阅读说明：本技术 一种快速确定甜味剂配方及使用量的方法及其应用 (Method for rapidly determining formula and using amount of sweetener and application thereof ) 是由 李亚军 卢传礼 刘娟 于 2020-06-04 设计创作，主要内容包括：本发明公开了一种快速确定甜味剂配方及使用量的方法,该方法通过口感曲线结合化学计量学方法,建立复合甜味剂配比及总用量的定量预测模型,可根据产品口感要求快速逆向确定甜味剂配方及用量。获取操作简单、快速,节省了大量人力物力。为食品开发过程中的食品添加剂的应用及产品质量控制提供了有力的支持,在实际工业生产中具有重要应用价值。(The invention discloses a method for rapidly determining a formula and a using amount of a sweetening agent. The acquisition operation is simple and quick, and a large amount of manpower and material resources are saved. Provides powerful support for the application of food additives and the quality control of products in the process of food development, and has important application value in the actual industrial production.)

1. A method for rapidly determining the formula and amount of a sweetener, comprising the steps of:

s1 compounding sweeteners with different types, proportions and total dosage, adding the sweeteners into a system to be applied, carrying out sensory evaluation, and measuring a taste curve;

s2, thinning the taste curve data interval collected in the step S1, interpolating, and increasing the curve resolution;

s3, dividing the data collected in the step S2 into a training set and a verification set by using an SPXY method;

s4, preprocessing the taste curve data of the training set and optimizing the modeling interval;

s5, correlating actual values of the sweetener proportion and the use amount in the step S1 with the training set taste curve data obtained in the step S4 by using a chemometric method, and establishing a quantitative prediction model of the sweetener proportion and the use amount;

and S6, verifying the prediction model by using the verification set sample data, and determining the optimal quantitative prediction model of the compound sweetener proportion and the dosage according to the model parameters.

2. The method of claim 1, wherein the sweetener material of step S1 is selected from the group consisting of rebaudioside a, stevioside, monopotassium glycyrrhizinate, tripotassium glycyrrhizinate, ammonium glycyrrhizinate, mogroside, sucralose, sodium saccharin, cyclamate, calcium cyclamate, acesulfame potassium, alitame, aspartame, xylitol, erythritol, sorbitol, maltitol, lactitol, isomaltulose, and sucrose.

3. The method of claim 1, wherein the taste profile determined in step S1 is a time-intensity profile of taste.

4. The method of claim 1, wherein the system to be applied in step S1 comprises milk and dairy products, beverages, preserves, preserved fruits, baked goods, and condiments.

5. The method of claim 1, wherein the interpolation method of step S2 includes lagrangian interpolation, piecewise linear interpolation, spline interpolation, Hermite interpolation and fractal interpolation.

6. The method of claim 1, wherein the step S3 uses sample set partitioning based on joint x-y distance training set samples to group the taste profile data collected in step S2.

7. The method of claim 1, wherein the preprocessing method involved in step S4 is at least one of first order differentiation, second order differentiation, detrending, baseline, orthonormal transformation, and s.g. smoothing.

8. The method of claim 1, wherein the preferred method for modeling intervals in step S4 comprises at least one of genetic algorithm, spaced partial least squares, competitive adaptive reweighing algorithm, and monte carlo non-informative variable elimination.

9. The method of claim 1, wherein the step of establishing quantitative predictive models in step S5 includes linear models and non-linear models; the method for establishing the quantitative prediction model in the step S5 is a partial least square method, a support vector machine or an artificial neural network; the model parameters in step S6 include correlation coefficient, correction error root mean square, cross validation error root mean square, and validation error root mean square.

10. An application of the model obtained by the method for rapidly determining the formula and the usage amount of the sweetener according to claim 1 is characterized in that a target taste curve is set according to an actual application scene and is led into a prediction model, so that the optimal proportion and the usage amount of the compound sweetener are obtained.

Technical Field

The invention belongs to the technical field of food additives, and particularly relates to a method for rapidly determining a formula and a using amount of a sweetener, in particular to a method for quantitatively predicting the composition and the using amount of the sweetener by combining a taste time-intensity curve with chemometrics.

Background

Sweeteners are a very important class of food additives and can be divided into nutritional and non-nutritional types. The nutritive sweetener has a sweetness equivalent to that of sucrose, has a caloric value of 2% higher than that of sucrose, and includes various sugars (such as glucose, fructose, maltose, etc.) and sugar alcohols (sorbitol, xylitol, etc.). The non-nutritive sweetener has a sweetness equal to that of sucrose, has a calorific value lower than that of sucrose by 2%, and includes natural substances such as stevioside and glycyrrhizin, and chemically synthesized substances such as aspartame, saccharin, cyclamate, acesulfame potassium and sucralose. These sweeteners have been widely used in the food, beverage and like industries.

