A kind of multilayer elite role method excavated for traditional Chinese medical science case history diagnostic rule
Technical field:
The present invention relates to traditional Chinese medical science information intelligent process field, relate in particular to a kind of for traditional Chinese medical science case history diagnostic rule
The multilayer elite role method of excavation.
Background technology:
The traditional Chinese medical science has great contribution to the world as the distinctive culture of the Chinese nation and scientific heritage.Tcm diagnosis master
If being carried out by processes such as " hope, hear, ask, cut " according to patient symptom, due to crisscross intersection between TCM Syndrome Type and symptom, according to
By computer system carry out the traditional Chinese medical science it is accurate it is dialectical recognize control relatively difficult, but a large amount of useful letters are store in traditional Chinese medical science medical records system
Breath, extracts wherein important tcm diagnosis rule, carries out classification and the rule digging of disease of tcm, examines improving traditional Chinese medical science disease
Disconnected final result and outcomes important in inhibiting and value.
Due to there is bulk redundancy case history category in ambiguity, the uncertain and traditional Chinese medical science database of case history of traditional Chinese medical science case history
Property, it is relatively difficult to want set up a traditional Chinese medical science case history intellectuality assistant diagnosis system to carry out illness rule digging.Particularly
, there is complicated correlation and interdependency between its attribute, using general rule method for digging centering in some traditional Chinese medical science case histories
Cure the desease to go through and pre-processed and classified, it is difficult to excavate wherein actually useful diagnostic rule.Other general rule method for digging
The complicated inside-connecting relation of various dimensions between traditional Chinese medical science case history attribute cannot be often processed, and it cannot be guaranteed that the final traditional Chinese medical science case history extracted is examined
Disconnected rule is the optimal rules feature set of required target, and these excavate rule such as doctor aid in diagnosis using by
Its ordinary conditions can be disturbed to judge, it is also possible to causing mistaken diagnosis when serious.Therefore diagnostic rule is not being influenceed to excavate precision premise
Under, we must seek a kind of traditional Chinese medical science case history rule digging side that can effectively process complicated correlation and interdependency attribute
Method.
The content of the invention:
Deficiency instant invention overcomes more than provides a kind of multilayer elite angle excavated for traditional Chinese medical science case history diagnostic rule
Color method.
The present invention is realized by following technical scheme:
A kind of multilayer elite role method excavated for traditional Chinese medical science case history diagnostic rule, comprises the following steps that:
A, traditional Chinese medical science case history attribute is assigned in different " common-elite " role's Advanced group species, in carrying out traditional Chinese medical science case history
The related and attribute pretreatment that interdepends, by the elite concentration ECD (Elitist for designing elite role populationi) and selection is generally
Rate P (xi), ElitistiIt is i-th elite role population, xiIt is for i & lt carries out traditional Chinese medical science case history selection, attribute is related and mutual
The traditional Chinese medical science case history of dependence is distributed to elite role population, will be without selected probability P (xi) the general traditional Chinese medical science case history attribute chosen
In distribution to conventional character population;
A kind of multilayer elite role's dynamic equalization strategy excavated for traditional Chinese medical science case history diagnostic rule of B, construction, by i-th
J-th elite in the individual sub- population of elite roleWith dynamic rate of moving intoAnd emigrationElite self adaptation is carried out to move
Move, reach the global equilibrium point Mbest of multilayer elite role;
C, the traditional Chinese medical science case history diagnostic rule excavation for realizing being based on the sub- population vector of multilayer elite role, from multilayer elite role
Concentration selects the most strong elite subset optimization vector of n optimization abilityRealize n traditional Chinese medical science case history property set
MPR1,MPR2,...,MPRnGlobal search and local excavation of refining, traditional Chinese medical science case history attribute set MPR is extracted respectively1,
MPR2,...,MPRnEach self-corresponding diagnostic rule excavates collection DRS1,DRS2,...,DRSn;
