Cis-trans diethyl cyclomalonate nondestructive testing method based on feature screening

Abstract

The invention discloses a non-destructive testing method for cis-trans-diethyl cyclomalonate based on feature screening, and relates to the technical field of testing or analyzing materials by means of measuring chemical or physical properties of the materials. Comprising the following steps: respectively taking a plurality of mixtures formed by a cis-isomer and a trans-isomer of diethyl cyclomalonate under various mass ratios, and taking Raman characteristic peak parameters of the plurality of mixtures; based on the Raman characteristic peak parameters of the plurality of mixtures, constructing a plurality of Raman characteristic peak proportion variables between cis-trans isomers; constructing a prediction model; on the basis of the prediction model, predicting the mass ratio of cis-trans isomers of the diethyl cyclomalonate; the method has the advantages of rapidness, no damage, no need of separation, high recognition precision and the like, and is suitable for the fields of isomer component analysis, structure identification, precise quantitative detection and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of testing or analyzing materials by means of measuring the chemical or physical properties of materials, and in particular to a non-destructive detection method of cis-trans diethyl cyclopropane diester based on feature screening. Background Art

[0002] Diethyl cyclopropanediol (DIE) is an important organic synthesis intermediate widely used in pharmaceuticals, pesticides, and polymer materials. This compound exists in two stable spatial configurations: cis (cis) and trans (trans)—the cis- and trans-isomers of DIE. Although identical in molecular composition, the strain effects of the three-membered ring structure and the relative orientation of the ester groups lead to significant differences in physicochemical properties such as reactivity, thermal stability, and stereoselectivity. The cis configuration, due to the close spatial position of the two ester groups, exhibits higher reactivity and is often used to construct chiral drug precursors and highly selective functional monomers. The trans configuration, on the other hand, is more suitable for constructing rigid backbone structures, thermally stable materials, and the synthesis of some agricultural compounds due to its conformational stability and greater steric hindrance. Therefore, accurately identifying and quantitatively analyzing the ratio of these two isomers is crucial for ensuring synthetic selectivity, optimizing product performance, and improving quality control in fine chemical production.

[0003] The current analysis of the cis and trans isomers of diethyl cyclopropanediol mainly relies on chromatography or mass spectrometry. Although chromatography has good separation and quantitative capabilities, it usually requires the use of chiral chromatographic columns or complex derivatization pre-treatment, making it difficult to achieve rapid response and on-site in-situ detection. Mass spectrometry is difficult to effectively distinguish between cis and trans configurations due to the same molecular mass and extremely similar structures, and it also requires high-efficiency separation methods to achieve analysis. For example, traditional separation methods often use solvent selective crystallization technology, using acetonitrile or nitromethane to obtain pure trans or pure cis isomers, respectively. This type of method is cumbersome, time-consuming, and highly dependent on operating conditions, making it difficult to adapt to high-throughput or real-time monitoring needs.

[0004] Raman spectroscopy, a nondestructive testing method based on molecular vibrations, has shown great potential in organic molecule identification, material sorting, and pharmaceutical analysis due to its high selectivity, high resolution, and structural sensitivity. However, because the differences between cis and trans isomers in Raman spectra are often subtle, traditional methods based on single peaks or subjective empirical ratios have difficulty achieving high-precision identification and quantification. Summary of the Invention

[0005] In order to solve the above problems, the present invention provides a nondestructive detection method for cis-trans diethyl cyclopropane diester based on feature screening.

[0006] A nondestructive detection method for cis-trans diethyl cyclopropane diester based on feature screening comprises the following steps: taking multiple mixtures of cis-isomers and trans-isomers of diethyl cyclopropanediol at various mass ratios, and obtaining Raman characteristic peak parameters of the multiple mixtures; constructing a plurality of Raman characteristic peak ratio variables between cis isomers and trans isomers based on Raman characteristic peak parameters of the plurality of mixtures; the Raman characteristic peak ratio variables are used to describe the difference in Raman characteristic peaks between the cis isomers and the trans isomers; Constructing a prediction model, wherein the input of the prediction model is the Raman characteristic peak ratio variable, and the output of the prediction model is the mass ratio of the cis isomer to the trans isomer; Based on the prediction model, the mass ratio of the cis isomer and the trans isomer of diethyl cyclopropanate is predicted.

