CN102650630A - Method for building distinguishing model for evaluating odor type of flue-cured tobacco raw material on the basis of note components - Google Patents
Method for building distinguishing model for evaluating odor type of flue-cured tobacco raw material on the basis of note components Download PDFInfo
- Publication number
- CN102650630A CN102650630A CN2012101392913A CN201210139291A CN102650630A CN 102650630 A CN102650630 A CN 102650630A CN 2012101392913 A CN2012101392913 A CN 2012101392913A CN 201210139291 A CN201210139291 A CN 201210139291A CN 102650630 A CN102650630 A CN 102650630A
- Authority
- CN
- China
- Prior art keywords
- odor type
- flue
- note
- cured tobacco
- raw material
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention belongs to the technical field of flue-cured tobacco raw material odor type evaluation and particularly relates to a method for evaluating flue-cured tobacco raw material odor type. The invention discloses a method for building a distinguishing model for evaluating flue-cured tobacco raw material odor type on the basis of note components, wherein senses are used for evaluating and quantifying note indexes (tobacco aroma, faint aroma, sweet aroma, fruital aroma, potpourri, picees, costustoot and scorch aroma); with the combination of step-by-step discriminant analysis, the indexes are further screened, and an odor type distinguishing model with less independent variables and good effects is built; and the method is used for evaluating and identifying odor types of flue-cured tobaccos (strong, middle, clean and fresh and sweet types). The method is used for identifying the odor type of the flue-cured tobacco and has the advantages of simplicity in operation, convenience in calculation, high distinguishing rate and the like.
Description
Technical field
The invention belongs to flue-cured tobacco raw material odor type assessment technique field, be specifically related to a kind of method of criticism flue-cured tobacco raw material odor type.
Background technology
The tobacco leaf odor type is the important evidence and the leading indicator of quality of tobacco evaluation and use; The sixties in 20th century, Zhu's honor power, Wang Chenghan etc. were divided into delicate fragrance type, middle odor type, Luzhou-flavor etc. to China's tobacco leaf according to odor type; But how the fragrance style of raw tobacco material is estimated, still do not had the evaluation method of system so far, at present main is main with the sensory evaluating smoking of aroma characteristic still; This evaluation method can only qualitative description; And can not quantitative evaluation, the raw tobacco material sensory evaluating smoking method of domestic employing varies in addition, and it is inconsistent often to cause same raw tobacco material fragrance style to estimate qualification result.Therefore set up the raw tobacco material fragrance style evaluation method of a cover science, system; For accurate location raw tobacco material fragrance style and quality; Show the raw tobacco material style characteristic; Rational Application its in moulding the cigarette product style, the development and application that fundamentally promotes the characteristic sound tobacco effectively has crucial meaning.
Summary of the invention
The objective of the invention is to deficiency to existing flue-cured tobacco raw material evaluation method; Provide a kind of and make up the method that discrimination model is estimated flue-cured tobacco raw material odor type based on the note component; Use the sensory evaluating smoking and carry out note evaluation of indexes and quantification, and combine progressively discriminatory analysis structure discrimination model that the odor type of flue-cured tobacco is estimated discriminating.
The object of the invention is realized through following technical scheme:
A kind of method based on note component structure discrimination model evaluation flue-cured tobacco raw material odor type may further comprise the steps:
S1, flue-cured tobacco note estimate: adopt the mode of secretly commenting by the technician that smokes panel test, the note index of different flavor raw tobacco material is evaluated and quantized;
The foundation of S2, odor type discrimination model: based on the evaluation and the quantized data of step S1 gained note index; Adopt progressively diagnostic method to the note index screening and carry out the Bayes discriminatory analysis; Make up flue-cured tobacco odor type discrimination model, confirm decision rule, and the differentiation effect of model is estimated;
S3, odor type are differentiated: with each discriminant score of calculation sample in the corresponding note quantification of targets of the sample to be identified data difference substitution discrimination model, and then the odor type of definite sample.
