JP2017032326A - Method for predicting quality of coffee product by utilizing metabolomic analysis, and coffee product - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide: a method for easily predicting the quality of a coffee product which includes as prediction targets, the coffee products of different forms such as coffee beans, coffee liquid and instant coffee and also includes auxiliary raw material derived from material other than coffee beans such as dairy material, saccharide and pH adjuster; and a coffee product in which the predicted quality is reproduced.SOLUTION: There is provided a method for predicting the quality of a coffee product which includes a step of collating a quality prediction model of coffee products whose qualities are known, with a numerical data of a coffee product which is obtained by pretreating the coffee product, obtaining an analysis sample therefrom and converting an analysis result for the analysis sample. The quality prediction model is obtained by: pretreating plural coffee products whose qualities are known and obtaining individual analysis samples; performing instrumental analysis on the individual samples to obtain individual analysis results; converting the analysis results into numerical data; and performing multivariate analysis on a relationship between the converted numerical data and the known qualities. There is also provided a coffee product in which the predicted quality is reproduced.SELECTED DRAWING: None

Description

The present invention relates to a coffee product quality prediction method utilizing metabolomic analysis, and a coffee product manufactured using the quality prediction method.

With the trend of shortening the product life cycle, the lead time for coffee product development tends to be shorter, but the varieties (Arabica and Robusta) used in the target product group, their blend ratio, and roasting degree It is useful for product design from the viewpoint of improving the flavor design accuracy of the developed product and reducing the lead time. However, the product forms of coffee products such as coffee beans, coffee liquor, and instant coffee are not uniform, and there are products that use various secondary ingredients such as dairy ingredients, sugars, and high sweetness sweeteners, and have a complicated composition. For some reason, it is still not easy to predict design information related to the quality of the coffee product.
On the other hand, with the remarkable development of computer technology in recent years, metabolome analysis, which is a research method for comprehensive analysis of low molecular weight metabolites related to life phenomena, has been rapidly and globally utilized,・ There are many applications in the industrial field. In the agriculture, forestry and fisheries / food field, metabolome analysis is an applicable technology for agricultural products such as taste and aroma, and low molecular weight metabolites related to food quality. It is required to strongly promote the application of (Non-patent Document 1).

Examples of reports that apply metabolomic analysis to the food field include: a step of pre-treating green tea to obtain an analysis sample; a step of subjecting the analysis sample to instrumental analysis to obtain an analysis result; and a step of converting the analysis result into numerical data A multivariate analysis step; and a green tea quality prediction method (Patent Document 1) including a step of predicting the quality from the obtained analysis result, or a pretreatment step of pretreating cheese with a known ripening quality to obtain an analysis sample Instrument analysis step of subjecting the analysis sample to instrument analysis to obtain instrument analysis data; multivariate analysis using instrument analysis data for a plurality of the cheeses and ripening quality data representing each ripening quality of the cheese; A multivariate analysis step of creating a cheese quality prediction model that represents the relationship between the device analysis data and the ripening quality predicted from the device analysis data The a, a quality prediction method for cheese (Patent Document 2) are known.
However, these reports are quality evaluation methods related to green tea or cheese, and no mention is made of methods for predicting coffee bean varieties, blend ratios, and roasting degrees.
As a method for identifying the variety of coffee beans and predicting the blend ratio, for example, a method using DIP / IA-MS (Direct Inlet Probe / Ion Attachment Ionization Mass Spectrometry) is known (Non-patent Document 2). . However, in this method, since the roasting degree is limited to a specific value, it is difficult to predict the quality of products on the market. Further, there is no mention of the validity in the presence of auxiliary ingredients such as milk components and the predictability of design information related to the degree of roasting.
In addition, a method using Raman spectroscopy is also known as a method for predicting coffee bean varieties and blend ratios (Non-Patent Document 3). However, in this method, the pretreatment of the instrumental analysis is complicated and not easy, and further, the validity in the presence of secondary ingredients such as milk components and the predictability of design information related to the roasting degree are mentioned. It has not been.
Further, as a method for simultaneously predicting the roasting degree and the variety, a method using near infrared spectroscopy is known (Non-Patent Document 4). However, this method also mentions a method for predicting design information of coffee ingredients used in coffee liquor, not coffee beans, and a method for predicting the quality of coffee products in the presence of auxiliary ingredients such as milk components. Absent.

