Monomer fragrance raw material concentration prediction model construction method, prediction method, terminal and medium

Abstract

The application discloses a monomer fragrance raw material concentration prediction model construction method, a prediction method, a terminal and a medium, wherein the model construction method comprises the following steps: acquiring near-infrared spectrum data of a certain monomer fragrance raw material pure solution, a certain solvent pure solution and a series of different concentration solutions of the monomer fragrance raw material in the solvent; pre-processing all the near-infrared spectrum data, then correlating and fitting each pre-processed near-infrared spectrum with the actual solvent concentration corresponding to the near-infrared spectrum, and constructing a solvent concentration prediction model corresponding to the monomer fragrance raw material and the solvent; acquiring near-infrared spectrum of other types of monomer fragrance raw material pure solution and pre-processing the near-infrared spectrum; and obtaining a solvent concentration prediction model corresponding to other types of monomer fragrance raw material and the solvent by using a model transfer algorithm. Only the model construction based on a single type of fragrance raw material is needed, and then the concentration of other types of fragrance raw material can be quickly and more accurately predicted, so that the modeling difficulty is greatly reduced, and the modeling efficiency and operability are improved.

Description

Technical Field

[0001] This invention relates to the field of fragrance and flavor quality testing technology, and in particular to a method for constructing a model for predicting the concentration of single fragrance raw materials, a prediction method, a terminal, and a medium. Background Technology

[0002] The near-infrared spectral region was discovered in 1800, but its widespread application was limited by the technology available at the time. In the mid-1990s, the NIR spectrometer was developed. In the mid-to-late 1970s, analytical techniques developed rapidly, and NIR began to be applied to the determination of moisture, oil, and protein content in agricultural products and food. With advancements in metrology and the rapid development of instrument and computer hardware, near-infrared spectroscopy quickly evolved into an independent analytical technique. Its applications expanded from traditional agricultural and sideline product analysis to numerous fields such as petrochemicals, fine chemicals, light industry and food, environment, hygiene, polymer synthesis and processing, clinical medicine, and textiles. As a rapidly developing high-tech field in analytical chemistry, near-infrared spectroscopy has brought about another revolution in analytical techniques. In recent Pittsburgh conferences, "near-infrared" has been separated from "infrared and Raman" and listed as a separate category, thus receiving continued attention.

[0003] Near-infrared (NIR) light refers to electromagnetic waves between visible and mid-infrared light, with a wavelength range of 780-2526 nm (wavenumber 12820-3959 cm⁻¹). NIR spectroscopy is generated when molecular vibrations transition from the ground state to higher energy levels due to the anharmonicity of molecular vibrations. Different functional groups, or even the same functional group in different chemical environments, exhibit significant differences in NIR absorption wavelengths and intensities. Therefore, NIR spectroscopy provides rich structural and compositional information, making it suitable for analyzing components in natural products that have direct or indirect relationships with organic functional groups.

[0004] Infrared spectroscopy is widely used in industry due to its speed, accuracy, and non-destructive nature. Near-infrared spectroscopy primarily focuses on the overtone and combination frequency absorptions of hydrogen-containing group vibrations, containing compositional information for most types of organic compounds and providing rich information related to the chemical composition of fragrances and flavorings. Since the introduction of near-infrared spectroscopy into my country, the rapid development of computer technology, the gradual improvement of chemometric methods, and the implementation of relevant standards have actively promoted its widespread application and development in agriculture, chemical industry, medicine, tobacco, feed, and food. The emergence of near-infrared spectroscopy has provided new research ideas for the quality control of fragrances and flavorings.

[0005] Currently, the concentration testing of single fragrance raw materials mainly employs GC-MS fingerprinting or spectral analysis combined with standard curves. This method is not only time-consuming in terms of pretreatment and detection, but also places certain demands on the analysts. Furthermore, existing methods are all designed for the concentration detection of single fragrance raw materials and single solvents. However, fragrance raw materials are diverse, and solvents are also varied. Therefore, predicting the concentration of different fragrance raw materials in different solvents requires the analysis and processing of large amounts of data, resulting in a heavy workload and low efficiency. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the present invention aims to provide a method for constructing a concentration prediction model for single fragrance raw materials, a prediction method, a terminal, and a medium. Based on near-infrared spectral data of fragrances and flavorings, a corresponding concentration prediction model is constructed and the model is transferred to achieve rapid construction of various models, thereby enabling rapid determination of the concentration of various fragrance raw materials.