Desirable sweetener requirements are: safe and nontoxic, has pure sweet taste similar to sucrose, high sweetness, low calorific value or no calorific value, high stability, no caries and reasonable price. Each sweetener has a sweet taste and texture different from sucrose, and often produces an unpleasant flavor and aftertaste when used in large amounts, whereas the compound sweetener overcomes the above-mentioned disadvantages. The compound sweetener refers to a sweetener which is prepared by compounding two or more natural or artificial sweeteners. The sweet taste sweetener utilizes the synergistic interaction and taste physiological characteristics of various sweeteners to achieve the comprehensive sweet taste effect, and has the following remarkable advantages: (1) bad taste is reduced, and flavor is increased; (2) shortening the gustatory deficit at the onset of taste; (3) the stability of the sweet taste is improved; (4) the total usage amount of the sweetening agent is reduced, the cost is reduced, and the like. Compound sweeteners have become an important development direction for sweetener development and application.

As sweeteners gradually warm up, objective evaluation of sweetness and flavor of the sweetener becomes important. The evaluation index of the sweetener can be divided into four aspects: evaluation of sweetness values, nuance measurements, test of the sensitivity of the assessor to sweetness and descriptive analysis. The Time-Intensity (T-I) curve is used as the evaluation method of the integral flavor of the sweetener, the limitation that the equal sweetness method only focuses on a single sweetness value is overcome, the continuous change rule of the sensory index of the sweetener in the mouth along with the Time can be obtained, more taste evaluation indexes are obtained, and the flavor evaluation of the product is more comprehensive.

According to the requirements of the mouthfeel of products such as food, beverage and the like, the formula and the dosage of the sweetening agent are determined to be crucial to the development and popularization of the products. At present, in the process of food development, sweetener selection and use amount are determined, and the sweetener is obtained by preparing sweetener samples with different formulas and proportions one by one for sensory evaluation. The method has the disadvantages of complicated operation, long time consumption, large amount of manpower and material resources, and is not beneficial to product development. There is no scientific method for rapidly and reversely determining the formula and the dosage of the sweetening agent according to the taste requirement of the product in the field.

Disclosure of Invention

In view of the defects of the prior art, the invention aims to provide a method for inversely determining the formula and the dosage of a sweetener according to the taste requirement of a product and application thereof.

In order to achieve the purpose, the invention adopts the following technical scheme:

a method for rapidly determining the formula and amount of sweetener to be used, said method comprising the steps of:

s1 compounding sweeteners with different types, proportions and total dosage, adding the sweeteners into a system to be applied, carrying out sensory evaluation, and measuring a taste curve;

s2, thinning the taste curve data interval collected in the step S1, interpolating, and increasing the curve resolution;

s3, dividing the data collected in the step S2 into a training set and a verification set by using an SPXY method;

s4, preprocessing the taste curve data of the training set and optimizing the modeling interval;

s5, correlating actual values of the sweetener proportion and the use amount in the step S1 with the training set taste curve data obtained in the step S4 by using a chemometric method, and establishing a quantitative prediction model of the sweetener proportion and the use amount;

and S6, verifying the prediction model by using the verification set sample data, and determining the optimal quantitative prediction model of the compound sweetener proportion and the dosage according to the model parameters.

Preferably, the sweetener raw material of step S1 is selected from rebaudioside a, stevioside, monopotassium glycyrrhizinate, tripotassium glycyrrhizinate, ammonium glycyrrhizinate, mogroside, sucralose, sodium saccharin, cyclamate, calcium cyclamate, acesulfame potassium, alitame, aspartame, xylitol, erythritol, sorbitol, maltitol, lactitol, isomaltulose, and sucrose.

Preferably, the mouthfeel curve measured in the step S1 is a time-intensity curve of mouthfeel.

Preferably, the system to be applied in step S1 includes milk and dairy products, beverages, preserves, preserved fruits, baked foods, and seasonings.

Preferably, the interpolation method in step S2 includes lagrange interpolation, piecewise linear interpolation, spline interpolation, Hermite interpolation, and fractal interpolation.

Preferably, in the step S3, the taste curve data collected in the step S2 is grouped by using a sample set partitioning on joint x-distance training set sample selection method.

Preferably, the preprocessing method involved in step S4 is at least one of first order differentiation, second order differentiation, detrending, baseline, orthonormal transformation, and s.g. smoothing.

Preferably, the preferred method for modeling the interval in step S4 includes at least one of a genetic algorithm, an interval partial least squares method, a competitive adaptive re-weighting algorithm, and a monte carlo information-free variable elimination method.

Preferably, the establishing of the quantitative prediction model in step S5 includes a linear model and a nonlinear model.

Preferably, the method for establishing the quantitative prediction model in step S5 is a partial least squares method, a support vector machine or an artificial neural network.

Preferably, the model parameters in step S6 include correlation coefficient, correction error root mean square, cross validation error root mean square, and validation error root mean square.