Promotion Policy for Industry-Academia-Government Collaboration Research in the Agriculture, Forestry and Fisheries / Food Field Utilizing Metabolome Analysis January 14, 2015 Industry-Academia-Government Collaboration Research Study Group in Agriculture, Forestry and Fisheries / Food Field Utilizing Metabolome Analysis Analytical chemistry vol. 63, 2014, no. 10, pp825-830 Thomas Wermelinger, et al. , Quantification of the Robusta Fraction in a Coffee Blend via Raman Spectroscopy: Proof of Principal, Journal of Agricultural 59 E. Bertone, et al. , Simulaneous determination by NIR spectroscopy of the roasting degree and Arabic / Robusta ratio in roasted and ground coffe, Food Control 68, 68 (9)

JP 2009-14700 A JP 2013-7732 A

The present invention includes coffee products having different forms such as coffee beans, coffee liquor, and instant coffee as prediction targets, and also includes auxiliary materials derived from sources other than coffee beans such as milk materials, sugars, and pH adjusters. An object is to provide a method for easily predicting quality.

The present invention is a method for predicting the quality of a coffee product comprising:
[1] A step of pretreating a coffee product to obtain an analytical sample;
Subjecting the analysis sample to instrumental analysis to obtain an analysis result;
Converting the analysis result into numerical data; and the obtained numerical data,
Pre-treating a plurality of coffee products of known quality to obtain individual analytical samples;
Subjecting the individual analytical samples to instrumental analysis to obtain individual analytical results;
Converting the individual analysis results into numerical data; and multivariate analysis of the relationship between the individual converted numerical data and the quality;
Checking with a coffee product quality prediction model obtained by
Quality prediction method for coffee products including
[2] The method according to [1], wherein the coffee product is coffee beans, coffee liquor, or instant coffee.
[3] The method according to [1], wherein the coffee product is a coffee product which contains, in addition to [2], auxiliary materials such as milk materials, sugars and pH adjusters.
[4] The pretreatment includes a step of extracting and filtering coffee beans with an aqueous solvent, a step of filtering coffee liquid, or a step of dissolving and filtering instant coffee with an aqueous solvent. [1] to [3] The method described.
[5] The method according to [1] to [4], wherein the instrumental analysis is provided with a diode array detector as a detector after the high performance liquid chromatography. [6] The multivariate analysis is pattern recognition. The method according to [1] to [5], which is selected from a method or a multivariate regression analysis method.
[7] The method according to [1] to [6], wherein the quality is selected from coffee bean varieties, blend ratio of coffee bean varieties, and roasted degree of coffee beans.
[8] The method according to [1] to [7], wherein the quality is a blend ratio of coffee bean varieties and / or coffee bean varieties and / or a roast degree of coffee beans.
[9] A coffee product that reproduces the quality predicted by the method according to [1] to [8].

ADVANTAGE OF THE INVENTION According to this invention, the quality prediction of the coffee product which was difficult conventionally can be performed with an accurate and simple method.

(Coffee products subject to quality prediction)
The coffee products that are subject to quality prediction in the present invention include coffee beans, coffee liquor, and instant coffee.

