix corresponding to the corn starch process site and the DE value of the starch milk product to obtain a sorting sequence; determine the difference between the corn starch process site and the DE value according to the sequence of the sorting sequence. influence level.

Preferably, the degree of influence includes positive correlation and negative correlation;

When the element in the correlation matrix corresponding to the corn starch processing site and the starch milk product DE value is negative, it is determined that the influence degree of the corn starch processing site corresponding to the element on the starch milk DE value is negative correlation;

When the element in the correlation degree matrix corresponding to the corn starch processing site and the starch milk product DE value is positive, it is determined that the influence degree of the corn starch processing site corresponding to the element on the starch milk DE value is positive correlation.

Sorting the absolute values of the elements in the total weight matrix W _c , the characteristic sites that have a greater impact on the DE value target of starch milk products can be obtained, which can provide certain guidance for the next step of process parameter tuning. The results are as follows in Table 1 shown.

Table 1: Correlation table of feature site tag number and DE value

characteristic site Correlation with DE value LIA2103_7 negative correlation LIA2103_3 positive correlation LIA2103_5 negative correlation LIA2103_8 positive correlation LIA2103_1 positive correlation PID1\LIC1401_2_5-PV positive correlation LEVEL1\LIA_1473_1 positive correlation LEVEL2\LIA_1590 negative correlation PID1\LIC_302A-PV negative correlation PID1\LIC_502A-PV positive correlation PID1\LIC1401_2_12-PV negative correlation PID1\LIC_501A-PV negative correlation PID1\LIC_301B-PV positive correlation PID1\LIC_502B-PV negative correlation LEVEL1\LIA_1401_1_5 positive correlation LEVEL1\LIA_1684 positive correlation LEVEL2\LIA_1401_3_2 negative correlation LEVEL1\LIA_1401_1_4 negative correlation LIA2103_2 positive correlation LEVEL1\LIA_1631 positive correlation

Adjust the feature sites with greater influence to optimize the production process parameters. For example, the parameters of the characteristic sites that have a large positive and negative impact on the DE value of starch milk products are adjusted by 10% respectively. The results of a certain period of time are shown in Figure 5, and the DE value has increased by 21%.

It can be seen from the above technical solutions that the present invention has the following advantages: the corn starch raw materials and process data are analyzed to obtain characteristic parameters that affect the DE value of starch milk products for different corn raw materials, and the results have good universality. sex. By adjusting the characteristic parameters of the excavation, the DE value of the product is improved, which provides a better auxiliary effect for the processing and production of corn starch.

Through the above five steps of data processing, principal component analysis, convolutional neural network modeling, model weight extraction and determination of optimal characteristic site parameters, the correlation degree between corn starch process parameters and starch milk products DE value was analyzed. Feature selection and dimensionality reduction for corn starch production data not only avoids the dimensional disaster problem of big data modeling, but also ensures the independence of input machine learning model data. At the same time, feature extraction and model training are performed twice. While mining and extracting the time series features of the process data, the results obtained by this technical solution are more reliable. This technical solution can also be used for big data mining of other processes to adjust production process parameters. Thereby optimizing product production, improving product quality, and maintaining enterprise production safety.

The optional embodiments of the embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details of the above-mentioned embodiments. A variety of simple modifications are made to the technical solution of the invention, and these simple modifications all belong to the protection scope of the embodiments of the present invention.

In addition, it should be noted that each specific technical feature described in the above-mentioned specific implementation manner may be combined in any suitable manner under the circumstance that there is no contradiction. To avoid unnecessary repetition, various possible combinations are not further described in this embodiment of the present invention.

In addition, various implementations of the embodiments of the present invention can also be combined arbitrarily, as long as they do not violate the ideas of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

Claims (10)

1. A method for analyzing the correlation degree of corn starch process parameters and a starch milk DE value is characterized by comprising the following steps:

obtaining corn raw material data associated with a site in a corn starch process and production monitoring raw data in the corn starch process to jointly form initial input data;

performing feature extraction processing on the initial input data by using a principal component analysis method to obtain a feature vector of the initial input data; performing dimensionality reduction on the initial input data by using a principal component analysis method to obtain principal component input data of the initial input data;

training the principal component input data by using a convolutional neural network model, and calculating a total weight matrix of the trained convolutional neural network model; and a matrix formed by sequentially combining the characteristic vectors of the initial input data is used as a weight matrix of the initial input data;

and multiplying the total weight matrix of the convolutional neural network model by the weight matrix of the initial input data, and taking the obtained matrix as an association matrix corresponding to the corn starch process site and the starch milk DE value.