[0007] Firstly, a method for constructing a prediction model for the concentration of single fragrance raw materials is provided, including:

[0008] S1: Obtain near-infrared spectral data of a pure solution of a certain monomeric fragrance raw material, a pure solution of a certain solvent, and a series of solutions of the same monomeric fragrance raw material at different concentrations in the solvent;

[0009] S2: Preprocess all the acquired near-infrared spectra, and then perform correlation fitting between each preprocessed near-infrared spectrum and its corresponding actual solvent concentration to construct a solvent concentration prediction model for the corresponding monomeric fragrance raw material and solvent.

[0010] S3: Obtain the near-infrared spectrum of pure solutions of other types of monomeric fragrance raw materials and perform preprocessing;

[0011] S4: Based on the near-infrared spectral data of the pretreated pure solutions of other types of single fragrance raw materials and the pure solvent solution in step S1, the solvent concentration prediction model obtained in step S2 is transferred using the model transfer algorithm to obtain the corresponding solvent concentration prediction model for other types of single fragrance raw materials and the solvent, which is used to predict the concentration of the solvent in the solutions of other types of single fragrance raw materials, and thus obtain the concentration of other types of single fragrance raw materials.

[0012] Furthermore, the solvent includes one of water, propylene glycol, glycerol, alcohol, triacetin, and triethyl citrate.

[0013] Furthermore, the preprocessing methods include one or more combinations of sample normalization, sample centering, sample scaling, standard normal variable transformation, Gaussian filtering smoothing, Savitzky-Golay differentiation, multivariate scattering correction, and removal of uninformative variables.

[0014] Furthermore, the preprocessing method was optimized by using the cross-validation correlation coefficient and root mean square error of the cross-validation model as evaluation indicators.

[0015] Furthermore, in step S2, partial least squares method is used to correlate and fit each preprocessed near-infrared spectrum with its corresponding actual solvent concentration to construct a solvent concentration prediction model for the corresponding monomeric fragrance raw material.

[0016] Furthermore, the model transfer algorithm is the S / B (slope / intercept) model transfer algorithm.

[0017] Furthermore, it also includes:

[0018] Obtain the near-infrared spectrum of at least one solution sample of other types of monomeric fragrance raw materials at a known concentration in this solvent and preprocess it;

[0019] The solvent concentration prediction model obtained by the model transfer algorithm was validated based on the near-infrared spectrum of the pretreated solution sample with known concentration and the corresponding actual solvent concentration.

[0020] Secondly, a method for predicting the concentration of single fragrance raw materials is provided, including:

[0021] To determine the type of analyte fragrance raw material and solvent in the analyte fragrance raw material solution of unknown concentration;

[0022] Select a solvent concentration prediction model corresponding to the type of single fragrance raw material and solvent type to be tested, constructed using the single fragrance raw material concentration prediction model construction method described above.

[0023] Near-infrared spectra of the analyte monomer fragrance raw material solution with unknown concentrations were obtained and preprocessed. The solvent concentration C was then predicted using a selected solvent concentration prediction model. 溶剂 ;

[0024] The concentration C of the single fragrance raw material to be tested is obtained based on the predicted solvent concentration. 单体香原料 =1-C 溶剂 .

[0025] Thirdly, an electronic terminal is provided, including:

[0026] A memory that stores computer programs;

[0027] A processor is used to load and execute the computer program to implement the steps of the method for constructing a single fragrance ingredient concentration prediction model or the method for predicting single fragrance ingredient concentration as described above.

[0028] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the method for constructing a single fragrance raw material concentration prediction model or the method for predicting single fragrance raw material concentration as described above.