(Definition of coffee beans)
In the present invention, the types of coffee beans that are listed as the quality prediction target coffee products are Arabica, Robusta, and Revelica, and may be a blend of two or more of these. The coffee beans may be green beans, and in the case of roasted beans or roasted and ground beans, the roasting conditions and roasting equipment can be performed by any known equipment and conditions, and those appropriately selected are subject to quality prediction. It becomes. The roasting degree can be expressed by brightness such as a Hunter L value in the Lab color space and an L star value in the CIE color space. For example, in the case of the Hunter L value, the range of 13 to 32 can be exemplified. As a method for measuring the roasting degree, for example, roasted beans are pulverized while removing chaff as appropriate, the pulverized beans are put into a cell, sufficiently tapped, and then measured with a color difference meter. As the color difference meter, ZE-2000, ZE-6000 (manufactured by Nippon Denshoku Industries Co., Ltd.) and the like can be used. Moreover, as a roasting apparatus, a direct fire type, a semi-hot air type, a hot air type etc. can be illustrated, for example. In the case of roasted and ground beans, the roasted beans are produced by pulverizing roasted beans by any known method, and there is no particular limitation.

(Definition of coffee liquid)
In the present invention, the coffee liquid that is subject to quality prediction includes a coffee extract obtained by bringing a roasted ground bean and an aqueous solvent into contact with each other, and a coffee extract having a higher soluble solid content than the normal drinking concentration. Extract, liquid in which instant coffee is dissolved, coffee liquid containing finely pulverized roasted coffee beans that have been finely pulverized to an average particle size of about 300 μm or less, a container that is packed in a container and refrigerated and stored at room temperature Includes stuffed beverages. Furthermore, auxiliary materials such as milk materials and sugars may be added to these coffee liquids as necessary, and different types of coffee liquids may be mixed. The extraction method of the coffee extract and the coffee extract is not particularly limited, and examples thereof include a drip method, an espresso method, a fixed bed column method, a slurry method, a continuous countercurrent method, and an extractive distillation method. Extraction conditions such as extraction temperature, extraction pressure, extraction time, and bean / extraction solvent ratio are not particularly limited, and those manufactured under arbitrary conditions are targeted. As a container-packed beverage, for example, sugar-free black coffee; sweetened black coffee to which sucrose, liquid sugar, sweeteners, etc. are added; milk components such as milk, skim milk powder, and fresh cream are added to sugar-free or sweetened coffee beverages Café olea type coffee drinks etc. are included in the quality prediction targets. When producing a container-packed beverage, any component that can be used in a coffee beverage can be added to the coffee liquid as an auxiliary material. For example, a coffee extract obtained by a conventional method, an antioxidant, pH Examples thereof include regulators, emulsifiers, and fragrances. After normal pH adjustment, homogenize, heat to 80-95 ° C using a plate heat exchanger, etc., then fill and tighten, and sterilize at 120-125 ° C for 10-40 minutes And a retort sterilized coffee drink is obtained. In the case of UHT sterilization, a plate-type heat exchanger or the like is used, and after sterilization treatment at 130 to 145 ° C. for 15 to 60 seconds, it is filled into a container such as a can or a plastic bottle to obtain a container-packed beverage. The container-packed beverage is included as a quality prediction target.

(Definition of instant coffee)
In the present invention, instant coffee (including regular soluble coffee, hybrid coffee, etc.) that is a quality prediction target is a coffee extract that is a coffee extract having a higher soluble solid content concentration than the normal drinking concentration, or an average particle size of 300 μm. Includes coffee extract containing finely pulverized roasted coffee beans that have been finely pulverized to the following extent, in which water has been removed by spray drying or freeze drying, and sugars, milk materials, etc. Also includes instant coffee products with added and mixed ingredients.