The global optimum's diagnostic rule collection of D, output traditional Chinese medical science case history under most strong elite optimization
Further improvement of the present invention is:Step A is specially:By traditional Chinese medical science case history attribute be assigned to it is different " it is common-
In elite " role's Advanced group species, the related and attribute pretreatment that interdepends in traditional Chinese medical science case history is carried out, by designing elite role's kind
The elite concentration ECD (Elitist of groupi) and select probability P (xi), ElitistiIt is i-th elite role population, xiIt is i & lt
Traditional Chinese medical science case history selection is carried out, the related and complementary traditional Chinese medical science case history of attribute is distributed to elite role population, will be without selected
Select probability P (xi) the general traditional Chinese medical science case history attribute chosen distributed into conventional character population, it is comprised the following steps that:
A, the elite role concentration ECD (Elitist for calculating each elite populationi) be
Wherein ElitistiIt is i-th elite role population, f (xj) it is j-th adaptation of elite in elite role population
Degree,It is i-th average fitness of elite role population;
B, i-th elite role population of structure select the select probability P (x of traditional Chinese medical science case history attributei) be
C, with select probability P (xi) the related and complementary traditional Chinese medical science case history attribute of attribute in traditional Chinese medical science case history is assigned to essence
English role population ElitistiIn, will be without selected probability P (xi) the general traditional Chinese medical science case history attribute chosen is assigned to conventional character
In population;
D, each traditional Chinese medical science case history property set that will be assigned in elite role populationIt is mapped to [0,1] Advanced group species
Space, is normalized, and mapping equation is as follows:
Wherein f (xj) it is j-th elite fitness in elite role population,It is average for the i-th elite role population
Fitness, fmaxAnd fminIt is the maximum adaptation degree and minimum fitness of the i-th elite role population;
IfThen MPRiIt is 0;
If f (xj) andIt is equal, then MPRiIt is 1.
Further improvement of the present invention is:Step B is specially:Construction it is a kind of for traditional Chinese medical science case history diagnostic rule excavate
Multilayer elite role's dynamic equalization strategy, by elite in the sub- population of different elite rolesWith dynamic rate of moving intoWith move
Extracting rateElite adaptive-migration is carried out, the global equilibrium point Mbest of multilayer elite role is reached, it is comprised the following steps that:
A, evolution space is divided into the n pentagon Pe of area equation1,Pe2,...,Pen, all evolution elite roles
Population is evenly distributed in different pentagons, and each pentagon then is divided into several equal-sized triangle microevolutions
Region;
B, select the elite with adaptive optimal control degree in i-th pentagon evolution region and evolve individual, i.e. elite roleAnd be ranked up them to low according to fitness size, so as to form multilayer elite angle in i-th pentagonal regions
Color collection EriIt is as follows:
Wherein NeltIt is the maximum quantity of multilayer elite role in i-th pentagonal regions;
C, calculate each elite roleIn EriIn dynamic move into rateAnd emigrationComputing formula
It is as follows:
Wherein IiIt is elite role's maximum dynamic move into rate, E in i-th pentagonal regionsiFor in i-th pentagonal regions
The maximum dynamic emigration of elite role, j is multilayer elite role set EriThe serial number of middle elite role's arrangement;
Each elite role in d, multilayer elite role setWill be according to the rate of moving intoAnd emigrationRespective
Dynamic self-adapting migration is carried out in adjacent pentagonal regions, the average adaptation of each pentagon evolution region elite role is calculated
Degree, is then transferred to the elite role population of adjacent pentagonal regions, so as to adjust each dynamic rate of moving intoAnd emigration
E, repeat the above steps, until the elite role being divided in each pentagonal regions no longer migrates, reach many
The overall equilibrium state of layer elite role;
F, build multilayer elite role global equilibrium point Mbest it is as follows:
Adjusted by the above-mentioned dynamic equalization based on multilayer elite role, in each pentagonal regions space PeiMiddle multilayer
Elite role will can find each best rate of moving intoAnd emigrationThen the dynamic of each elite role is carried out
Adaptive-migration, obtains the global equilibrium point of elite role in each pentagon evolution region, so as to pick out more excellent essence
English role population carries out the related and complementary traditional Chinese medical science case history diagnostic rule of attribute and excavates.