The invention also provides application of the model obtained by the method for rapidly determining the formula and the usage amount of the sweetener, and a target taste curve is set according to an actual application scene and is led into the prediction model, so that the optimal proportion and the usage amount of the compound sweetener are obtained.

The invention has the beneficial effects that:

1. the invention provides a method for rapidly determining a sweetener formula and the amount of the sweetener according to taste requirements.

2. Compared with the existing determination method for selecting and using the compound sweetener in the food development process, namely a method for performing sensory evaluation on sweetener samples with different formulas and different proportions one by one, the method is established on the basis of a large number of data models in the previous period, a quantitative prediction model is established, the optimal formula can be directly obtained through target reverse simulation, the acquisition operation is simple and rapid, and a large amount of manpower and material resources are saved. Provides powerful support for the application of food additives and the quality control of products in the process of food development, and has important application value in the actual industrial production.

Drawings

FIG. 1 is a flow chart of the quantitative prediction model building and prediction of the present invention;

figure 2 is the mouthfeel time-intensity curve (T-I) for the sweetener of example 1;

FIG. 3 is a schematic illustration of interpolation according to example 1;

FIG. 4 is a graph of the optimization modeling interval of the quantitative prediction model of aspartame mass fraction in example 1;

FIG. 5 is a graph relating the predicted mass fraction to actual mass fraction of aspartame in comparative example 1;

FIG. 6 is a graph of the optimized modeling interval of the quantitative prediction model of the compound sweetener dosage in comparative example 2;

FIG. 7 is a graph relating predicted value to actual value of the amount of the compound sweetener in comparative example 2;

FIG. 8 is a plot of the predicted value versus the actual value of the mass fraction of rebaudioside A for comparative example 3;

FIG. 9 is a graph showing the correlation between the predicted value and the actual value of the mogroside mass fraction in comparative example 3;

figure 10 is a graph relating predicted and actual compound sweetener usage values for comparative example 3.

Detailed Description

The present invention will be further described with reference to the accompanying drawings, and it should be noted that the following examples are provided to illustrate the detailed embodiments and specific operations based on the technical solutions of the present invention, but the scope of the present invention is not limited to the examples.

As shown in fig. 1, the present invention is a method for rapidly determining the formula and amount of sweetener to be used, comprising the steps of:

s1 compounding sweeteners with different types, proportions and total dosage, adding the sweeteners into a system to be applied, carrying out sensory evaluation, and measuring a taste curve;

s2, thinning the taste curve data interval collected in the step S1, interpolating, and increasing the curve resolution;

s3, dividing the data collected in the step S2 into a training set and a verification set by using an SPXY method;

s4, preprocessing the taste curve data of the training set and optimizing the modeling interval;

s5, correlating actual values of the sweetener proportion and the use amount in the step S1 with the training set taste curve data obtained in the step S4 by using a chemometric method, and establishing a quantitative prediction model of the sweetener proportion and the use amount;

and S6, verifying the prediction model by using the verification set sample data, and determining the optimal quantitative prediction model of the compound sweetener proportion and the dosage according to the model parameters.

Preferably, the sweetener raw material of step S1 is selected from rebaudioside a, stevioside, monopotassium glycyrrhizinate, tripotassium glycyrrhizinate, ammonium glycyrrhizinate, mogroside, sucralose, sodium saccharin, cyclamate, calcium cyclamate, acesulfame potassium, alitame, aspartame, xylitol, erythritol, sorbitol, maltitol, lactitol, isomaltulose, and sucrose.

Preferably, the mouthfeel curve measured in the step S1 is a time-intensity curve of mouthfeel.

Preferably, the system to be applied in step S1 includes milk and dairy products, beverages, preserves, preserved fruits, baked foods, and seasonings.

Preferably, the interpolation method in step S2 includes lagrange interpolation, piecewise linear interpolation, spline interpolation, Hermite interpolation, and fractal interpolation.

Preferably, in the step S3, the taste curve data collected in the step S2 is grouped by using a sample set partitioning on joint X-distance training set sample selection method.

Preferably, the preprocessing method involved in step S4 is at least one of first order differentiation, second order differentiation, detrending, baseline, orthonormal transformation, and s.g. smoothing.

Preferably, the preferred method for modeling the interval in step S4 includes at least one of a genetic algorithm, an interval partial least squares method, a competitive adaptive re-weighting algorithm, and a monte carlo information-free variable elimination method.

Preferably, the establishing of the quantitative prediction model in step S5 includes a linear model and a nonlinear model.

Preferably, the method for establishing the quantitative prediction model in step S5 is a partial least squares method, a support vector machine or an artificial neural network.

Preferably, the model parameters in step S6 include correlation coefficient, correction error root mean square, cross validation error root mean square, and validation error root mean square.

The invention also provides application of the model obtained by the method for rapidly determining the formula and the usage amount of the sweetener, and a target taste curve is set according to an actual application scene and is led into the prediction model, so that the optimal proportion and the usage amount of the compound sweetener are obtained.