[0007] Description: The above method achieves accurate and efficient differentiation and detection of the mass ratio of cis and trans isomers by constructing a variable for the ratio of Raman characteristic peaks between isomers and establishing a prediction model. To address the problem that the spatial configuration differences between cis and trans isomers are small and difficult to distinguish using traditional methods, the characteristic peak ratio variables between isomers are used to amplify the molecular response differences and significantly improve the configuration resolution ability. The mass ratio is directly output by the prediction model, achieving accurate prediction from Raman characteristic differences to mass ratios without repeated calibration. This method is suitable for fields such as pharmaceuticals and materials science that require precise control of configuration ratios.

[0008] Note: Since the difference in the ratio of cis- and trans-diethyl cyclopropanediolate is of great significance for ensuring the selectivity of the synthesis route, optimizing product performance, and improving the quality control of fine chemical production, it is necessary to quickly distinguish these two isomers to facilitate subsequent operations.

[0009] Furthermore, the method for obtaining Raman characteristic peak parameters of multiple mixtures includes: The Raman characteristic peaks of the cis isomer and the trans isomer of diethyl cyclopropanediolate were determined respectively; According to the Raman characteristic peaks of the cis isomer and the trans isomer, Raman characteristic peak parameters of the multiple mixtures are determined, wherein the Raman characteristic peak parameters are the Raman characteristic peak areas of the cis isomer and the Raman characteristic peak areas of the trans isomer.

[0010] Note: The above method clarifies the Raman characteristic peaks of cis and trans isomers themselves and calculates the Raman characteristic peak areas, providing data support for the subsequent construction of proportional variables.

[0011] Furthermore, the method for respectively determining the Raman characteristic peaks of the cis isomer and the trans isomer of diethyl cyclopropanediolate comprises: The cis isomer and trans isomer standard samples of diethyl cyclopropanediol ester are respectively taken, and a standard Raman spectrum of the cis isomer and a standard Raman spectrum of the trans isomer are respectively obtained; the Raman characteristic peak of the cis isomer is extracted from the standard Raman spectrum of the cis isomer, and the Raman characteristic peak of the trans isomer is extracted from the standard Raman spectrum of the trans isomer.

[0012] Note: The above method uses standard samples to obtain standard Raman spectra, which ensure the accuracy and standardization of the measured data. The standard Raman spectra can truly reflect the characteristics of cis and trans isomers. Extracting Raman characteristic peaks from the standard Raman spectra makes the entire process operational and systematic.

[0013] Furthermore, the range of the multiple mass ratios is 1 to 9:10, and the Raman characteristic peak area of ​​the cis isomer and the Raman characteristic peak area of ​​the trans isomer are calculated using a Gaussian fitting method.

[0014] Note: The above range ensures that the data coverage is wide enough and representative. By using the Gaussian fitting method to calculate the Raman characteristic peak area, the peak shape can be accurately fitted, effectively reducing noise interference and measurement errors, and improving the accuracy and reliability of the characteristic peak area calculation.

[0015] Furthermore, the Raman characteristic peak ratio variable includes the original variable and the constructed variable after the original variable is combined; the original variable includes the single peak ratio, weighted combination ratio, logarithmic transformation, multi-peak product ratio, symmetry index, ratio of single peak to multi-peak weighted sum, symmetric peak and asymmetric peak combination structure, cross ratio of homologous vibration modes in Raman spectrum, and logarithmic or square root transformation of peak area.

[0016] Note: The above content lists the composition of the ratio-type variables of Raman characteristic peaks. The ratio-type variables can fully reflect the complex relationship between Raman characteristic peaks and thus capture the difference information between cis-isomers and trans-isomers.

[0017] Furthermore, the method for obtaining the constructed variables after combining the original variables includes: The original variables are fully traversed in two-dimensional and three-dimensional combination spaces to generate multiple two-dimensional combinations and multiple three-dimensional combinations; the multiple two-dimensional combinations and multiple three-dimensional combinations are used as construction variables.

[0018] Note: The above method generates constructed variables by performing a full combination traversal of the original variables in two-dimensional and three-dimensional combination spaces, which greatly expands the dimensions and combination forms of the variables. It can mine more complex and deeper correlation information between the original variables, and help to more fully capture the subtle differences in Raman characteristics between cis-isomers and trans-isomers.