Further improvement of the present invention is:Step C is specially:Realize the traditional Chinese medical science based on multilayer elite role population vector
Case history diagnostic rule is excavated, and the most strong elite subset optimization vector of n optimization ability is selected from multilayer elite role setRealize n traditional Chinese medical science case history property set MPR1,MPR2,...,MPRnGlobal search and local digging of refining
Pick, extracts each self-corresponding diagnostic rule and excavates collection DRS respectively1,DRS2,...,DRSn, it is comprised the following steps that:
A, will be in multilayer elite role overall situation equilibrium point Mbest each elite role set EriMiddle elite role according to
Its optimization ability carries out descending sort from high to low, is followed successively by
B, select from multilayer elite role set the most strong elite subset optimization vector of n optimization abilityIt is sequentially allocated n traditional Chinese medical science case history attribute set MPR1,MPR2,...,MPRnIn, for diagnostic rule
Global search is excavated, and it is expressed as:
C, by the most strong elite subset vector of optimization abilityRespectively according to traditional Chinese medical science case history attribute set
Dimension be launched into n × n dimension arrays, form most strong elite optimization array;
D, will contain n dimension most strong elite optimization array carry out traditional Chinese medical science case history attribute set MPR successively1,MPR2,...,MPRn
The part of middle diagnostic rule is refined excavation, extracts its corresponding tcm diagnosis rule set for DRS1,DRS2,...,DRSn;
E, judge that above-mentioned elite subset vector excavates traditional Chinese medical science case history diagnostic rule collection DRSiPrecision whether meet requirement,
As met, then traditional Chinese medical science case history global optimum diagnostic rule collection is exported;Such as it is unsatisfactory for, then repeats local excavation of refining, until
Rule digging precision meets requirement.
The present invention has the following advantages that compared with prior art:
1st, the balance of multilayer elite role:
This method build for traditional Chinese medical science case history diagnostic rule excavate multilayer elite role when, using multilayer elite role
Dynamic equalization strategy, by elite role in pentagon evolution space in multilayer elite role set EriInside carry out dynamic self-adapting
Move into and move out, reach the global equilibrium point Mbest of multilayer elite role, the multilayer elite role for so being constructed is more beneficial for
The related and complementary traditional Chinese medical science case history of processing attribute, reaches preferable diagnostic rule mining effect.
2nd, global search and local excavation optimization performance of refining:
This method realize based on multilayer elite role's subset vector traditional Chinese medical science case history diagnostic rule excavate optimization process when,
Carry out that global search is excavated, local refining and performs local excavation three phases of refining and optimize excavation repeatedly respectively, so
The uncertainty that related and complementary traditional Chinese medical science case history attribute is brought to actual excavation result can be preferably processed, this is mesh
Tcm diagnosis rule digging method as previous is beyond one's reach actual excavation effect.
3rd, stronger traditional Chinese medical science case history diagnostic rule applicability:
This method can preferably overcome that special Traditional Chinese medical electronic case history attribute is fuzzy, complicated inline of various dimensions between case history attribute
The problems such as relation, it is ensured that the traditional Chinese medical science case history diagnostic rule collection of final extraction is the optimal characteristics rule set of required target, is effectively carried
Senior middle school cures the desease and goes through the digging efficiency of diagnostic rule.Practical clinical shows:The tcm diagnosis excavated using the method are advised
Preferable consistent effect is then had according to the diagnostic rule that diagnostic experiences are inferred to actual clinical doctor, the method has stronger
Objectivity and science, traditional Chinese medical science case history computer intelligent auxiliary diagnosis and tcm clinical practice decision-assisting analysis field will with compared with
Strong application value.
Brief description of the drawings
Fig. 1 is overall construction drawing of the invention;
Fig. 2 is multilayer elite role's dynamic equalization strategy schematic diagram;
Fig. 3 is that the traditional Chinese medical science case history diagnostic rule based on multilayer elite role's subset vector excavates schematic diagram;
Fig. 4 is most strong elite optimization array.
Specific embodiment
In order to deepen the understanding of the present invention, below in conjunction with embodiment and accompanying drawing, the invention will be further described, should
Embodiment is only used for explaining the present invention, is not intended to limit the scope of the present invention..
