CN107341613A - A kind of method for aiding in tobacco leaf formulation balance to replace - Google Patents
A kind of method for aiding in tobacco leaf formulation balance to replace Download PDFInfo
- Publication number
- CN107341613A CN107341613A CN201710549990.8A CN201710549990A CN107341613A CN 107341613 A CN107341613 A CN 107341613A CN 201710549990 A CN201710549990 A CN 201710549990A CN 107341613 A CN107341613 A CN 107341613A
- Authority
- CN
- China
- Prior art keywords
- raw material
- replaced
- tobacco leaf
- attribute
- replaceable
- 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.)
- Granted
Links
- 241000208125 Nicotiana Species 0.000 title claims abstract description 41
- 235000002637 Nicotiana tabacum Nutrition 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 37
- 239000000203 mixture Substances 0.000 title claims abstract description 34
- 238000009472 formulation Methods 0.000 title claims abstract description 27
- 239000002994 raw material Substances 0.000 claims abstract description 164
- 239000000463 material Substances 0.000 claims abstract description 26
- 238000011156 evaluation Methods 0.000 claims abstract description 13
- 238000010801 machine learning Methods 0.000 claims abstract description 10
- 238000004519 manufacturing process Methods 0.000 claims description 35
- 235000019504 cigarettes Nutrition 0.000 claims description 12
- 238000003066 decision tree Methods 0.000 claims description 10
- 230000000694 effects Effects 0.000 claims description 8
- 238000005194 fractionation Methods 0.000 claims description 8
- 238000010835 comparative analysis Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 2
- 230000001788 irregular Effects 0.000 claims description 2
- 238000013461 design Methods 0.000 abstract description 5
- 238000012423 maintenance Methods 0.000 abstract description 2
- 238000005457 optimization Methods 0.000 abstract description 2
- 238000012827 research and development Methods 0.000 abstract description 2
- 238000012216 screening Methods 0.000 abstract description 2
- 230000015572 biosynthetic process Effects 0.000 abstract 1
- 238000012360 testing method Methods 0.000 description 11
- 238000004458 analytical method Methods 0.000 description 10
- 230000000391 smoking effect Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 235000019506 cigar Nutrition 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 239000003205 fragrance Substances 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 241000196324 Embryophyta Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000796 flavoring agent Substances 0.000 description 1
- 235000013355 food flavoring agent Nutrition 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06315—Needs-based resource requirements planning or analysis
-
- A—HUMAN NECESSITIES
- A24—TOBACCO; CIGARS; CIGARETTES; SIMULATED SMOKING DEVICES; SMOKERS' REQUISITES
- A24C—MACHINES FOR MAKING CIGARS OR CIGARETTES
- A24C5/00—Making cigarettes; Making tipping materials for, or attaching filters or mouthpieces to, cigars or cigarettes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Educational Administration (AREA)
- Animal Husbandry (AREA)
- General Health & Medical Sciences (AREA)
- Mining & Mineral Resources (AREA)
- Marine Sciences & Fisheries (AREA)
- Agronomy & Crop Science (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Manufacture Of Tobacco Products (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of method for aiding in tobacco leaf formulation balance to replace, and artificial posterior infromation of replacing is converted into screening conditions, the raw material sorting technique of formation rule;Interchangeable material is chosen from on-hand inventory material, screen and compare with clustering algorithm, the replaceable material that will be provided with most common traits enters row information submission as optional replacement raw material, while replaces formula record using history, and the deviation being likely to occur in Feature Selection is modified and filtered;Comprehensive two kinds of the selection results, substitute into the environment of complete formula demand, with the method for evaluation and forecast of machine learning, by raw material data correlation complete formula design data and specialty evaluation data, scoring is predicted to the replaceable raw material filtered out, final optimization pass selects reliable replaceable material selective listing.The present invention provides for staff and intuitively replaces selection, is replaced for formula, composition maintenance and formula research and development provide scientific, information-based data prediction support.
Description
Technical field
The present invention relates to a kind of method for aiding in tobacco leaf formulation balance to replace, belong to cigarette tobacco leaf formulation technical field.
Background technology
Tobacco is as a kind of agricultural product, by itself gene, cultivation step, edaphic condition, climatic factor and modulation methods
Many influences such as method, different cultivars, the different regions even tobacco leaf of same plant different parts, in quality and style all
There is larger difference.Therefore, the various fine quality factors of tobacco leaf, it is impossible to exist only in same or a few tobacco leaf
In.The tobacco leaf of one kind, its quality factor are often conflicting.Such as some tabacco fragrances are sufficient, but miscellaneous gas is heavier, thorn
It is larger to swash property;Some tobacco leaf strength are moderate, but fragrance is not sufficient enough.Use volume made of the tobacco leaf of single variety or single-grade
Cigarette, there is the mass defect that can not be overcome.Even high-grade tobacco leaf, it is also difficult to obtain completely satisfactory effect.This is just
The different qualities using various tobacco leaves are needed, optimal tobacco leaf formulation combination is chosen, the various tobacco leaves of participation formula is made the best use of the advantages
Keep away it is short, complement each other, and in concert with play respective effect.