[0019] Furthermore, the above method for constructing a prediction model includes: performing correlation analysis on a Raman characteristic peak area ratio variable and a plurality of cis-isomer and trans-isomer mass ratios, and selecting at least one Raman characteristic peak area ratio variable having the highest correlation with the plurality of mass ratios; At least one Raman characteristic peak area ratio variable with the highest correlation with multiple ratios is used as the input of the prediction model, and multiple cis-isomer and trans-isomer mass ratios are used as output. A machine learning model is used for model training. After the training is completed, a prediction model is constructed.

[0020] Furthermore, the prediction model adopts one of a support vector regression model, a random forest model and a neural network model.

[0021] Description: The above method first accurately selects Raman characteristic peak variables that are highly correlated with the cis / trans isomer mass ratio through correlation analysis as input, and then combines them with machine learning models for training to ensure the physical meaning of the input features and the representativeness of the data. The prediction accuracy and computational efficiency are ensured through model calculation, and a prediction model with high generalization ability and interpretability is constructed.

[0022] The beneficial effects of the present invention are as follows: the method of the present invention realizes accurate and efficient discrimination and detection of the mass ratio of cis and trans isomers by constructing Raman characteristic peak ratio variables between isomers and establishing a prediction model; in view of the problem that the spatial configuration differences between cis and trans isomers are slight and difficult to distinguish by traditional methods, the characteristic peak ratio variables between isomers are used to amplify the molecular response differences and significantly improve the configuration resolution capability; the mass ratio is directly output by the prediction model, thereby realizing accurate prediction from Raman characteristic differences to mass ratios without repeated calibration, and is suitable for fields such as pharmaceuticals and materials science that require precise control of configuration ratios; specifically, by combining construction and variable screening mechanisms, the expression capability of spatial configuration differences in Raman responses is significantly enhanced, thereby realizing rapid, separation-free identification and quantitative analysis of cis and trans isomers. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is the molecular structural formula of the cis-trans isomers of diethyl cyclopropanediolate in the embodiment of the present invention; Figure 2 This is a standard Raman spectrum of cis / trans diethyl cyclopropane diolate in the embodiment of the present invention; Figure 3 is a Raman spectrum of a mixture of cis / trans isomers in an embodiment of the present invention; Figure 4 It is the linear regression model and equation with the best linear fitting condition in the embodiment of the present invention; Figure 5This is a Raman spectrum of a mixture with unknown mass fraction in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to further illustrate the approach and effects achieved by the present invention, the technical solution of the present invention will be clearly and completely described below in conjunction with experiments.

[0025] In view of the problems described in the background technology, the present invention proposes a non-destructive detection method for cis-trans diethyl cyclopropane ester based on feature screening. By constructing area combination variables of multiple groups of Raman characteristic peaks and introducing a combination space traversal and regression screening strategy, it achieves the first accurate distinction of isomer configurations and mass fraction prediction without separation conditions, significantly improving the distinguishing detection capability of Raman technology in complex cis-trans isomer systems, and providing an efficient and reliable new solution for the configuration control and quality evaluation of pharmaceutical intermediates and fine chemicals. The following examples take the cis-isomers and trans-isomers of diethyl cyclopropane ester as an example for detailed explanation.