A kind of to cooperate with equalization methods for the multilayer elite role method that traditional Chinese medical science case history diagnostic rule is excavated, specific steps are such as
Under:
A, traditional Chinese medical science case history attribute is assigned in different " common-elite " role's Advanced group species, in carrying out traditional Chinese medical science case history
The related and attribute pretreatment that interdepends, by the elite concentration ECD (Elitist for designing elite role populationi) and selection is generally
Rate P (xi), ElitistiIt is i-th elite role population, xiIt is for i & lt carries out traditional Chinese medical science case history selection, attribute is related and mutual
The traditional Chinese medical science case history of dependence is distributed to elite role population, will be without selected probability P (xi) the general traditional Chinese medical science case history attribute chosen
In distribution to conventional character population;
A kind of multilayer elite role's dynamic equalization strategy excavated for traditional Chinese medical science case history diagnostic rule of B, construction, by i-th
J-th elite in the individual sub- population of elite roleWith dynamic rate of moving intoAnd emigrationElite self adaptation is carried out to move
Move, reach the global equilibrium point Mbest of multilayer elite role;
C, the traditional Chinese medical science case history diagnostic rule excavation for realizing being based on the sub- population vector of multilayer elite role, from multilayer elite role
Concentration selects the most strong elite subset optimization vector of n optimization abilityRealize n traditional Chinese medical science case history property set
MPR1,MPR2,...,MPRnGlobal search and local excavation of refining, traditional Chinese medical science case history attribute set MPR is extracted respectively1,
MPR2,...,MPRnEach self-corresponding diagnostic rule excavates collection DRS1,DRS2,...,DRSn;
The global optimum's diagnostic rule collection of D, output traditional Chinese medical science case history under most strong elite optimization
Step A is specially:Traditional Chinese medical science case history attribute is assigned in different " common-elite " role's Advanced group species, is carried out
The related and attribute pretreatment that interdepends in traditional Chinese medical science case history, by the elite concentration ECD for designing elite role population
(Elitisti) and select probability P (xi), ElitistiIt is i-th elite role population, xiFor i & lt carries out traditional Chinese medical science case history choosing
Select, the related and complementary traditional Chinese medical science case history of attribute is distributed to elite role population, will be without selected probability P (xi) choose
General traditional Chinese medical science case history attribute distribute into conventional character population, it is comprised the following steps that:
A, the elite role concentration ECD (Elitist for calculating each elite populationi) be
Wherein ElitistiIt is i-th elite role population, f (xj) it is j-th adaptation of elite in elite role population
Degree,It is i-th average fitness of elite role population;
B, i-th elite role population of structure select the select probability P (x of traditional Chinese medical science case history attributei) be
C, with select probability P (xi) the related and complementary traditional Chinese medical science case history attribute of attribute in traditional Chinese medical science case history is assigned to essence
English role population ElitistiIn, will be without selected probability P (xi) the general traditional Chinese medical science case history attribute chosen is assigned to conventional character
In population;
D, each traditional Chinese medical science case history property set that will be assigned in elite role populationIt is mapped to [0,1] Advanced group species
Space, is normalized, and mapping equation is as follows:
Wherein f (xj) it is j-th elite fitness in elite role population,It is average for the i-th elite role population
Fitness, fmaxAnd fminIt is the maximum adaptation degree and minimum fitness of the i-th elite role population;
IfThen MPRiIt is 0;
If f (xj) andIt is equal, then MPRiIt is 1.
Step B is specially:A kind of multilayer elite role's dynamic equalization plan excavated for traditional Chinese medical science case history diagnostic rule of construction
Slightly, by elite in the sub- population of different elite rolesWith dynamic rate of moving intoAnd emigrationCarry out elite adaptive
Should migrate, reach the global equilibrium point Mbest of multilayer elite role, it is comprised the following steps that:
A, evolution space is divided into the n pentagon Pe of area equation1,Pe2,...,Pen, all evolution elite roles
Population is evenly distributed in different pentagons, and each pentagon then is divided into several equal-sized triangle microevolutions
Region;
B, select the elite with adaptive optimal control degree in i-th pentagon evolution region and evolve individual, i.e. elite roleAnd be ranked up them to low according to fitness size, so as to form multilayer elite angle in i-th pentagonal regions
Color collection EriIt is as follows:
Wherein NeltIt is the maximum quantity of multilayer elite role in i-th pentagonal regions;
C, calculate each elite roleIn EriIn dynamic move into rateAnd emigrationComputing formula
It is as follows:
Wherein IiIt is elite role's maximum dynamic move into rate, E in i-th pentagonal regionsiFor in i-th pentagonal regions
The maximum dynamic emigration of elite role, j is multilayer elite role set EriThe serial number of middle elite role's arrangement;
Each elite role in d, multilayer elite role setWill be according to the rate of moving intoAnd emigrationRespective
Dynamic self-adapting migration is carried out in adjacent pentagonal regions, the average adaptation of each pentagon evolution region elite role is calculated
Degree, is then transferred to the elite role population of adjacent pentagonal regions, so as to adjust each dynamic rate of moving intoAnd emigration
E, repeat the above steps, until the elite role being divided in each pentagonal regions no longer migrates, reach many
The overall equilibrium state of layer elite role;
F, build multilayer elite role global equilibrium point Mbest it is as follows:
Adjusted by the above-mentioned dynamic equalization based on multilayer elite role, in each pentagonal regions space PeiMiddle multilayer
Elite role will can find each best rate of moving intoAnd emigrationThen the dynamic of each elite role is carried out
Adaptive-migration, obtains the global equilibrium point of elite role in each pentagon evolution region, so as to pick out more excellent essence
English role population carries out the related and complementary traditional Chinese medical science case history diagnostic rule of attribute and excavates.