For obtaining note index evaluation and quantized result preferably, preferably among the step S1 adopt the mode of secretly commenting by 9 sensory evaluating smoking technician that are good at perception flue-cured tobacco note at least, the note index of different flavor raw tobacco material is evaluated and quantized.
Wherein, the flue-cured tobacco note index among the said step S1 comprises: the fragrant (y of tobacco
1), delicate fragrance (y
2), fragrant and sweet (y
3), fruital (y
4), the fragrance of a flower (y
5), Xin Xiang (y
6), the banksia rose (y
7) and burnt odor (y
8); Said each note index is divided into 6 class: outstanding, outstanding, obviously, obviously, have and do not have, and correspondence was given score value 10 minutes, 8 minutes, 6 minutes, 4 minutes, 2 minutes and 0 minute respectively.
Wherein, the raw tobacco material odor type among the said step S1 comprises Luzhou-flavor, middle odor type, delicate fragrance type and fresh and sweet odor type.
Wherein, the discrimination model among the said step S2 is:
F
Giving off a strong fragrance=-2.703y
2+ 21.561y
3+ 2.319y
7+ 7.094y
8-62.881
F
Middle fragrant=8.494y
2+ 14.239y
3+ 7.34y
7+ 1.767y
8-60.158
F
Delicate fragrance=11.696y
2+ 15.889y
3+ 5.268y
7+ 3.092y
8-84.124
F
Fresh and sweet perfume (or spice)=11.764y
2+ 20.719y
3+ 6.05y
7+ 3.574y
8-113.497.
Wherein, the definite decision rule described in the step S2 is: if F in the model
k(x
0)-max{F
k(x
0)=0, k ∈ { giving off a strong fragrance, middle fragrant, delicate fragrance, fresh and sweet perfume (or spice) } then thinks sample x
0Be under the jurisdiction of F relatively
kThe type of representative; Adopt effect verification methods such as validation-cross and self checking that the differentiation effect of model is estimated among the step S2.
With respect to prior art, the present invention has following advantage:
1, is the basis with sensory evaluating smoking, solved the complex physical perception issue that general physico-chemical analysis cann't be solved;
2, through the sensory evaluating smoking note index of flue-cured tobacco raw material is estimated and quantize; The evaluation that combines progressively discriminatory analysis to make up the odor type evaluation model simultaneously and be applied to the flue-cured tobacco odor type is differentiated; Made up a cover and be applicable to Luzhou-flavor, middle odor type, delicate fragrance type and fresh and sweet odor type flue-cured tobacco are differentiated, the objective raw tobacco material odor type of science decision method, operate simple and easy, calculate quick, differentiation rate height.
Embodiment
Below in conjunction with embodiment the present invention is done to specify further, but they are not to qualification of the present invention.
1. the flue-cured tobacco note is estimated
Sample source in the present embodiment 202 flue-cured tobacco raw materials that group provides in red cloud Red River, wherein sample comprises modeling sample: Luzhou-flavor (38), middle odor type (41), delicate fragrance type (51), fresh and sweet odor type (51); 21 in sample to be identified.
Organize 10 personnel that smoke panel test that are good at perception flue-cured tobacco note, adopt and secretly to comment mode, tobacco perfume of raw tobacco material, delicate fragrance, fragrant and sweet, fruital, the fragrance of a flower, Xin Xiang, the banksia rose and 8 note indexs of burnt odor are evaluated and quantized.Every kind of note is divided into 6 class: outstanding, outstanding, obviously, obviously, have and do not have, and gave score value respectively 10 minutes, 8 minutes, 6 minutes, 4 minutes, 2 minutes and 0 minute.