Numerous studies and application have been done in tobacco leaf formulation design aspect by China, but do not obtain practical application, or rely on
The Conventional wisdom of Yu designer designs cigarette product, and this Conventional wisdom thinks:Can according to same area, same breed,
Same area, different grades of tobacco leaf carry out mixture;Can also be stronger to different zones, same breed, same area, suitability
Tobacco leaf mixture etc..But it was verified that this traditional type empirical formula is unfavorable for the exploitation of cigarette new product, so being set in experience
On the basis of meter, research algorithm design cigarette product is the inexorable trend of cigarette development.
The content of the invention
In view of the above-mentioned deficiencies in the prior art, it is an object of the present invention to provide a kind of side for aiding in tobacco leaf formulation balance to replace
Method, complicated to solve traditional raw material tobacco leaf formulation balance replacement method process, combined influence factor need to be taken into consideration, test result
Accuracy rate fluctuation is larger, the problem of wasting time and energy.
The present invention adopts the following technical scheme that:A kind of method for aiding in tobacco leaf formulation balance to replace, is by large batch of
Historical data and current creation data, including the related data such as raw material, formula, inherent quality evaluation, to data cleansing and are extracted
Incidence relation is studied, the influence angle analysis with material composition change to formula adjustment, while analyzes composition of raw materials change, record
To keep the constant of endoplasm after analysis formula change, essence and flavoring agent, smoking material and processing parameter adjustment are commented for endoplasm
The positive and negative influence algorithm application of valency, it is eventually found material quality information and replacement method is balanced to raw material leaf group.Specifically include following
Step:
(1)Rule-based raw material classification:All data that raw material history is replaced are collected, and comparative analysis history replaces number
The frequency of each attribute change, obtains inventory information according to the frequency and raw material attribute identity information is interchangeable to raw material heavy in
The sequence of the property wanted degree;Further according to the sequence of importance degree, decision tree is established, using the first important attribute as to prediction effect
Improved maximum independent variable, is first used fractionation node, and the second important attribute then is carried out into second again splits, with this
Analogize, eventually find interchangeable raw material;
(2)Raw material classification based on clustering algorithm:The existing available stock raw material of summarizing, establishes hyperspace system, with original
The characteristic identity information of material generates multiple space coordinates points as dimension, each raw material according to the dimension of different characteristic attribute,
The mutual dimension of coordinate points is more close, and the characteristic attribute of raw material is then more close, and the accuracy being replaced is higher;Calculated by clustering
Method selects the approximate convergent raw material of various dimensions feature as replaceable alternative in the characteristic attribute of existing raw material;Then using going through
History replaces record amendment cluster deviation, obtains the replaceable raw material in part;
(3)Evaluation and foreca based on machine learning algorithm:By step(1)With step(2)In replaceable raw material result combine take
Orthogonal set, gained set are final replaceable materials list.
The step(1)The frequency in the frequency it is higher, then the influence degree that the attributive character is replaced to raw material is lower, weight
The property wanted is also low.
The step(1)Inventory information refer to whether possess stock.Stock is the deciding factor for influenceing to replace, and is had no
Query is consistent with actual conditions, is that can not carry out raw material replacement at all in the case of no stock.
The step(1)Raw material attribute identity information include whether associate the trade mark, kind information, the raw material two level place of production,
Time, the raw material three-level place of production and price.
Wherein, if the association trade mark is the data that the trade mark associated according to raw material generates, and raw material is given birth to available for multiple trades mark
Production, then it is the association trade mark;Raw material does not associate then only for the production of a trade mark cigarette.Next to that the influence of material trade mark association,
If certain raw material replaced is relevant with other multiple trades mark, in the case of considering the preferential use of other trades mark when replacing
Do not replace as far as possible;3rd is the factor that kind is replaced, and raw material attribute has certain similitude between some kinds, for
Appropriate selection should be given when changing;The influence in the certain place of production, time and price is also very important, adjacent or close place of production material quality
More approximate, replacement is can also to pay the utmost attention to.
The step(2)The characteristic identity information of raw material include time, the place of production, grade, position etc..
Described step(2)Clustering algorithm refer to hyperspace system algorithm, this hyperspace system algorithm is a kind of
Fuzzy, quick algorithm;Specific algorithm is:With the characteristic identity information of raw material needed for formula replacement(Time, position, place of production etc.)