Step C is specially:Realize that the traditional Chinese medical science case history diagnostic rule based on multilayer elite role population vector is excavated, from multilayer
The most strong elite subset optimization vector of n optimization ability is selected in elite role setCured the desease in realizing n
Go through property set MPR1,MPR2,...,MPRnGlobal search and local excavations of refining, each self-corresponding diagnosis is extracted respectively and is advised
Then excavate collection DRS1,DRS2,...,DRSn, it is comprised the following steps that:
A, will be in multilayer elite role overall situation equilibrium point Mbest each elite role set EriMiddle elite role according to
Its optimization ability carries out descending sort from high to low, is followed successively by
B, select from multilayer elite role set the most strong elite subset optimization vector of n optimization ability
It is sequentially allocated n traditional Chinese medical science case history attribute set MPR1,MPR2,...,MPRnIn, excavated for diagnostic rule global search, its
It is expressed as:
C, by the most strong elite subset vector of optimization abilityRespectively according to traditional Chinese medical science case history attribute set
Dimension be launched into n × n dimension arrays, form most strong elite optimization array;
D, will contain n dimension most strong elite optimization array carry out traditional Chinese medical science case history attribute set MPR successively1,MPR2,...,MPRn
The part of middle diagnostic rule is refined excavation, extracts its corresponding tcm diagnosis rule set for DRS1,DRS2,...,DRSn;
E, judge that above-mentioned elite subset vector excavates traditional Chinese medical science case history diagnostic rule collection DRSiPrecision whether meet requirement,
As met, then traditional Chinese medical science case history global optimum diagnostic rule collection is exported;Such as it is unsatisfactory for, then repeats local excavation of refining, until
Rule digging precision meets requirement.
The present invention discloses a kind of multilayer elite role method excavated for traditional Chinese medical science case history diagnostic rule.The method sets first
Traditional Chinese medical science case history attribute is assigned to different classes of " common-elite " role's kind by the meter population concentration select probability of elite role
In group, the related and attribute pretreatment that interdepends in traditional Chinese medical science case history is carried out;Then construct a kind of dynamic based on multilayer elite role
State balance policy, moves into and moves out to form multilayer elite by different elite role's population dynamics in each pentagon evolution region
Role's overall situation equilibrium point;The last essence selected from multilayer elite role with global search and local most strong optimization ability of refining
English subset vector, constructs most strong elite optimization array and is quickly excavated with realizing traditional Chinese medical science case history diagnostic rule.
The present invention can preferably process the feature that there is complicated correlation and interdependency between traditional Chinese medical science case history attribute, gram
Take diagnostic rule and be difficult to the problems such as excavating, effectively improve traditional Chinese medical science case history diagnostic rule digging efficiency, with stronger robustness and
Practicality.The diagnosis that the tcm diagnosis rule excavated using the method is inferred to actual clinical doctor according to diagnostic experiences
Rule has preferably consistent effect, and the method has stronger objectivity and science, in traditional Chinese medical science case history Computer-aided Intelligent
Diagnosis and tcm clinical practice decision-assisting analysis field will have stronger application value.