2. the foundation of odor type discrimination model
Adopt progressively techniques of discriminant analysis, with the fragrant (y of above-mentioned tobacco
1), delicate fragrance (y
2), fragrant and sweet (y
3), fruital (y
4), the fragrance of a flower (y
5), Xin Xiang (y
6), the banksia rose (y
7) and burnt odor (y
8) totally 8 note achievement datas be that independent variable carries out discriminatory analysis to four kinds of odor types (Luzhou-flavor, middle odor type, delicate fragrance type and fresh and sweet odor type) flue-cured tobacco raw material, make up the discriminant function of flue-cured tobacco odor type.
Adopt Bayes discriminant function method,, from 8 note indexs, filter out the delicate fragrance that can characterize the tobacco leaf odor type, fragrant and sweet, the banksia rose and 4 variablees of burnt odor and get into discriminant function (being characteristic index) (seeing table 1) through the multistep discriminatory analysis.
Table 1 Bayes discriminant function coefficient table
The discrimination model that can be obtained Luzhou-flavor, middle odor type, delicate fragrance type and fresh and sweet odor type flue-cured tobacco raw material by table 1 is:
f
Giving off a strong fragrance=-2.703y
2+ 21.561y
3+ 2.319y
7+ 7.094y
8-62.881
f
Middle fragrant=8.494y
2+ 14.239y
3+ 7.34y
7+ 1.767y
8-60.158
f
Delicate fragrance=11.696y
2+ 15.889y
3+ 5.268y
7+ 3.092y
8-84.124
f
Fresh and sweet perfume (or spice)=11.764y
2+ 20.719y
3+ 6.05y
7+ 3.574y
8-113.497
When doing the evaluation of raw tobacco material odor type, with sample x
0Corresponding note quantification of targets data are 4 discriminant function computational discriminations of substitution functional value respectively, if F
k(x
0)-max{F
k(x
0)=0, k ∈ { giving off a strong fragrance, middle fragrant, delicate fragrance, fresh and sweet perfume (or spice) } then thinks x
0Be under the jurisdiction of F relatively
kThe type of representative, that is: which functional value is maximum, and sample just belongs to the odor type of corresponding representative.
3. model is differentiated effect assessment
Utilize self checking modeling sample is returned to declare, to verify the differentiation effect with validation-cross.The utilization cross verification often can draw than the higher False Rate of self proof method, and the confidence level that the result is declared in its time is higher.This instance carries out self checking and validation-cross to 105 modeling samples.It is differentiated the result and sees table 2.
The evaluation of table 2 classification results
Annotate: be actual sample number before judging in the bracket
Self checking and validation-cross return to declare and come to the same thing in the table 2; Differentiation accuracy rate to the Luzhou-flavor sample is 100%; The differentiation accuracy rate of middle odor type sample is 95.12%; The differentiation accuracy rate of delicate fragrance type sample is 96.08%, and the differentiation accuracy rate of fresh and sweet odor type sample is 96.1%, shows that the discrimination model of being set up is better to the differentiation effect of odor type.
4. the tobacco leaf odor type is differentiated
Further adopting said method that 21 flue-cured tobacco samples are carried out odor type differentiates; And identification result compared the evaluation result of sample odor type with the expert that smokes panel test, the result shows that this method is 100% to the coincidence rate of the result of determination of sample and the expert's that smokes panel test evaluation result.
Can know that to sum up this exemplary application sensory evaluating smoking carries out sample note evaluation of indexes and the odor type discrimination model that quantizes to combine progressively discriminatory analysis to make up and comprise delicate fragrance, fragrant and sweet, the banksia rose and four target variables of burnt odor, and model is differentiated effect estimate.The result shows: the model of setting up through this method has higher accuracy rate to distinguishing Luzhou-flavor, middle odor type, delicate fragrance type and fresh and sweet odor type flue-cured tobacco raw material.Therefore, the method can to the index of known sample screen and set up independent variable still less, the odor type decision model of better effects if, for the criticism of flue-cured tobacco odor type provides a kind of quick, accurate and effective new method.