(Pretreatment process)
In the present invention, the pretreatment step is a step performed to obtain a form suitable for instrumental analysis in order to obtain information related to the quality of coffee used in the coffee product that is the target of quality prediction. Unit operations such as extraction, fractionation, concentration, drying, and filtration are included. These unit operations may be appropriately combined. In particular, when the instrumental analysis is high-performance liquid chromatography (HPLC), an eluent of high-performance liquid chromatography (HPLC) containing an organic solvent such as methanol or acetonitrile is added to the analytical sample to precipitate insoluble matter, and then chromatographed. A pretreatment step of microfiltration using a disk or the like is suitable.
When the instrumental analysis is gas chromatography (GC), a solvent extraction method using a known organic solvent such as diethyl ether, isopentane or pentane, or a volatile component of coffee that has reached equilibrium in a sealed container Pretreatment step used for headspace gas analysis collected or adsorbed on adsorbent, pretreatment step used for solid phase extraction method after elution with solvent after adsorbing to non-polar porous polymer beads, steam distillation with rotary evaporator etc. Well-known pre-processing processes, such as the pre-processing process used for the vacuum steam distillation method which isolate | separates a volatile component by performing, are mentioned.

(Instrument analysis process)
In the present invention, the instrument analysis step is a step of performing analysis / measurement using an analysis instrument, and the analysis method includes high performance liquid chromatography (HPLC), gas chromatography (GC), ion chromatography. , Mass spectrometry (MS), near infrared spectroscopy (NIR), Fourier transform infrared spectroscopy (FT-IR), nuclear magnetic resonance analysis (NMR), Fourier transform nuclear magnetic resonance analysis (FT-NMR), inductive coupling A plasma mass spectrometer (ICP-MS) or the like can be exemplified.
A detector for high performance liquid chromatography (HPLC) is not particularly limited, and a known one can be used as appropriate. For example, a diode array (DAD) detector can be suitably used.
These instrumental analysis methods may be used in combination of two or more.
The measurement conditions can be arbitrarily selected as long as sufficient reproducibility according to the target prediction accuracy is obtained.

(Process to convert to numerical data)
The data obtained in the instrument analysis process is obtained as waveform data obtained by sampling the result of the change in signal intensity converted into a normal electrical signal at regular time intervals, or as an integrated value obtained by integrating electrical signals for a certain time. Feature values such as slope, periodicity, and amplitude obtained from waveform data, and smoothing processing of waveform data to detect peaks and extracting numerical values such as area value, peak height, peak width, etc., or a specific time The step of converting into numerical data such as using the signal intensity in the range as it is or taking out a numerical value quantified by paying attention to a specific compound can be arbitrarily selected from known methods.
In particular, when performing an instrumental analysis in which a diode array detector is provided as a detector after the high performance liquid chromatography (HPLC), a peak is detected from the obtained chromatogram data, and the area value is obtained as numerical data. Is preferable from the viewpoint of high reproducibility of numerical values. Therefore, although it is possible to create a quality prediction model without specifying the target peak component, use the numerical data of the compound that is common to the training and prediction target samples and is relatively easy to quantify. Is not particularly limited as long as it is a compound contained in a known coffee, but chemistry and technology of coffee roasting, published by Kogaku Publishing Co., Ltd. and well-known conventional technology collection (fragrance) part II, January 2000 On the 14th, compounds listed in the JPO issue are listed.

(Multivariate analysis process)
In the present invention, the multivariate analysis step is performed in order to learn and predict quality using the obtained numerical data. When the quality is categorical classification such as varieties and ratings, k-neighbor method, decision tree, SIMCA (Soft Independent Modeling of Class Analysis), linear discriminant analysis, logistic discriminant analysis, stepwise variable selection discriminant analysis, MT (Mahalanobis Taguchi) method, Fisher discriminant analysis, random forest discriminant analysis, support vector machine, bagging, learning vector quantization, robust discriminant analysis using t-distribution, neural network discriminant analysis, PLS (Partial Last Squares) A pattern recognition method including a discriminant analysis method such as discriminant analysis and a machine learning method can be exemplified.
In the case of the roasting degree (L value) and the blend ratio of the varieties where the quality is expressed as a continuous value, principal component regression analysis, PLS (Partial Last Squares) regression analysis, generalized linear regression analysis, bagging, support vector machine A multivariate regression analysis method including machine learning / regression analysis methods such as random forest and neural network regression analysis can be exemplified. Depending on the quality to be predicted and the nature of the numerical data used for the description, these methods can be used in appropriate combination in order to improve the prediction accuracy.
An analysis tool for executing these multivariate analyzes can be appropriately selected from a large number of commercially available software and freeware analysis tools.
In addition, various analysis parameters such as the number of repetitions of calculation can be appropriately adjusted in order to improve the calculation time and prediction accuracy.