As different latitude discrete point, the set that packet count is carried out under same latitude is given to discrete data point and is divided, formed irregular vertical
Body figure, and a numerical point k is randomly selected in the data set under same latitude, k is calculated into next random number distance
Heart point, and the central point is set to k;Next point is randomly selected out in the form of local weighted(3rd point), k is calculated to newly
The central point ko of point;Successively calculate k to all non-midpoints distance and L and ko to all non-midpoints distance and Lo;If L>
Lo, then assigning ko to k (k=ko), complete to calculate when to all set, final return of each set obtains a midpoint and preserved,
The central point of last each dimension correspondingly feeds back to the characteristic identity information that each dimension defines, according to the central value calculated in stock
Raw material traversal searches identical or closest value, and the characteristic identity information of comprehensive each dimension, which provides, there may be raw materials inventory, or
With the most similar raw materials inventory of characterization factor in result of calculation, and it is ranked up according to phase close values.
The step(3)Gained set substitute into the environment of complete formula replacement demand, the corresponding volume for replacing raw material
Cigarette formula carries out specialty evaluation score value and compared, and is verified;Or with reference to historical data, using step(2)Middle clustering algorithm,
Algorithm is included to gained set and carries out constantly circulation checking, to determine the accuracy of balance formula replacement method and stably
Property, and Substitution Rules are constantly summarized in a manner of machine learning during replacement, optimize the method model of replacement, lifting is auxiliary
Help the accurate and reliability of tobacco leaf formulation balance replacement method.
Beneficial effects of the present invention:The present invention is selected in on-hand inventory, analyzed, evaluation and foreca, and optimal screening goes out optimal
Range of choice.Provided for staff and intuitively replace selection, replaced for formula, composition maintenance and formula research and development provide science
Change, the support of information-based data prediction, solve the problems such as blindness, efficiency is low, improve the service efficiency of formula material,
And then the design of tobacco leaf formulation is optimized, so as to reach the purpose of optimization lifting cigarette inherent quality.
Brief description of the drawings
Fig. 1 is that the raw material of embodiment 1 replaces variable importance prognostic chart;
Fig. 2 is that the raw material of embodiment 1 replaces variable frequency prognostic chart;
Fig. 3 is the decision tree of embodiment 1;
Fig. 4 is the specification A of embodiment 1, No. 1 module artificial experience raw material replacement example;
Fig. 5 is the step of embodiment 1(3)The schematic diagram for taking orthogonal set;
Fig. 6 is that the raw material of embodiment 2 replaces variable importance prognostic chart;
Fig. 7 is that the raw material of embodiment 2 replaces variable frequency prognostic chart;
Fig. 8 is the decision tree of embodiment 2;
Fig. 9 is the specification B of embodiment 2, No. 1 module artificial experience raw material replacement example.
Embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples, but embodiment is not to the present invention
The restriction of technical scheme.
Embodiment 1
With xx cigar mills specification A, " the red big/WBBSF/FL/P/ " raw materials in 2013 flue-cured tobaccos/Kunming 2/ enter exemplified by replacing in No. 1 module
Row explanation:
(1)Rule-based raw material classification:All data that raw material history is replaced are collected, and comparative analysis history replaces number
The frequency of each attribute change, obtains inventory information according to the frequency and raw material attribute identity information is interchangeable to raw material heavy in
The sequence of the property wanted degree;Further according to the sequence of importance degree, decision tree is established, using the first important attribute as to prediction effect
Improved maximum independent variable, is first used fractionation node, and the second important attribute then is carried out into second again splits, with this
Analogize, eventually find interchangeable raw material;
Laterally to analysis:All data that raw material history is replaced are collected, each category in comparative analysis history replacement data
Property change the frequency, replace the frequency it is higher, illustrate that the influence degree that the attributive character is replaced to raw material is lower, importance is low.
Analysis result:Inventory information and raw material attribute identity information are followed successively by the interchangeable importance degree of raw material to be possessed
Stock, the trade mark, kind information, the raw material two level place of production, time, the raw material three-level place of production and price whether are associated, wherein whether associating
The trade mark is the data that the trade mark that system associates according to raw material generates, and raw material produces available for multiple trades mark, then is the association trade mark;It is former
Expect the production only for a trade mark cigarette, then do not associate.Wherein stock is the deciding factor for influenceing to replace, certainly with reality
Border situation is consistent, and is that can not carry out raw material replacement at all in the case of no stock;Next to that the shadow of material trade mark association
Ring, if certain raw material replaced is relevant with other multiple trades mark, the feelings that are preferentially used in view of other trades mark when replacing
Do not replaced as far as possible under condition;3rd is the factor that kind is replaced, and raw material attribute has certain similitude between some kinds,
Appropriate selection should be given when replacing;The influence in the certain place of production, time and price is also very important, adjacent or close place of production raw material
Quality is more approximate, and replacement is can also to pay the utmost attention to.Such as Fig. 1 and 2.