The above only is a preferred implementation of the present invention; Should be pointed out that for those skilled in the art, under the prerequisite that does not break away from know-why of the present invention; Can also make some improvement and distortion, these improvement and distortion also should be regarded as protection scope of the present invention.
Claims (5)
1. the method based on note component structure discrimination model evaluation flue-cured tobacco raw material odor type is characterized in that, may further comprise the steps:
S1, flue-cured tobacco note estimate: adopt the mode of secretly commenting by the technician that smokes panel test, the note index of different flavor raw tobacco material is evaluated and quantized;
The foundation of S2, odor type discrimination model: based on the evaluation and the quantized data of step S1 gained note index; Adopt progressively diagnostic method to the note index screening and carry out the Bayes discriminatory analysis; Make up flue-cured tobacco odor type discrimination model, confirm decision rule, and the differentiation effect of model is estimated;
S3, odor type are differentiated: with each discriminant score of calculation sample in the corresponding note quantification of targets of the sample to be identified data difference substitution discrimination model, and then the odor type of definite sample.
2. make up the method that discrimination model is estimated flue-cured tobacco raw material odor type according to claim 1 is said based on the note component, it is characterized in that the flue-cured tobacco note index among the said step S1 comprises: tobacco perfume, delicate fragrance, fragrant and sweet, fruital, the fragrance of a flower, Xin Xiang, the banksia rose and burnt odor; Said each note index is divided into 6 class: outstanding, outstanding, obviously, obviously, have and do not have, and correspondence was given score value 10 minutes, 8 minutes, 6 minutes, 4 minutes, 2 minutes and 0 minute respectively.
3. make up the method that discrimination model is estimated flue-cured tobacco raw material odor type according to claim 1 is said based on the note component, it is characterized in that the raw tobacco material odor type among the said step S1 comprises Luzhou-flavor, middle odor type, delicate fragrance type and fresh and sweet odor type.
4. make up the method that discrimination model is estimated flue-cured tobacco raw material odor type according to claim 1 is said based on the note component, it is characterized in that the discrimination model among the said step S2 is:
F
Giving off a strong fragrance=-2.703y
2+ 21.561y
3+ 2.319y
7+ 7.094y
8-62.881
F
Middle fragrant=8.494y
2+ 14.239y
3+ 7.34y
7+ 1.767y
8-60.158
F
Delicate fragrance=11.696y
2+ 15.889y
3+ 5.268y
7+ 3.092y
8-84.124
F
Fresh and sweet perfume (or spice)=11.764y
2+ 20.719y
3+ 6.05y
7+ 3.574y
8-113.497.
5. make up the method that discrimination model is estimated flue-cured tobacco raw material odor type according to claim 4 is said based on the note component, it is characterized in that the definite decision rule described in the step S2 is: if F in the model
k(x
0)-max{F
k(x
0)=0, k ∈ { giving off a strong fragrance, middle fragrant, delicate fragrance, fresh and sweet perfume (or spice) } then thinks sample x
0Be under the jurisdiction of F relatively
kThe type of representative; Adopt effect verification methods such as validation-cross and self checking that the differentiation effect of model is estimated among the step S2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101392913A CN102650630A (en) | 2012-05-08 | 2012-05-08 | Method for building distinguishing model for evaluating odor type of flue-cured tobacco raw material on the basis of note components |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101392913A CN102650630A (en) | 2012-05-08 | 2012-05-08 | Method for building distinguishing model for evaluating odor type of flue-cured tobacco raw material on the basis of note components |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102650630A true CN102650630A (en) | 2012-08-29 |