[0026] Example 1: A nondestructive detection method for cis-trans diethyl cyclopropane diester based on feature screening, comprising the following steps: S1. Multiple mixtures of cis-isomers and trans-isomers of diethyl cyclopropanediol are prepared at various mass ratios, and Raman characteristic peak parameters of the multiple mixtures are obtained. The structural formula of diethyl cyclopropanediol is as follows: Figure 1 As shown, it includes cis-diethyl cyclopropanediate and trans-diethyl cyclopropanediate; The method for obtaining Raman characteristic peak parameters of multiple mixtures includes: S1-1. A method for respectively determining the Raman characteristic peaks of the cis isomer and the trans isomer of diethyl cyclopropanediolate comprises: Taking standard samples of the cis isomer and trans isomer of diethyl cyclopropanediol respectively, and obtaining a standard Raman spectrum of the cis isomer and a standard Raman spectrum of the trans isomer respectively; extracting the Raman characteristic peak of the cis isomer from the standard Raman spectrum of the cis isomer, and extracting the Raman characteristic peak of the trans isomer from the standard Raman spectrum of the trans isomer; S1-2. Determining Raman characteristic peak parameters of the plurality of mixtures based on the Raman characteristic peaks of the cis isomer and the Raman characteristic peaks of the trans isomer, wherein the Raman characteristic peak parameters are the Raman characteristic peak areas of the cis isomer and the Raman characteristic peak areas of the trans isomer; For example, combined Figure 2 and Figure 3 As shown, the specific implementation process of S1-1~S1-2 includes (1)~(3): (1) Establish standard Raman spectra of cis-diethyl cyclopropionate and trans-diethyl cyclopropionate; take 1 mL of cis-diethyl cyclopropionate or trans-diethyl cyclopropionate standard sample on the surface of a clean silicon wafer, place the silicon wafer on the sample stage of the Raman spectrometer, focus the 532 nm laser on the surface of the standard sample, use a laser power of 10 mW, and set the spectrum acquisition range to 150~2000 cm -1 , use 1200 lines / mm grating to optimize the spectral resolution, set the integration time to 30s, perform 3 scans, and obtain its standard Raman spectrum. The results are as follows Figure 2 The standard Raman spectra of cis- and trans-cyclopropanediol diethyl ester are shown in ; (2) Extract the Raman characteristic peaks of cis-cyclopropanoic acid diethyl ester and trans-cyclopropanoic acid diethyl ester; extract the characteristic peaks of the obtained cis- and trans-isomer Raman spectra; combine Figure 2 As shown, select 860 cm -1 , 980 cm -1 , 1194 cm -1 As the Raman characteristic peak of diethyl cis-cyclopropanediol, 252 cm -1 , 752 cm -1 、875 cm -1 As the Raman characteristic peak of trans-cyclopropanediol diethyl ester; (3) Obtain Raman spectra and characteristic peak areas of isomer mixtures with different mass fractions; Prepare a standard sample of a mixture of cis- and trans-diethylcyclopropanediol ester with a mass fraction of 10 to 90%, collect Raman spectral data in the same manner as step (1), and calculate the characteristic peak area in step (2) by Gaussian fitting; The range of the above-mentioned multiple mass ratios is 1 to 9:10, and the Raman characteristic peak areas of the cis isomer and the Raman characteristic peak areas of the trans isomer are calculated using the Gaussian fitting method in the prior art; For example, the specific implementation process of various mass ratios is as follows: prepare standard solutions with a total mass ratio of cis-cyclopropanediol diethyl ester to trans-cyclopropanediol diethyl ester of 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, and 90%, and perform the measurement according to step (1) in the above content to obtain the Raman spectra of the samples of each mass ratio, as shown in FIG. Figure 3 As shown; and the Raman characteristic peak of diethyl cis-cyclopropanediol 860 cm -1 , 980 cm -1 、1194 cm -1 and the Raman characteristic peak of trans-cyclopropanediol diethyl ester at 252 cm -1 , 752 cm -1 、875 cm -1Gaussian fitting was performed to calculate the peak area, and the results are shown in Table 1; Table 1 Peak area calculation results under various mass ratios

[0027] S2. constructing multiple Raman characteristic peak ratio variables between cis isomers and trans isomers based on Raman characteristic peak parameters of the multiple mixtures; the Raman characteristic peak ratio variables are used to describe the differences in Raman characteristic peaks between cis isomers and trans isomers; The above-mentioned Raman characteristic peak ratio variables include original variables and constructed variables after combining the original variables; and the original variables include single peak ratio, weighted combination ratio, logarithmic transformation, multi-peak product ratio, symmetry index, ratio of single peak to multi-peak weighted sum, combination structure of symmetric peak and asymmetric peak, cross ratio of homologous vibration modes in Raman spectrum, logarithmic or square root transformation of peak area; Methods for obtaining constructed variables after combining original variables include: Perform a full combination traversal of the above original variables in the two-dimensional and three-dimensional combination spaces to generate multiple two-dimensional combinations and multiple three-dimensional combinations; use the multiple two-dimensional combinations and multiple three-dimensional combinations as construction variables; the specific operation steps include: Define the number of variables and their value sets for each dimension, traverse the values ​​of each dimension in turn, and obtain all possible combinations: Two-dimensional output ordered pairs: [(unimodal ratio, weighted combination ratio), (weighted combination ratio, logarithmic transformation)]; Three-dimensional output triplet: [(unimodal ratio, weighted combination ratio, logarithmic transformation), (unimodal ratio, weighted combination ratio, multimodal product ratio)]; What you get is a list containing all combinations, where each element is a tuple. For one dimension, it is a single-element tuple of the original variable itself. For two dimensions, it is a two-tuple formed by the ordered pairs of variables. For three dimensions, it is a triple formed by the three variable elements. Specifically, the format for selecting Raman characteristic peak area ratio variables is as follows: the original variables or constructed variables are combined into a characteristic matrix: one-dimensional (the original variable itself) is a one-dimensional matrix; two-dimensional / three-dimensional forms an n×kn×k matrix (k is 2 or 3); the complete characteristic matrix and the target vector (the mass ratio of diethyl cis-cyclopropane diolate) are input into the linear regression model for global fitting: Then calculate the R of the fitted curve 2 The values ​​are sorted and at least one Raman characteristic peak area ratio variable with the best fitting effect is screened out; for example, a multivariate linear regression model is established by the least squares method to calculate the R of each combination. 2 value and output the corresponding regression coefficient b, output multiple R 2The results of the model implementation with a linear fit of ≥ 0.99 are shown in Table 2. The proportional variables with the best linear fit are Figure 4 As shown; Table 2 Proportional variable fitting results