Draw a conclusion:When replacing, time replacement is the factor that relative priority considers, next to that kind and three-level production
The replacement on ground, finally just consider to replace the two level place of production of raw material when other situations can not all meet.
Establish decision tree forecast model:By the analysis to the artificial replacement experience of history, according to raw material attribute to replacement
Influence degree size sorts, and establishes decision tree(Such as Fig. 3):Possess stock as to the improved maximum independent variable of prediction effect,
It is first used fractionation node;Then according to whether association, other trades mark split for the second time again, and then again to two level production
Ground carries out third time fractionation, has eventually found interchangeable raw material.It can be seen that top-priority factor is respectively when the raw material is replaced
Stock, trade mark related information and the two level place of production, are screened layer by layer in such a manner, will artificially replace empirical conversion as can
The raw material of replacement becomes more regular, filters out the replaceable material in part.
As shown in figure 4, replaced for artificial experience.It is as follows with replacing raw material contrast to be replaced raw material:
It can be seen that this time manually replace selection is same kind, same grade, the same place of production, the raw material of different year is replaced
Change.And a kind of method for aiding in tobacco leaf formulation balance to replace, just using similar artificial replacement experience as rule, with reference to replacing for history
Record is changed, obtains the optional alternate material in part.
(2)Raw material classification based on clustering algorithm:The existing available stock raw material of summarizing, establishes hyperspace system,
With the characteristic identity information of raw material(Time, the place of production, grade, position etc.)As dimension, each raw material is according to different characteristic category
Property dimension generate multiple space coordinates points, the mutual dimension of coordinate points is more close, and the characteristic attribute of raw material is then more close, quilt
The accuracy of replacement is higher;If the time is a dimension, grade is a dimension, and position is dimension etc. again.And these are former
The specifying information of material characteristic attribute is exactly scale in this dimension, i.e., the scale in time dimension be 2009,2010,
2012 ..., the scale of position dimension had top, middle part, bottom ....
Illustrate, the dimension in hyperspace system is not transverse and longitudinal coordinate, if transverse and longitudinal coordinate have to be expressed as, then
Any one dimension can be abscissa or ordinate in hyperspace system.That is, hyperspace system
In, transverse and longitudinal coordinate is relative, as long as two dimension perpendiculars, then the two dimensions can serve as the x in the side, y-axis.
Point in space:A raw material is chosen, the specific descriptions of its identity information can be generated in space according to different scales
One point, what this point represented is exactly this raw material.
Selecting in the characteristic attribute of existing raw material that the approximate convergent raw material of various dimensions feature is used as by clustering algorithm can
Replace alternative;Then record amendment cluster deviation is replaced using history, obtains the replaceable raw material in part;That is, by establishing multidimensional
Three-dimensional system, it is determined that meet the raw material position in space of production inventory requirement, then by calculating in space two not
With the beeline put, to judge the similarity of raw material.
Make same latitude(Same time, same part etc.)Coordinate points as one set, randomly select a numerical point K, count
K is calculated to next random data distance center point, and the central point is set to k;Under being randomly selected out in the form of local weighted
One point(3rd point), calculate k to the central point ko newly put;K is calculated successively to arrive to the distance and L and ko at all non-midpoints
The distance and Lo at all non-midpoints.If L>Lo, then assign K=Ko;If L<Lo, then K=K.That is, take in distance and minimum
Relative central point K of the heart point as the set.
Finally draw a conclusion:When it is all set completions central point is calculated, begin to calculate the distance of these central points,
Obtain new central point.Constantly repeat until finally returning the set more concentrated, the represented raw material letter of this set
Breath is namely based on the replaceable material in part that the raw material sorting technique of clustering algorithm filters out.
(3)Evaluation and foreca based on machine learning algorithm:By step(1)With step(2)In replaceable raw material result knot
The orthogonal set of conjunction, such as Fig. 5, gained set are final replaceable materials list.
Measuring and calculating is combined with machine learning algorithm by empirical rule and draws replaceable raw material, and the special of raw material is replaced with it
Industry evaluation is established score value and compared, and wherein the source of specialty evaluation is that internal specialty is smoked panel test the daily work of smokeing panel test of Shi Jinhang, knot of smokeing panel test
Result data is recorded and uploaded by specialty software of smokeing panel test by fruit.Can be replaced raw material smoke panel test data with it is actually replacing after
Smoking result data carry out fraction contrast, to verify the Stability and veracity of balance formula replacement method.Such as following table:
It can be seen that occur in the replaceable raw material filtered out by a kind of method for aiding in tobacco leaf formulation balance to replace shown in Fig. 4
The raw material that artificial experience is replaced, and its preferred sequence comes second, illustrates that the method that this auxiliary tobacco leaf formulation balance is replaced has
Certain accuracy.