[0029] Given the hundreds or even thousands of varieties of single fragrance raw materials, existing methods of modeling each single fragrance raw material individually are not only costly and cumbersome, but also difficult to implement in practice due to future model maintenance concerns. However, the solvents typically used in the preparation of single fragrance raw materials are relatively simple. Therefore, this invention proposes a method for constructing a single fragrance raw material concentration prediction model, a prediction method, a terminal, and a medium. It transforms the traditional model establishing the relationship between near-infrared spectral data and single fragrance raw material concentration into a model establishing the relationship between near-infrared spectral data and solvent concentration. The concentration of the corresponding single fragrance raw material can be inferred by predicting the solvent concentration, greatly reducing the modeling difficulty. Moreover, considering that modeling requires the preparation of a large number of samples and the processing of a large amount of near-infrared spectral data, this invention combines model transfer technology. It only requires constructing several solvent concentration prediction models for a single type of single fragrance raw material in various solvents. The model transfer technology can then be used to construct solvent concentration prediction models for various types of single fragrance raw materials in various solvents, further greatly reducing the modeling difficulty and improving modeling efficiency and operability. Furthermore, using the above-mentioned rapid modeling technology, the concentration of various unknown single fragrance raw materials can be quickly determined, which can be used to assist in the quality control process of fragrance raw materials. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a flowchart of the method for constructing a concentration prediction model for single fragrance raw materials provided in an embodiment of the present invention;

[0032] Figure 2 This is a flowchart of the method for predicting the concentration of single fragrance raw materials provided in the embodiments of the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0034] This invention provides a method for constructing a model to predict the concentration of single fragrance raw materials, such as... Figure 1 As shown, it includes the following steps:

[0035] S1: Obtain near-infrared spectral data of a pure solution of a certain monomeric fragrance raw material, a pure solution of a certain solvent, and a series of solutions of the same monomeric fragrance raw material at different concentrations in the same solvent.

[0036] First, a single fragrance ingredient was selected, and a series of solutions of this single fragrance ingredient at different concentrations in common solvents were prepared. Then, near-infrared spectroscopy was used to scan the near-infrared spectral data of the pure solution of the single fragrance ingredient, the series of solutions of this single fragrance ingredient at different concentrations in common solvents, and the pure solutions in the corresponding solvents. The single fragrance ingredient had a certain degree of fluidity (this embodiment uses near-infrared transmission spectroscopy, so the sample needs to have a certain degree of fluidity to ensure normal sample injection and analysis); common solvents included water, propylene glycol, glycerol, alcohol, triacetyl ester, triethyl citrate, and other commonly used solvents in the industry. It should be noted that the number of solutions of different concentrations can be configured according to the modeling effect. Theoretically, the more single fragrance ingredient solutions of different concentrations prepared, the better; however, considering the actual work efficiency, the selection can be targeted based on the subsequent model construction effect.

[0037] In this embodiment and in the other embodiments described below, anhydrous ethanol is used as an example for illustration.

[0038] In this embodiment, β-ionone was selected as the experimental subject. It was diluted with anhydrous ethanol as the solvent to prepare β-ionone solutions with concentrations of 0% (anhydrous ethanol pure solution), 5%, 10%, ..., 90%, 95%, and 100% (β-ionone pure solution) using a 5% dilution gradient. The corresponding solvent (anhydrous ethanol) concentrations were 100%, 95%, ..., 10%, 5%, and 0%. Near-infrared spectroscopy was used to obtain the transmission near-infrared spectral data of the β-ionone pure solution, β-ionone solutions of different concentrations, and the anhydrous ethanol pure solution.

[0039] S2: Preprocess all the acquired near-infrared spectra, and then perform correlation fitting between each preprocessed near-infrared spectrum and its corresponding actual solvent concentration to construct a solvent concentration prediction model for the given monomeric fragrance raw material and solvent.