Family
ID=46692677
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012101392913A Pending CN102650630A (en) | 2012-05-08 | 2012-05-08 | Method for building distinguishing model for evaluating odor type of flue-cured tobacco raw material on the basis of note components |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102650630A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103217408A (en) * | 2013-05-14 | 2013-07-24 | 贵州大学 | Method for identifying two flue-cured tobaccos with different odor types in Guizhou |
| CN103226136A (en) * | 2012-10-15 | 2013-07-31 | 红塔烟草(集团)有限责任公司 | Detection and determination method for flavor type of cigarette |
| CN103226135A (en) * | 2012-10-15 | 2013-07-31 | 红塔烟草(集团)有限责任公司 | Detection and determination method for flavor type of flue-cured tobacco |
| CN103344713A (en) * | 2013-06-18 | 2013-10-09 | 湖南中烟工业有限责任公司 | Method for analyzing importance of flavoring in flue-cured tobacco fragrance type classification and forecasting fragrance type based on random forest method |
| CN105241976A (en) * | 2015-10-29 | 2016-01-13 | 云南中烟工业有限责任公司 | Method for characterizing cigarette spicy aroma characteristic by using index |
| CN106047483A (en) * | 2016-06-15 | 2016-10-26 | 河南中烟工业有限责任公司 | Spicy tobacco essence based on spice threshold and mixing method thereof |
| CN106093246A (en) * | 2016-08-11 | 2016-11-09 | 云南中烟工业有限责任公司 | A kind of quantitative forecasting technique of Medicated cigarette spicy |
| CN106198809A (en) * | 2016-07-27 | 2016-12-07 | 河南中烟工业有限责任公司 | A kind of flat-flavor flue-cured tobacco authentication method based on activity threshold |
| CN106290610A (en) * | 2016-07-27 | 2017-01-04 | 河南中烟工业有限责任公司 | A kind of delicate fragrance type flue-cured tobacco authentication method based on activity threshold |
| CN106290611A (en) * | 2016-07-27 | 2017-01-04 | 河南中烟工业有限责任公司 | A kind of Luzhou-flavor flue-cured tobacco authentication method based on activity threshold |
| CN110850032A (en) * | 2019-11-21 | 2020-02-28 | 云南中烟工业有限责任公司 | Method for detecting cigarette quality |
-
2012
- 2012-05-08 CN CN2012101392913A patent/CN102650630A/en active Pending
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103226136A (en) * | 2012-10-15 | 2013-07-31 | 红塔烟草(集团)有限责任公司 | Detection and determination method for flavor type of cigarette |
| CN103226135A (en) * | 2012-10-15 | 2013-07-31 | 红塔烟草(集团)有限责任公司 | Detection and determination method for flavor type of flue-cured tobacco |
| CN103226135B (en) * | 2012-10-15 | 2014-12-03 | 红塔烟草(集团)有限责任公司 | Detection and determination method for flavor type of flue-cured tobacco |
| CN103217408A (en) * | 2013-05-14 | 2013-07-24 | 贵州大学 | Method for identifying two flue-cured tobaccos with different odor types in Guizhou |
| CN103217408B (en) * | 2013-05-14 | 2014-12-31 | 贵州大学 | Method for identifying two flue-cured tobaccos with different odor types in Guizhou |
| CN103344713A (en) * | 2013-06-18 | 2013-10-09 | 湖南中烟工业有限责任公司 | Method for analyzing importance of flavoring in flue-cured tobacco fragrance type classification and forecasting fragrance type based on random forest method |
| CN103344713B (en) * | 2013-06-18 | 2014-08-06 | 湖南中烟工业有限责任公司 | Method for analyzing importance of flavoring in flue-cured tobacco fragrance type classification and forecasting fragrance type based on random forest method |
| CN105241976A (en) * | 2015-10-29 | 2016-01-13 | 云南中烟工业有限责任公司 | Method for characterizing cigarette spicy aroma characteristic by using index |
| CN106047483A (en) * | 2016-06-15 | 2016-10-26 | 河南中烟工业有限责任公司 | Spicy tobacco essence based on spice threshold and mixing method thereof |