(Process to compare with the quality prediction model)
In the present invention, the quality of the coffee product refers to the type of coffee beans used, the blend ratio of the coffee bean varieties, the origin of the coffee beans, the blend ratio of coffee beans of different origins, the roasting degree of roasted beans, and the raw beans It includes quality related to coffee flavor design and quality related to human sensory quantity, such as conversion value, soluble solid content extraction yield, flavor rating, and sensory evaluation score. In particular, from the viewpoint of high prediction accuracy, coffee beans varieties, blend ratios of coffee bean varieties, and roasting degree of roasted beans can be preferably exemplified.
In the present invention, a plurality of coffee products of known quality are pre-processed to obtain individual analysis samples, the individual analysis samples are subjected to instrument analysis to obtain individual instrument analysis data, and the individual instrument analysis data is further obtained. And a multivariate analysis using each of the known qualities of a plurality of coffee products to create a coffee product quality prediction model that represents the relationship between the device analysis data and the quality predicted from the device analysis data To do.
The prediction accuracy of the quality prediction model is determined by dividing the device analysis data of known quality for learning and testing, creating a quality model using only the learning data, and using the test data for the quality prediction. Evaluation can be made by comparing the result of predicting the quality of the test data against the model and the known quality of the test data. By evaluating the high prediction accuracy of the quality prediction model, it is possible to select an appropriate multivariate analysis method and instrument analysis data.

The higher the accuracy of the quality prediction model, the better. However, it varies depending on the type of target quality, the cost of training data collection for creating the quality prediction model, and time constraints. The range of the prediction accuracy can be arbitrarily set according to the purpose.

  In the same manner, the coffee product of unknown quality is pre-processed to obtain an analysis sample, the analysis sample is subjected to instrument analysis, instrument analysis data is obtained, and the instrument analysis data is obtained from the quality of the coffee product prepared above. By comparing with the prediction model, the quality of the coffee product whose quality is unknown can be predicted.

  Based on the predicted coffee product quality, obtain Arabica, Rob and Revelica coffee beans from the predicted origin, roast to the expected degree of roasting, and the predicted blend ratio By further blending so as to be, and further appropriately extracting and processing according to the form of the coffee product, a coffee product that reproduces the predicted quality can be obtained.

Hereinafter, the present invention will be described more specifically with reference to examples. However, the present invention is not limited to these examples.

Example 1
(Prediction of variety)

<Preparation of coffee liquor and pretreatment process>
Arabica roasted beans 191 samples (Brazil 158 samples, Colombia 33 samples, L value 15-24, obtained from Takasago Coffee Co., Ltd.), Robusta (Vietnam) roasted beans 58 samples (L value 16-23, Takasago coffee) (Obtained from Co., Ltd.) using a commercially available dipping mill (manufactured by Ditting Machinen AG), and then brought into contact with 10 to 20 parts by weight of hot water for 1 to 10 minutes per 1 part by weight of the ground beans Extraction and filtration with filter paper were performed to obtain a coffee extract.
Of the obtained coffee extract, 174 samples were directly subjected to pretreatment for HPLC analysis.
Of the obtained coffee extract, 75 samples were adjusted to pH 5.9 using sodium bicarbonate water, and then the ion exchanged water was used to obtain a soluble solid content concentration of 1.2% by weight. Diluted, filled into a can and wound, and then retort sterilized under the condition of F0 = 10 to obtain a black type container-packed beverage. Of the obtained container-packed beverages, 36 samples were stored as they were at room temperature until they were subjected to pretreatment for HPLC analysis (black type container-packed beverages). The remaining 39 specimens were stored at 60 ° C. for 1 week, and then stored at room temperature until they were subjected to HPLC analysis pretreatment (warmed black type container-packed beverage).
1.5 g of the obtained coffee liquid (coffee extract 174 samples, black type container-packed beverage 36 samples, warmed black type container-packed beverage 39 sample) was weighed, and HPLC eluent (liquid A: 0.05M acetic acid, 5% by volume acetonitrile aqueous solution) was added and the volume was adjusted.