[0040] First, within the near-infrared spectral range, all near-infrared spectra obtained in step S1 are uniformly preprocessed. Then, using partial least squares (PLS) method, the preprocessed near-infrared spectra are correlated and fitted with their corresponding actual solvent concentrations to construct a solvent concentration prediction model for the specific fragrance raw material and solvent. The preprocessing methods include one or more combinations of sample normalization, mean averaging, sample centering, sample scaling, standard normal variable transformation, Gaussian filtering smoothing, Savitzky-Golay differentiation, multivariate scattering correction, and removal of uninformative variables. During implementation, the correlation coefficient (Q) of the solvent concentration prediction model is cross-validated. 2 The root mean square error of cross-validation (RMSECV) and other metrics were used to optimize the preprocessing method. In this embodiment, through repeated experimental calculations of the preprocessing method, the optimal preprocessing method was finally obtained as follows: smoothing the derivative using the Savitzky-Golay derivative algorithm, using average + standard deviation scaling, with Savitzky-Golay parameters of window size of 5 and derivative of order 2; the correlation coefficient (Q) of the constructed β-ionone solvent concentration prediction model in anhydrous ethanol system was cross-validated. 2 The mean square error of cross-validation (RMSECV) was 0.996, and the root mean square error of cross-validation (RMSECV) was 0.0196.

[0041] S3: Obtain the near-infrared spectrum of pure solutions of other types of monomeric fragrance raw materials and perform preprocessing.

[0042] In this embodiment, cinnamaldehyde, phenylethanol, sweet orange oil, and oakmoss extract were selected as experimental subjects. Near-infrared spectra of pure cinnamaldehyde solution, pure phenylethanol solution, pure sweet orange oil solution, and pure oakmoss extract solution were obtained by scanning with a near-infrared spectrometer and preprocessed.

[0043] S4: Based on the near-infrared spectral data of the pretreated pure solutions of other types of single fragrance raw materials and the pure solvent solution in step S1, the solvent concentration prediction model obtained in step S2 is transferred using the model transfer algorithm to obtain the corresponding solvent concentration prediction model for other types of single fragrance raw materials and the solvent, which is used to predict the concentration of the solvent in the solutions of other types of single fragrance raw materials, and thus obtain the concentration of other types of single fragrance raw materials.

[0044] When using an anhydrous alcohol concentration model based on β-ionone to predict the anhydrous alcohol concentration of other single fragrance ingredients in anhydrous alcohol solutions, a fixed systematic bias always exists, making it unsuitable for direct prediction of other single fragrance ingredient concentrations. Therefore, model transfer needs to be considered. In this embodiment, the S / B (slope / intercept) model transfer algorithm is selected to directly transfer the anhydrous alcohol concentration model based on β-ionone, which can significantly improve the model's prediction performance. When using the S / B (slope / intercept) model transfer algorithm, only the near-infrared spectra of the pure solutions of the single fragrance ingredient and solvent in the solution to be tested need to be obtained. In step S3, the near-infrared spectra of the pure solutions of cinnamaldehyde, phenylethanol, sweet orange oil, and oakmoss extract have been obtained, and in step S1, the near-infrared spectrum of the pure anhydrous alcohol solution has been obtained. Based on this, the slope and intercept values ​​can be obtained, enabling rapid model transfer and the construction of the corresponding solvent concentration prediction model for the single fragrance ingredient to be tested.

[0045] To verify the feasibility of the solvent concentration prediction model obtained using the model transfer algorithm, at least one sample of a known concentration of another type of monomeric fragrance ingredient in the solvent can be prepared. The near-infrared spectrum is then scanned and preprocessed, and the model prediction effect is verified. In this embodiment, a sample of a known concentration is selected for verification. Anhydrous ethanol is used for dilution, and 50% concentration solutions of cinnamaldehyde, phenylethanol, sweet orange oil, and oakmoss extract in anhydrous ethanol are prepared, resulting in 50% anhydrous ethanol solutions of cinnamaldehyde, phenylethanol, sweet orange oil, and oakmoss extract. 2 The root mean square error of cross-validation (RMSECV) is as follows: After model transfer, the anhydrous alcohol concentration prediction model Q used in cinnamaldehyde is... 2 The value is 0.999, and the RMSECV value is 0.0106; the anhydrous alcohol concentration prediction model Q used in phenylethanol 2 The value is 0.998, and the RMSECV value is 0.0140; the anhydrous ethanol concentration prediction model Q used in sweet orange oil 2 The value is 0.999, and the RMSECV value is 0.0046; the prediction model Q for the anhydrous ethanol concentration used in oakmoss extract. 2 The value is 0.999, and the RMSECV value is 0.0102.