| CN106047483B (en) * | 2016-06-15 | 2019-08-27 | 河南中烟工业有限责任公司 | A kind of pungent fragrant odor type flavouring essence for tobacco and its concocting method based on fragrance threshold value |
| CN106198809A (en) * | 2016-07-27 | 2016-12-07 | 河南中烟工业有限责任公司 | A kind of flat-flavor flue-cured tobacco authentication method based on activity threshold |
| CN106290610A (en) * | 2016-07-27 | 2017-01-04 | 河南中烟工业有限责任公司 | A kind of delicate fragrance type flue-cured tobacco authentication method based on activity threshold |
| CN106290611A (en) * | 2016-07-27 | 2017-01-04 | 河南中烟工业有限责任公司 | A kind of Luzhou-flavor flue-cured tobacco authentication method based on activity threshold |
| CN106093246A (en) * | 2016-08-11 | 2016-11-09 | 云南中烟工业有限责任公司 | A kind of quantitative forecasting technique of Medicated cigarette spicy |
| CN110850032A (en) * | 2019-11-21 | 2020-02-28 | 云南中烟工业有限责任公司 | Method for detecting cigarette quality |
| CN110850032B (en) * | 2019-11-21 | 2023-03-31 | 云南中烟工业有限责任公司 | Method for detecting cigarette quality |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102650630A (en) | Method for building distinguishing model for evaluating odor type of flue-cured tobacco raw material on the basis of note components | |
| Qiongyan et al. | Detecting spikes of wheat plants using neural networks with Laws texture energy | |
| CN106897570A (en) | A kind of COPD test system based on machine learning | |
| CN107796782B (en) | Redrying quality stability evaluation method based on tobacco leaf characteristic spectrum consistency measurement | |
| CN103278609A (en) | Meat product freshness detection method based on multisource perceptual information fusion | |
| Kaefer et al. | Beyond the river: underlying determinants of population acoustic signal variability in Amazonian direct-developing Allobates (Anura: Dendrobatoidea) | |
| CN108845085A (en) | A method of tobacco leaf odor type is predicted based on tobacco leaf chemical composition | |
| CN104713849A (en) | Method for quickly predicting tobacco mildew based on near-infrared spectrum analysis technique | |
| CN103091282A (en) | Method for detecting quality of tobacco essence perfume | |
| CN105092789B (en) | Tobacco style changing features trend analysis in a kind of sheet nicotinyl alcohol process | |
| CN103760110A (en) | Method for rapidly identifying meat with different animal sources | |
| CN109297534B (en) | Environmental parameter weight determination method and system for evaluating indoor environmental quality | |
| CN113298373A (en) | Financial risk assessment method, device, storage medium and equipment | |
| CN103892435A (en) | Method for preparing tobacco powder raw materials and stabilizing quality of reconstituted tobacco | |
| CN108414471B (en) | Method for distinguishing sensory characterization information based on near infrared spectrum and sensory evaluation mutual information | |
| CN102435709A (en) | Method for accurately identifying flue-cured tobacco flavor types | |
| CN110132879A (en) | A kind of flue-cured tobacco note discrimination method based near infrared spectrum | |
| CN109324016A (en) | A kind of determination method of redried odor type style | |
| CN104680314A (en) | Method for carrying out sensory evaluation on style characteristic of tobacco leaf | |
| CN117407661B (en) | Data enhancement method for equipment state detection | |
| CN109567746A (en) | A kind of grain roughness method for quantitatively evaluating based on Signal Detection Theory | |
| CN113030007B (en) | Method for rapidly testing quality stability of tobacco essence based on similarity learning algorithm | |
| CN107480630B (en) | Method for zoning forest ecological function by using remote sensing technology | |
| CN113936795A (en) | Quantitative analysis method for risk factors of senile lung cancer onset | |
| CN112016046B (en) | Construction method and device of tobacco shred structure prediction model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120829 |