[0028] S3. Constructing a prediction model, wherein the input of the prediction model is the Raman characteristic peak ratio variable, and the output is the mass ratio of the cis isomer to the trans isomer of diethyl cyclopropanediol; Methods for building predictive models include: Performing correlation regression analysis on the Raman characteristic peak area ratio variables and multiple mass ratios, selecting at least one ratio variable with the highest correlation with the multiple mass ratios and the corresponding regression coefficient, and obtaining a prediction model based on multiple linear regression; In the embodiment of the present invention, the prediction model adopts a multiple linear regression model; its function form is as follows Figure 4 As shown, it is a multiple linear regression model with the best ranking; It should be understood that in some other embodiments, any one of the regression models including ordinary least squares (OLS), partial least squares regression (PLSR), principal component regression (PCR), ridge regression, lasso regression, elastic net regression, and support vector regression (SVR) may be selected for support; this is not limited in the embodiments of the present invention.

[0029] S4. Predicting the mass ratio of the cis isomer to the trans isomer of diethyl cyclopropanate based on the above prediction model; For example, a mixture of cis-isomer and trans-isomer of diethyl cyclopropane diolate to be tested is taken, and a sample of the mixture is tested according to step (1) in Example 1, and the obtained Raman spectrum characteristic peaks are analyzed, such as Figure 5 As shown; the Raman characteristic peak area was calculated by Gaussian fitting and R3=0.156, R6=1.138, R7=0.047 were calculated, and then the cis mass fraction was calculated to be 38.7% through the prediction model Y=90.1294-82.2639R3-45.4856R6+280.6175R7.

[0030] Example 2: This example differs from Example 1 in that the method for constructing the prediction model in S4 is different, specifically including: performing correlation analysis on a Raman characteristic peak area ratio variable and a plurality of cis-isomer and trans-isomer mass ratios, and selecting at least one Raman characteristic peak area ratio variable having the highest correlation with the plurality of mass ratios; At least one Raman characteristic peak area ratio variable with the highest correlation with multiple mass ratios is used as the input of the prediction model, and multiple cis-isomer and trans-isomer mass ratios are used as outputs. A machine learning model is used for model training. The model training method is the same as the existing training method. After the training is completed, a prediction model is constructed; the prediction model adopts a random forest model; in other embodiments, one of the support vector regression model, the neural network model, etc. can be used.

[0031] In summary, the present invention can achieve rapid and accurate detection of cis- and trans-isomer mixtures without sample separation. At the same time, it constructs various forms of combination variables and performs automatic traversal and screening in the combination space to establish an optimal regression model to achieve high-precision quantitative prediction of isomer mass fractions, with non-destructive and rapid in situ detection. Compared with traditional detection methods, Raman spectroscopy technology can directly analyze samples without complex pretreatment steps, thereby achieving ultra-fast and non-destructive in situ detection; it also has high-precision quantitative modeling capabilities: a linear regression model is combined with a combination variable traversal screening strategy to systematically evaluate the fitting ability of each group of features with the cis mass fraction, establish a prediction model, and perform high-sensitivity and high-accuracy quantitative prediction of unknown samples; multi-component compatibility and strong versatility: the proposed method does not rely on specific wavelengths or specific structural types, and can be extended to other molecular isomer systems with similar structures. It has good universality and promotion value and is suitable for the configuration identification and quantification of complex systems such as drug synthesis intermediates and functional material precursors.