[0046] The above embodiments illustrate the process of constructing a concentration prediction model for single fragrance raw materials using anhydrous ethanol as an example. The principle remains the same when other solvents are selected (such as water, propylene glycol, glycerol, triacetyl triacetate, triethyl citrate, etc.). A solvent concentration prediction model for that single fragrance raw material in that solvent is constructed, and then a model transfer algorithm is used to obtain solvent concentration prediction models for other types of single fragrance raw materials in the same solvent. The specific process can be found in the aforementioned embodiments and will not be repeated here. It is evident that by constructing only a few solvent prediction models for one single fragrance raw material in several solvents, hundreds or thousands of solvent prediction models for hundreds or thousands of single fragrance raw materials in several solvents can be obtained through model transfer.

[0047] Given the hundreds or even thousands of varieties of single fragrance raw materials, the existing method of modeling each single fragrance raw material individually is not only costly and cumbersome, but also difficult to implement in practice due to future model maintenance concerns. However, the solvents typically used in the preparation of single fragrance raw materials are relatively simple. Therefore, the above embodiments propose a method for constructing a single fragrance raw material concentration prediction model. This method transforms the traditional model of establishing the relationship between near-infrared spectral data and single fragrance raw material concentration into a model of establishing the relationship between near-infrared spectral data and solvent concentration. By predicting the solvent concentration, the concentration of the corresponding single fragrance raw material can be inferred, greatly reducing the modeling difficulty. Moreover, considering that modeling requires the preparation of a large number of samples and the processing of a large amount of near-infrared spectral data, this invention combines model transfer technology. It only requires the construction of several solvent concentration prediction models for a single type of single fragrance raw material in various solvents. The model transfer technology can then be used to construct solvent concentration prediction models for various types of single fragrance raw materials in various solvents, further greatly reducing the modeling difficulty and improving modeling efficiency and operability.

[0048] Based on the single fragrance ingredient concentration prediction model construction method provided in the above embodiments, this invention also provides a single fragrance ingredient concentration prediction method, such as... Figure 2 As shown, it includes the following steps:

[0049] Step 1: Determine the type of monomeric fragrance raw material and solvent in the unknown concentration of the monomeric fragrance raw material solution;

[0050] Step 2: Select the solvent concentration prediction model corresponding to the type of single fragrance raw material and solvent type to be tested, constructed using the single fragrance raw material concentration prediction model construction method described in the above embodiments.

[0051] Step 3: Obtain the near-infrared spectrum of the unknown concentration of the monomer fragrance raw material solution and preprocess it. Then, use the selected solvent concentration prediction model to predict the solvent concentration C. 溶剂 ;

[0052] Step 4: Obtain the concentration C of the single fragrance raw material to be tested based on the predicted solvent concentration. 单体香原料 =1-C 溶剂 .

[0053] In this embodiment, cinnamaldehyde, phenylethyl alcohol, sweet orange oil, and oakmoss extract were used as experimental subjects. Anhydrous ethanol was used for dilution, and a 40% concentration solution of the above-mentioned monomeric fragrance raw materials in anhydrous ethanol was artificially prepared as a sample of the unknown concentration of the monomeric fragrance raw material to be tested for model testing. The anhydrous ethanol concentration in the unknown concentration of the monomeric fragrance raw material was predicted by calling the new four solvent (anhydrous ethanol) concentration prediction models obtained after model transfer in step S4 of the aforementioned embodiment. The anhydrous ethanol concentrations of the corresponding unknown concentration monomeric fragrance raw materials are shown in Table 1.

[0054] Table 1

[0055]

[0056] Based on the anhydrous alcohol concentration in the unknown concentration monomeric fragrance raw materials obtained from Table 1, according to formula C... 单体香原料 =1-C 溶剂 The actual concentrations of the corresponding unknown concentrations of the tested monomeric fragrance raw materials are shown in Table 2.

[0057] Table 2

[0058]

[0059] As can be seen from the above model test results, overall, the single fragrance raw material concentration prediction model established in this invention has a good prediction effect.