<Instrument analysis process>
10 μL of the microfiltrated coffee liquid was subjected to instrumental analysis (1100 series, manufactured by Agilent Technologies) provided with a diode array detector as a detector after the HPLC.

[HPLC conditions]
Column: Capsule pack MGIII (4.6 × 150 mm, 3 μm, manufactured by Shiseido Co., Ltd.)
Solvent: A: 0.05 M acetic acid, 5% by volume acetonitrile aqueous solution, B: acetonitrile flow rate: 1 mL / min

<Process to convert to numerical data>
In the chromatogram obtained as a result of the HPLC analysis, a total of 29 peaks presumed to exist in common with the specimen donated to the analysis were selected from the retention time and the UV-Vis spectrum, and each area value was calculated ( Agilent OpenLAB chromatographic data system version C.01.07 (manufactured by Agilent Technologies) was used). In addition, the sum of these 29 peak area values was calculated. Next, a peak area ratio obtained by dividing each of the 29 peak area values by the sum of the 29 peak area values was calculated and subjected to a multivariate analysis step.

<Multivariate analysis process>
A total of 166 samples, 128 samples and 38 samples, were selected at random from Arabica type 191 samples and Robusta type 58 samples, respectively, and used as a sample group for training. The remaining 83 specimens were used as test specimen groups.
SIMCA (Soft Independent Modeling of Class Analog), linear, with 29 HPLC analysis peak area ratios of each sample belonging to the training sample group as explanatory variables, coffee varieties (that is, Arabica or Robusta) as response variables Discriminant analysis, stepwise variable selection discriminant analysis, Fisher discriminant analysis, random forest discriminant analysis, learning vector quantization, robust discriminant analysis using t-distribution, k-neighbor method, neural network discriminant analysis The analysis method was selected and nine types of quality prediction models were created.

<Process to compare with quality prediction model>
Using the 29 HPLC analysis peak area ratios of each sample belonging to the test sample group as explanatory variables, calculation is performed using nine kinds of quality prediction models, and the coffee varieties that are response variables are predicted, The accuracy rate was calculated. As a result, neural network discriminant analysis: 0.91, learning vector quantization: 0.94, linear discriminant analysis: 0.97, and other six types of quality prediction models (SIMCA, stepwise variable selection discriminant analysis, Fisher discriminant) Analysis, random forest discriminant analysis, robust discriminant analysis using t-distribution, k-neighbor method), a high accuracy rate of 1 and high prediction accuracy was obtained.
R (version 3.1.2), which is free statistical analysis software, was used as an analysis tool for the multivariate analysis process described in the examples and the process for matching with the quality prediction model.

(Example 2)
(Prediction of coffee varieties of sweetened and milk-filled container-packed beverages)

<Preparation of coffee liquor and pretreatment process>
Of the coffee extract obtained in Example 1, 30 randomly selected arabica species and 20 lobsta species were analyzed. 6% by weight of granulated sugar, 12% by weight of milk, and emulsifier (obtained from Mitsubishi Chemical Foods Co., Ltd.) Add ion-exchanged water to 0.25% by weight, homogenize, adjust pH to 6.9 using sodium bicarbonate water, and then use ion-exchanged water to add 1 soluble solid content derived from coffee. The sample was diluted to 3% by weight, filled into a can and wound, and then retort sterilized under the condition of F0 = 25 to obtain 50 samples of a sugared / milk-containing container-packed beverage.
The obtained sweetened / milk-filled type container-packed beverage was pretreated in the same manner as in Example 1 and subjected to HPLC analysis.