2. The analytical method of claim 1, wherein the obtaining corn feedstock data associated with a site in a corn starch process and raw production monitoring data in the corn starch process together form initial input data comprising:

denoising corn raw material data associated with a site in a corn starch process and production monitoring original data in the corn starch process to obtain denoising data;

and selecting the characteristics of the de-noised data by using a Lasso regression analysis method, and taking the data after characteristic selection as the initial input data.

3. The analysis method according to claim 1 or 2, wherein the performing feature extraction processing on the initial input data by using a principal component analysis method to obtain a feature vector of the initial input data comprises:

carrying out standardization conversion on initial input data to obtain a standardized matrix Z, and obtaining a correlation coefficient matrix R between the standardized matrix Z and a transposed matrix of the standardized matrix Z by using the standardized matrix Z;

application correlationThe characteristic equation of the coefficient matrix R is used for obtaining a unit characteristic vector

As a feature vector of the initial input data.

4. The analysis method according to claim 3, wherein the performing the dimensionality reduction on the initial input data by using the principal component analysis method to obtain principal component input data of the initial input data comprises:

the normalized matrix Z and the unit feature vector are combined

Using a matrix obtained by a correlation formula as principal component input data of the initial input data; the correlation formula is as follows:

m is the number of principal components.

5. The analysis method according to claim 3, wherein the sequentially combining the eigenvectors of the initial input data into a matrix as the weight matrix of the initial input data comprises:

unit feature vector

Sequentially used as column vectors, and the obtained matrix is used as a weight matrix W of initial input data_p：

m is the number of main components; p is the number of characteristic sites.

6. The analytical method of claim 1, wherein the convolutional neural network model comprises a plurality of one-dimensional convolutional kernels and a fully-connected neural network; the principal component input data comprises a training set and a validation set;

the training of the principal component input data by applying the convolutional neural network model comprises the following steps:

and performing iterative training on the data input fully-connected neural network by performing two times of convolution operation on the principal component input data by utilizing a plurality of one-dimensional convolution cores in the training set until the mean square errors of the training set and the verification set are converged, and finishing the training.

7. The analysis method according to claim 6, wherein the calculating the total weight matrix of the trained convolutional neural network model comprises:

1) adding the weights of the convolution layer with the same characteristic at different time to obtain the weight w of the convolution layer_jc：

Wherein s is the number of convolution kernels; p is the size of the convolution kernel; n is the size of the convolution kernel c after the convolution operation; wh is a weight coefficient in the convolution kernel; wq is the weight coefficient in the full connection layer: wq ═ wq₁wq₂,...,wq_a(ii) a Wherein a is the number of the full connecting layers;

2) the weight w of the convolution layer_jcThe combined matrix is used as the total weight matrix W of the convolutional neural network model_c：W_c＝(w_1c,...,w_jc)。

8. The analytical method of claim 7, wherein the correlation matrix W for the corn starch process site and the DE value of the starch milk product is:

W＝W_pW_c。

9. the analytical method of claim 1, further comprising:

sequencing the absolute values of the elements in the relevance matrix corresponding to the corn starch process site and the starch milk DE value to obtain a sequencing sequence; and determining the influence degree of the corn starch process site on the DE value of the starch milk according to the sequence of the sequencing sequence.

10. The assay of claim 9, wherein the degree of influence comprises a positive correlation and a negative correlation;

when the element in the correlation matrix corresponding to the corn starch process site and the starch milk DE value is a negative value, determining that the influence degree of the corn starch process site corresponding to the element on the starch milk DE value is negative correlation;

when the element in the correlation matrix corresponding to the corn starch process site and the starch milk DE value is a positive value, determining that the influence degree of the corn starch process site corresponding to the element on the starch milk DE value is positive correlation.

Images