[0060] Based on the single fragrance raw material concentration prediction model construction method provided in the foregoing embodiments, solvent concentration prediction models of various single fragrance raw materials in various solvents can be quickly constructed, which greatly reduces the modeling difficulty and improves the modeling efficiency. Furthermore, the above-mentioned rapid modeling technology can be used to quickly determine the concentration of various unknown single fragrance raw materials, which can be used to assist in the quality control process of fragrance raw materials.

[0061] This invention also provides an electronic terminal, comprising:

[0062] A memory that stores computer programs;

[0063] A processor is used to load and execute the computer program to implement the steps of the single fragrance raw material concentration prediction model construction method or single fragrance raw material concentration prediction method described in the above embodiments.

[0064] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the single fragrance raw material concentration prediction model construction method or the single fragrance raw material concentration prediction method described in the above embodiments.

[0065] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0066] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0067] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0068] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0069] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0070] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0071] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for constructing a model to predict the concentration of single fragrance raw materials, characterized in that, include: S1: Obtain near-infrared spectral data of a pure solution of a certain monomeric fragrance raw material, a pure solution of a certain solvent, and a series of solutions of the same monomeric fragrance raw material at different concentrations in the solvent; S2: All the acquired near-infrared spectra are preprocessed, and then the partial least squares method is used to correlate and fit each preprocessed near-infrared spectrum with its corresponding actual solvent concentration to construct a solvent concentration prediction model for the monomeric fragrance raw material and solvent. S3: Obtain the near-infrared spectrum of pure solutions of other types of monomeric fragrance raw materials and perform preprocessing; S4: Based on the near-infrared spectral data of the pre-processed pure solutions of other types of single fragrance raw materials and the pure solvent solution in step S1, the solvent concentration prediction model obtained in step S2 is transferred using a model transfer algorithm to obtain the corresponding solvent concentration prediction model for other types of single fragrance raw materials and the solvent, which is used to predict the concentration of the solvent in the solutions of other types of single fragrance raw materials, and thus obtain the concentration of other types of single fragrance raw materials; wherein the model transfer algorithm is a slope / intercept model transfer algorithm.

2. The method for constructing a concentration prediction model for single fragrance raw materials according to claim 1, characterized in that, The solvent includes one of water, propylene glycol, glycerol, alcohol, triacetin, and triethyl citrate.

3. The method for constructing a concentration prediction model for single fragrance raw materials according to claim 1, characterized in that, The preprocessing methods include one or more combinations of sample normalization, sample centering, sample scaling, standard normal variable transformation, Gaussian filtering smoothing, Savitzky-Golay differentiation, multivariate scattering correction, and removal of uninformative variables.

4. The method for constructing a concentration prediction model for monomeric fragrance raw materials according to claim 3, characterized in that, The preprocessing method was optimized by using the cross-validation correlation coefficient and root mean square error of the solvent concentration prediction model as evaluation indicators.

5. The method for constructing a concentration prediction model for monomeric fragrance raw materials according to any one of claims 1 to 4, characterized in that, Also includes: Obtain the near-infrared spectrum of at least one solution sample of other types of monomeric fragrance raw materials at a known concentration in this solvent and preprocess it; The solvent concentration prediction model obtained by the model transfer algorithm was validated based on the near-infrared spectra of pretreated solution samples with known concentrations and the corresponding actual solvent concentrations.

6. A method for predicting the concentration of monomeric fragrance raw materials, characterized in that, include: To determine the type of analyte fragrance raw material and solvent in the analyte fragrance raw material solution of unknown concentration; Select a solvent concentration prediction model corresponding to the type of single fragrance raw material and solvent type to be tested, constructed using the single fragrance raw material concentration prediction model construction method as described in any one of claims 1 to 5. Near-infrared spectra of the analyte monomer fragrance raw material solution with unknown concentrations were obtained and preprocessed. The solvent concentration C was then predicted using a selected solvent concentration prediction model. 溶剂 ; The concentration C of the single fragrance raw material to be tested is obtained based on the predicted solvent concentration. 单体香原料 =1-C 溶剂 .

7. An electronic terminal, characterized in that, include: A memory that stores computer programs; A processor for loading and executing the computer program to implement the steps of the method as described in any one of claims 1 to 6.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 6.

Images