<Instrument analysis process and conversion to numerical data>
In the same manner as in the method described in Example 1, the peak area ratio was calculated for 50 samples of a sugar-filled / milk-containing container-packed beverage.

<Process to compare with quality prediction model>
Using the quality prediction model of the stepwise variable selection discriminant analysis obtained in Example 1 using the peak area ratio for 50 samples of a sweetened / milk-containing container-packed beverage as an explanatory variable, the prediction of coffee varieties that are response variables As a result, the correct answer rate was 1, and high coffee variety prediction accuracy was obtained by using the same quality prediction model as that of the black type even in the case of a sweetened / milk-filled type container-packed beverage.

(Example 3)
(Prediction of blend ratio of varieties)

<Coffee liquor and pretreatment process>
18 Arabica roasted beans (9 Brazil samples, 9 Colombia samples, L value 15-24, obtained from Takasago Coffee Co., Ltd.), Robusta (Vietnam) 19 roasted beans (L value 16-23, Takasago coffee) (Obtained from Co., Ltd.), 34 specimens of Arabica and Robusta blend roasted beans (Robusta seed roasted beans blend ratios of 15 wt%, 25 wt%, 40 wt%, 50 wt%, 60 wt%, 75 In the same manner as in Example 1, grinding, extraction, and filter paper filtration were performed to obtain 71 coffee extract liquids. The obtained coffee extract was pretreated in the same manner as described in Example 1 and subjected to HPLC analysis.
Separately, 10 samples of coffee extract (Brix 15-55, 4 types of Arabica, 3 types of Robusta, 3 types of blends of Arabica and Robusta were obtained from Takasago Coffee Co., Ltd.) After dilution with, pretreatment was performed in the same manner as described in Example 1 and subjected to HPLC analysis.

<Instrument analysis process and conversion to numerical data>
In the same manner as in the method described in Example 1, the peak area ratio was calculated for 71 samples of coffee extract and 10 samples of coffee extract.

<Multivariate analysis process>
A total of 46 samples, 10 samples, 10 samples, 26 samples, were selected for training from 18 samples of Arabica seed extract, 19 samples of Robusta seed extract, and 34 samples of blend extract of Arabica and Robusta. Further, 4 samples were randomly selected from 4 Arabica coffee extracts and 3 Robusta coffee extracts, for a total of 4 samples. A total of 46 samples of the coffee extract and 4 samples of the coffee extract were used as a sample group for training. The remaining 25 samples of coffee extract and 6 samples of coffee extract, 31 samples in total, were used as test sample groups.
Principal component regression analysis, PLS (Partial Least Squares) regression analysis, support vector, with 29 HPLC analysis peak area ratios of each sample belonging to the training sample group as explanatory variables, and Robusta blend ratio as response variables Six types of quality prediction models were created by selecting six types of multivariate analysis methods: machine, random forest, neural network regression analysis, and generalized linear regression analysis.

<Process to compare with quality prediction model>
Calculation is performed using six quality prediction models using 29 HPLC analysis peak area ratios of each sample belonging to the test sample group as explanatory variables, and prediction of the blend ratio of Robusta species as response variables The prediction accuracy was evaluated using RMSEP (Root Mean Square Error of Prediction) and a coefficient of determination. As a result, principal component regression analysis: RMSEP = 89.7, determination coefficient = 0.95, PLS regression analysis: RMSEP = 37.8, determination coefficient = 0.98, support vector machine: RMSEP = 68.5, determination coefficient = 0.96, random forest: RMSEP = 56.8, coefficient of determination = 0.97, neural network regression analysis: RMSEP = 46.7, coefficient of determination = 0.97, generalized linear regression analysis: RMSEP = 30.7 A high prediction accuracy was obtained with a coefficient of determination = 0.98. Moreover, the tendency for the prediction accuracy of the coffee extract to decrease specifically was not observed, and the blend ratio of the varieties of the coffee extract and the coffee extract could be predicted with substantially the same prediction accuracy.