Embodiment 2
With xx cigar mills specification B, " 2014 flue-cured tobaccos/Qujing 2/K326/WDC3F/FW/P/ " raw materials enter exemplified by replacing in No. 1 module
Row explanation:
(1)Rule-based raw material classification:All data that raw material history is replaced are collected, and comparative analysis history replaces number
The frequency of each attribute change, obtains inventory information according to the frequency and raw material attribute identity information is interchangeable to raw material heavy in
The sequence of the property wanted degree;Further according to the sequence of importance degree, decision tree is established, using the first important attribute as to prediction effect
Improved maximum independent variable, is first used fractionation node, and the second important attribute then is carried out into second again splits, with this
Analogize, eventually find interchangeable raw material;
Laterally to analysis:The raw material history of formula is replaced all raw material data that are replaced in record to be collected, to score
Analyse the frequency of each attribute change in history replacement data, that is, after replacing raw material, frequency number that raw material attribute changes, institute
It is higher to replace the frequency, illustrates that the influence degree that the attributive character is replaced to raw material is lower, importance is low.
Analysis result:Inventory information and raw material attribute identity information are followed successively by the interchangeable importance degree of raw material to be possessed
Stock, the trade mark, kind information, the raw material two level place of production, time, the raw material three-level place of production and price whether are associated, wherein whether associating
The trade mark is the data that the trade mark that system associates according to raw material generates, and raw material produces available for multiple trades mark, then is the association trade mark;It is former
Expect the production only for a trade mark cigarette, then do not associate.Wherein stock is the deciding factor for influenceing to replace, certainly with reality
Border situation is consistent, and is that can not carry out raw material replacement at all in the case of no stock;Next to that the shadow of material trade mark association
Ring, if certain raw material replaced is relevant with other multiple trades mark, the feelings that are preferentially used in view of other trades mark when replacing
Do not replaced as far as possible under condition;3rd is the factor that kind is replaced, and raw material attribute has certain similitude between some kinds,
Appropriate selection should be given when replacing;The influence in the certain place of production, time and price is also very important, adjacent or close place of production raw material
Quality is more approximate, and replacement is can also to pay the utmost attention to.Such as Fig. 6 and 7.
Draw a conclusion:When replacing, time replacement is the factor that relative priority considers, next to that kind and three-level production
The replacement on ground, finally just consider to replace the two level place of production of raw material when other situations can not all meet.
Establish decision tree forecast model:By the analysis to the artificial replacement experience of history, according to raw material attribute to replacement
Influence degree size sorts, and establishes decision tree(Such as Fig. 8):Possess stock as to the improved maximum independent variable of prediction effect,
It is first used fractionation node;Then according to whether association, other trades mark split for the second time again, and then again to two level production
Ground carries out third time fractionation, has eventually found interchangeable raw material.It can be seen that top-priority factor is respectively when the raw material is replaced
Stock, trade mark related information and the two level place of production, are screened layer by layer in such a manner, will artificially replace empirical conversion as can
The raw material of replacement becomes more regular, filters out the replaceable material in part.
As shown in figure 9, replaced for people's experience.It is as follows with replacing raw material contrast to be replaced raw material:
It can be seen that this time manually replace selection is same kind, same grade, same time, the raw material of different sources is replaced
Change.And a kind of method for aiding in tobacco leaf formulation balance to replace, just using similar artificial replacement experience as rule, with reference to replacing for history
Record is changed, obtains the optional alternate material in part.
(2)Raw material classification based on clustering algorithm:The existing available stock raw material of summarizing, establishes hyperspace system,
With the characteristic identity information of raw material(Time, the place of production, grade, position etc.)As dimension, each raw material is according to different characteristic category
Property dimension generate multiple space coordinates points, the mutual dimension of coordinate points is more close, and the characteristic attribute of raw material is then more close, quilt
The accuracy of replacement is higher;If the time is a dimension, grade is a dimension, and position is dimension etc. again.And these are former
The specifying information of material characteristic attribute is exactly scale in this dimension, i.e., the scale in time dimension be 2009,2010,
2012 ..., the scale of position dimension had top, middle part, bottom ....
Illustrate, the dimension in hyperspace system is not transverse and longitudinal coordinate, if transverse and longitudinal coordinate have to be expressed as, then
Any one dimension can be abscissa or ordinate in hyperspace system.That is, hyperspace system
In, transverse and longitudinal coordinate is relative, as long as two dimension perpendiculars, then the two dimensions can serve as the x in the side, y-axis.