Claims

1. A nondestructive detection method for cis-trans diethyl cyclopropane diester based on feature screening, characterized in that: The following steps are involved: taking multiple mixtures of cis-isomers and trans-isomers of diethyl cyclopropanediol at various mass ratios, and obtaining Raman characteristic peak parameters of the multiple mixtures; constructing a plurality of Raman characteristic peak ratio variables between cis isomers and trans isomers based on Raman characteristic peak parameters of the plurality of mixtures; the Raman characteristic peak ratio variables are used to describe the difference in Raman characteristic peaks between the cis isomers and the trans isomers; Constructing a prediction model, wherein the input of the prediction model is the Raman characteristic peak ratio variable, and the output of the prediction model is the mass ratio of the cis isomer to the trans isomer of diethyl cyclopropane; Based on the prediction model, the mass ratio of the cis isomer and the trans isomer of diethyl cyclopropanate is predicted.

2. A nondestructive detection method for cis-trans diethyl cyclopropane diolate based on feature screening as claimed in claim 1, characterized in that: The method for obtaining Raman characteristic peak parameters of multiple mixtures includes: The Raman characteristic peaks of the cis isomer and the trans isomer of diethyl cyclopropanediolate were determined respectively; According to the Raman characteristic peaks of the cis isomer and the trans isomer, Raman characteristic peak parameters of the multiple mixtures are determined, wherein the Raman characteristic peak parameters are the Raman characteristic peak areas of the cis isomer and the Raman characteristic peak areas of the trans isomer.

3. A nondestructive detection method for cis-trans diethyl cyclopropane diolate based on feature screening as claimed in claim 2, characterized in that: The method for respectively determining the Raman characteristic peaks of the cis isomer and the trans isomer of diethyl cyclopropanediol comprises: The cis isomer and trans isomer standard samples of diethyl cyclopropanediol ester are respectively taken, and a standard Raman spectrum of the cis isomer and a standard Raman spectrum of the trans isomer are respectively obtained; the Raman characteristic peak of the cis isomer is extracted from the standard Raman spectrum of the cis isomer, and the Raman characteristic peak of the trans isomer is extracted from the standard Raman spectrum of the trans isomer.

4. A nondestructive detection method for cis-trans diethyl cyclopropane diolate based on feature screening as claimed in claim 2, characterized in that: The range of the multiple mass ratios is 1 to 9:10, and the Raman characteristic peak area of ​​the cis isomer and the Raman characteristic peak area of ​​the trans isomer are calculated using a Gaussian fitting method.

5. A nondestructive detection method for cis-trans diethyl cyclopropane diolate based on feature screening as claimed in claim 4, characterized in that: The Raman characteristic peak ratio variable includes the original variable and the constructed variable after the original variable is combined; the original variable includes the single peak ratio, weighted combination ratio, logarithmic transformation, multi-peak product ratio, symmetry index, ratio of single peak to multi-peak weighted sum, symmetric peak and asymmetric peak combination structure, homologous vibration mode cross ratio in Raman spectrum, and logarithmic or square root transformation of peak area.

6. A nondestructive detection method for cis-trans diethyl cyclopropane diolate based on feature screening as claimed in claim 5, characterized in that: The method for obtaining the constructed variables after combining the original variables includes: The original variables are fully traversed in two-dimensional and three-dimensional combination spaces to generate multiple two-dimensional combinations and multiple three-dimensional combinations; the multiple two-dimensional combinations and multiple three-dimensional combinations are used as construction variables.

7. A nondestructive detection method for cis-trans diethyl cyclopropane diolate based on feature screening as claimed in claim 1, characterized in that: The method for constructing a prediction model comprises: performing correlation analysis on a Raman characteristic peak area ratio variable and a plurality of cis-isomer and trans-isomer mass ratios, and selecting at least one Raman characteristic peak area ratio variable having the highest correlation with the plurality of mass ratios; At least one Raman characteristic peak area ratio variable with the highest correlation with multiple mass ratios is used as the input of the prediction model, and multiple cis-isomer and trans-isomer mass ratios are used as outputs. A machine learning model is used for model training. After the training is completed, a prediction model is constructed.

8. A nondestructive detection method for cis-trans diethyl cyclopropane diolate based on feature screening as claimed in claim 1, characterized in that: The prediction model adopts one of a support vector regression model, a random forest model and a neural network model.

Images