Example 4
(Prediction of roasting degree)

<Preparation of coffee liquor, pretreatment process, instrument analysis process, and conversion to numerical data>
Data similar to Example 1 was used.

<Multivariate analysis process>
128 samples were selected at random from 191 samples of the arabica species and used as a sample group for arabica species training, and the remaining 63 samples were used as a test sample group. Similarly, 38 samples were selected at random from 58 samples of Robusta, and used as a sample group for Robusta training, and the remaining 20 samples were used as a test sample group.
For each of the Arabica training sample group and Robusta training sample group, the 29 HPLC analysis peak area ratios of each sample belonging to the training sample group are used as explanatory variables, and the coffee used Using the roasting degree (L value) as a response variable, three types of multivariate analysis methods were selected: PLS regression analysis, neural network regression analysis, and random forest, and three types of quality prediction models were created.

<Process to compare with quality prediction model>
Calculation is performed using three quality prediction models for each of Arabica and Robusta, using 29 HPLC analysis peak area ratios of each sample belonging to the test sample group as explanatory variables, which are response variables. The degree of coffee roasting (L value) was predicted, and the prediction accuracy was evaluated using RMSEP and a coefficient of determination. As a result, for Arabica species, PLS regression analysis: RMSEP = 0.37, coefficient of determination = 0.86, neural network regression analysis: RMSEP = 0.25, coefficient of determination = 0.91, random forest: RMSEP = 0. 15. A high prediction accuracy was obtained with a coefficient of determination = 0.95. For Robusta species, PLS regression analysis: RMSEP = 0.34, coefficient of determination = 0.82, neural network regression analysis: RMSEP = 0.24, coefficient of determination = 0.82, random forest: RMSEP = 0.20, determination A prediction accuracy of coefficient = 0.87 was obtained.

According to the present invention, coffee products including coffee beans having different forms, such as coffee beans, coffee liquor, and instant coffee, are included as prediction targets, and also include auxiliary materials derived from other than coffee such as milk materials, sugars, and pH adjusters. The quality that is difficult or impossible to measure directly can be accurately and easily predicted.

Claims (7)

Pre-treating the coffee product to obtain an analytical sample;
Subjecting the analysis sample to instrumental analysis to obtain an analysis result;
Converting the analysis result into numerical data; and the obtained numerical data,
Pre-treating a plurality of coffee products of known quality to obtain individual analytical samples;
Subjecting the individual analytical samples to instrumental analysis to obtain individual analytical results;
Converting the individual analysis results into numerical data; and multivariate analysis of the relationship between the individual converted numerical data and the quality;
Checking with a coffee product quality prediction model obtained by
Quality prediction method for coffee products including The method of claim 1, wherein the coffee product is coffee beans or coffee liquor or instant coffee. The method according to claim 1 or 2, wherein the pretreatment comprises a step of extracting and filtering coffee beans with an aqueous solvent, a step of filtering coffee liquid, or a step of dissolving and filtering instant coffee with an aqueous solvent. 4. The method according to claim 1, wherein the instrumental analysis is performed by providing a diode array detector as a detector after the high performance liquid chromatography. The method according to any one of claims 1 to 4, wherein the multivariate analysis is selected from a pattern recognition method or a multivariate regression analysis method. 6. The method according to claim 1, wherein the quality is selected from coffee bean varieties, blend ratio of coffee bean varieties, and roasted degree of coffee beans. A coffee product that reproduces the quality predicted by the method according to claim 1.