Point in space:A raw material is chosen, the specific descriptions of its identity information can be generated in space according to different scales
One point, what this point represented is exactly this raw material.
Selecting in the characteristic attribute of existing raw material that the approximate convergent raw material of various dimensions feature is used as by clustering algorithm can
Replace alternative;Then record amendment cluster deviation is replaced using history, obtains the replaceable raw material in part;That is, by establishing multidimensional
Three-dimensional system, it is determined that meet the raw material position in space of production inventory requirement, then by calculating in space two not
With the beeline put, to judge the similarity of raw material.
Make same latitude(Same time, same part etc.)Coordinate points as one set, randomly select a numerical point K, count
K is calculated to next random data distance center point, and the central point is set to k;Under being randomly selected out in the form of local weighted
One point(3rd point), calculate k to the central point ko newly put;K is calculated successively to arrive to the distance and L and ko at all non-midpoints
The distance and Lo at all non-midpoints.If L>Lo, then assign K=Ko;If L<Lo, then K=K.That is, take in distance and minimum
Relative central point K of the heart point as the set.
Finally draw a conclusion:When it is all set completions central point is calculated, begin to calculate the distance of these central points,
Obtain new central point.Constantly repeat until finally returning the set more concentrated, the represented raw material letter of this set
Breath is namely based on the replaceable material in part that the raw material sorting technique of clustering algorithm filters out.
(3)Evaluation and foreca based on machine learning algorithm:By step(1)With step(2)In replaceable raw material result knot
The orthogonal set of conjunction, gained set are final replaceable materials list.
Measuring and calculating is combined with machine learning algorithm by empirical rule and draws replaceable raw material, and the special of raw material is replaced with it
Industry evaluation is established score value and compared, and wherein the source of specialty evaluation is that internal specialty is smoked panel test the daily work of smokeing panel test of Shi Jinhang, knot of smokeing panel test
Result data is recorded and uploaded by specialty software of smokeing panel test by fruit.Can be replaced raw material smoke panel test data with it is actually replacing after
Smoking result data carry out fraction contrast, to verify the Stability and veracity of balance formula replacement method.Such as following table:
It can be seen that occur in the replaceable raw material filtered out by a kind of method for aiding in tobacco leaf formulation balance to replace shown in Fig. 9
The raw material that artificial experience is replaced, and its preferred sequence comes first, illustrates that the method that this auxiliary tobacco leaf formulation balance is replaced has
Certain accuracy.
All in all, artificial experience is replaced mainly is carried out from the side such as the place of production, time.And the present invention combines artificial replace
Empirical rule and clustering algorithm, along with the method for machine learning forecast analysis, it is excellent to provide multiple replaceable raw materials for enterprise
Gather sequence.Meanwhile the replaceable raw material filtered out, by replacing experience contrast verification with artificial, the raw material for finding manually to replace exists
At first three in the replaceable raw material sequence filtered out, and a kind of method for aiding in tobacco leaf formulation balance to replace passes through machine learning
The fraction of prediction is about 80%-the 90% of true score, illustrate it is a kind of aid in tobacco leaf formulation balance replace method accuracy compared with
It is high.
Claims (7)
- A kind of 1. method for aiding in tobacco leaf formulation balance to replace, it is characterised in that comprise the following steps:(1)Rule-based raw material classification:All data that raw material history is replaced are collected, and comparative analysis history replaces number The frequency of each attribute change, obtains inventory information according to the frequency and raw material attribute identity information is interchangeable to raw material heavy in The sequence of the property wanted degree;Further according to the sequence of importance degree, decision tree is established, using the first important attribute as to prediction effect Improved maximum independent variable, is first used fractionation node, and the second important attribute then is carried out into second again splits, with this Analogize, eventually find interchangeable raw material;(2)Raw material classification based on clustering algorithm:The existing available stock raw material of summarizing, establishes hyperspace system, with original The characteristic identity information of material generates multiple space coordinates points as dimension, each raw material according to the dimension of different characteristic attribute; Selected by clustering algorithm in the characteristic attribute of existing raw material various dimensions feature approximate convergent raw material be used as it is alternatively alternative; Then record amendment cluster deviation is replaced using history, obtains the replaceable raw material in part;(3)Evaluation and foreca based on machine learning algorithm:By step(1)With step(2)In replaceable raw material result combine take Orthogonal set, gained set are final replaceable materials list.
- 2. the method that auxiliary tobacco leaf formulation balance according to claim 1 is replaced, it is characterised in that:The step(1)'s The frequency is higher in the frequency, then the influence degree that the attributive character is replaced to raw material is lower.
- 3. the method that auxiliary tobacco leaf formulation balance according to claim 1 is replaced, it is characterised in that:The step(1)'s Inventory information refers to whether possess stock.
- 4. the method that auxiliary tobacco leaf formulation balance according to claim 1 is replaced, it is characterised in that:The step(1)'s Raw material attribute identity information includes whether to associate the trade mark, kind information, the raw material two level place of production, time, the raw material three-level place of production and valency Lattice.
- 5. the method that auxiliary tobacco leaf formulation balance according to claim 1 is replaced, it is characterised in that:The step(2)'s The characteristic identity information of raw material includes time, the place of production, grade, position.
- 6. the method that auxiliary tobacco leaf formulation balance according to claim 1 is replaced, it is characterised in that:Described step(2) Clustering algorithm refer to hyperspace system algorithm, specific algorithm is:Using formula replace needed for raw material characteristic identity information as Different latitude discrete point, the set that packet count is carried out under same latitude is given to discrete data point and is divided, forms irregular stereogram, And a numerical point k is randomly selected in the data set under same latitude, calculating k to next random number distance central point, And the central point is set to k;Next point is randomly selected out in the form of local weighted, calculates k to the central point ko newly put;According to Secondary calculating k to all non-midpoints distance and L and ko to all non-midpoints distance and Lo;If L>Lo, then assign ko to k (k =ko), complete to calculate when to all set, final return of each set obtains a midpoint and preserved, in last each dimension Heart point correspondingly feeds back to the characteristic identity information that each dimension defines, identical in raw materials inventory traversal search according to the central value calculated Or closest value, the characteristic identity information of comprehensive each dimension, which provides, there may be raw materials inventory, or with feature in result of calculation The most similar raw materials inventory of the factor, and be ranked up according to phase close values.
- 7. the method that auxiliary tobacco leaf formulation balance according to claim 1 is replaced, it is characterised in that:The step(3)'s Gained set is substituted into the environment of complete formula replacement demand, and the corresponding cigarette composition for replacing raw material carries out specialty evaluation score value Compare, verified;Or with reference to historical data, using step(2)Middle clustering algorithm, algorithm progress is included to gained set Constantly circulation checking.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710549990.8A CN107341613B (en) | 2017-07-07 | 2017-07-07 | Method for assisting balance replacement of leaf group formula |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710549990.8A CN107341613B (en) | 2017-07-07 | 2017-07-07 | Method for assisting balance replacement of leaf group formula |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107341613A true CN107341613A (en) | 2017-11-10 |
CN107341613B CN107341613B (en) | 2021-05-25 |
Family
ID=60219610
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710549990.8A Active CN107341613B (en) | 2017-07-07 | 2017-07-07 | Method for assisting balance replacement of leaf group formula |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107341613B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110161011A (en) * | 2019-06-12 | 2019-08-23 | 浙江中烟工业有限责任公司 | Rapid detection method of the Semblance to the essence spice for cigarette of the amounts of mixing different in production |
CN110244670A (en) * | 2019-05-31 | 2019-09-17 | 山东中烟工业有限责任公司 | A kind of volume packet procedures technical parameter management-control method and system |
CN110250553A (en) * | 2019-06-25 | 2019-09-20 | 红云红河烟草(集团)有限责任公司 | Formula replacement method for maintaining stable quality of cigarette cut tobacco |
CN112395553A (en) * | 2019-08-16 | 2021-02-23 | 湖南中烟工业有限责任公司 | Digital flavoring method based on empirical formula network migration model |
CN112397156A (en) * | 2019-07-31 | 2021-02-23 | 湖南中烟工业有限责任公司 | Digital flavoring method based on K-means clustering |
CN112508470A (en) * | 2019-09-16 | 2021-03-16 | 比亚迪股份有限公司 | Method and device for changing bill of material, storage medium and electronic equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1524460A (en) * | 2003-02-25 | 2004-09-01 | 中国海洋大学 | Method for establishing mixed expert system of maintaining cigarette leaf group formulation |
CN101419454A (en) * | 2008-12-04 | 2009-04-29 | 哈尔滨工程大学 | Cigarette recipe maintenance method based on artificial immunity method |
US20120210303A1 (en) * | 2008-06-06 | 2012-08-16 | Apple Inc. | System and method for revising boolean and arithmetic operations |
CN103020765A (en) * | 2012-12-10 | 2013-04-03 | 红塔烟草(集团)有限责任公司 | Tobacco leaf raw material dynamic balancing method |
CN106096748A (en) * | 2016-04-28 | 2016-11-09 | 武汉宝钢华中贸易有限公司 | Entrucking forecast model in man-hour based on cluster analysis and decision Tree algorithms |
-
2017
- 2017-07-07 CN CN201710549990.8A patent/CN107341613B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1524460A (en) * | 2003-02-25 | 2004-09-01 | 中国海洋大学 | Method for establishing mixed expert system of maintaining cigarette leaf group formulation |
US20120210303A1 (en) * | 2008-06-06 | 2012-08-16 | Apple Inc. | System and method for revising boolean and arithmetic operations |
CN101419454A (en) * | 2008-12-04 | 2009-04-29 | 哈尔滨工程大学 | Cigarette recipe maintenance method based on artificial immunity method |
CN103020765A (en) * | 2012-12-10 | 2013-04-03 | 红塔烟草(集团)有限责任公司 | Tobacco leaf raw material dynamic balancing method |
CN106096748A (en) * | 2016-04-28 | 2016-11-09 | 武汉宝钢华中贸易有限公司 | Entrucking forecast model in man-hour based on cluster analysis and decision Tree algorithms |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110244670A (en) * | 2019-05-31 | 2019-09-17 | 山东中烟工业有限责任公司 | A kind of volume packet procedures technical parameter management-control method and system |
CN110161011A (en) * | 2019-06-12 | 2019-08-23 | 浙江中烟工业有限责任公司 | Rapid detection method of the Semblance to the essence spice for cigarette of the amounts of mixing different in production |
CN110250553A (en) * | 2019-06-25 | 2019-09-20 | 红云红河烟草(集团)有限责任公司 | Formula replacement method for maintaining stable quality of cigarette cut tobacco |
CN112397156A (en) * | 2019-07-31 | 2021-02-23 | 湖南中烟工业有限责任公司 | Digital flavoring method based on K-means clustering |
CN112397156B (en) * | 2019-07-31 | 2022-08-16 | 湖南中烟工业有限责任公司 | Digital flavoring method based on K-means clustering |
CN112395553A (en) * | 2019-08-16 | 2021-02-23 | 湖南中烟工业有限责任公司 | Digital flavoring method based on empirical formula network migration model |
CN112508470A (en) * | 2019-09-16 | 2021-03-16 | 比亚迪股份有限公司 | Method and device for changing bill of material, storage medium and electronic equipment |
Also Published As
Publication number | Publication date |
---|---|
CN107341613B (en) | 2021-05-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107341613A (en) | A kind of method for aiding in tobacco leaf formulation balance to replace | |
CN104063599B (en) | Index screening and processing method for evaluating quality of tobacco leaves | |
CN101414183B (en) | Cigarette working procedure quality overall evaluation system and method based on gray correlation analysis | |
CN105844300A (en) | Optimized classification method and optimized classification device based on random forest algorithm | |
CN103844344A (en) | Method for regulating and controlling tobacco shred quality uniformity of different batches of cigarettes and application of method | |
CN108132964A (en) | A kind of collaborative filtering method to be scored based on user item class | |
CN104820684B (en) | A kind of quick online analysis and processing method based on locus | |
CN110287423A (en) | A kind of farm Products Show system and method based on collaborative filtering | |
CN103070465A (en) | Tobacco leaf composition blending method based on compatibility | |
CN107897995A (en) | A kind of segmentation split based on former cigarette formula module is homogenized regulation and control method | |
CN110348480A (en) | A kind of non-supervisory anomaly data detection algorithm | |
CN106645530B (en) | A method of the multi-model based on tobacco leaf aroma component evaluates raw tobacco material similarity | |
CN108305195A (en) | A kind of comprehensive index system towards students in middle and primary schools' evaluation and theme attribute analysis | |
CN111652516A (en) | Tobacco base applicability evaluation method based on formula efficacy | |
CN104572900B (en) | The properties and characteristicses system of selection that a kind of crop breeding is evaluated | |
CN107767676A (en) | A kind of method and apparatus for contributing to Traffic signal control | |
CN109902898B (en) | Process capability evaluation method based on tip cutting, leaf threshing and redrying production | |
CN103593561B (en) | Method for representing style characteristics of tobacco leaves by using characteristic index | |
CN109858541A (en) | A kind of specific data self-adapting detecting method based on data integration | |
CN101866368A (en) | Method for carrying out computer assisted design of tobacco group formula by near infrared spectrum technology | |
CN115293444A (en) | Characterization method of health index of stored tobacco raw materials | |
CN114780599A (en) | Comprehensive analysis system based on wheat quality ratio test data | |
Achentalika et al. | Competitiveness of East African Exports: A Constant Market Share Analysis | |
CN110286663B (en) | Regional cigarette physical index standardized production improving method | |
CN114246356A (en) | Design method, system, medium and device of cigarette leaf group formula |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |