CN109062961A - A kind of expert's combination recommended method of knowledge based map - Google Patents
A kind of expert's combination recommended method of knowledge based map Download PDFInfo
- Publication number
- CN109062961A CN109062961A CN201810677224.4A CN201810677224A CN109062961A CN 109062961 A CN109062961 A CN 109062961A CN 201810677224 A CN201810677224 A CN 201810677224A CN 109062961 A CN109062961 A CN 109062961A
- Authority
- CN
- China
- Prior art keywords
- expert
- val
- project
- sim
- combination
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of experts of knowledge based map to combine recommended method, and the present invention is mainly based upon the domain knowledge map having had been built up, combines similar thought and association mining.History specialist examination data are handled, probabilistic logic data, which are obtained, using FP-Growth first formulates knowledge mapping, differences are excluded by similarity algorithm again, extract item close with recommendation, then it extracts expert and is combined according to rule, a kind of excellent method for assessing combination by frequent episode and confidence level is devised according to relationship knowledge.The method of the present invention effectively recommends a kind of careful expert's combination of the highest figure of scoring, so that expert, which cooperates with, examines that efficiency improves, and increases the use value that history item examination expert collects data.
Description
Technical field
The invention belongs to proposed algorithm and knowledge mapping technical field, in particular to a kind of expert group of knowledge based map
Close recommended method.
Background technique
Expert in the present invention combines proposed algorithm and recommends have important role and meaning to the expert group of facing project.
When needing to recommend a combinatorial problem in recommender system, researchers can select to be based on some group clusterings, phase
Like degree, classification analysis etc., or by user's rating matrix, the technologies such as semantic analysis incorporate in group's recommender system.How to send out
The potential relationship between group and item association, group individual is dug, to recommend the relationship for providing value to know for group
Know, optimizes group's proposed algorithm as related system and efficient knowledge services and personalized work allocation plan are provided.
The existing Research foundation of Feng Wanli, Zhu Quanyin et al. includes: Wanli Feng.Research of theme
statement extraction for chinese literature based on lexical
chain.International Journal of Multimedia and Ubiquitous Engineering,Vol.11,
No.6(2016),pp.379-388;Wanli Feng,Ying Li,Shangbing Gao,Yunyang Yan,Jianxun
Xue.A novel flame edge detection algorithm via a novel active contour
model.International Journal of Hybrid Information Technology,Vol.9,No.9
(2016),pp.275-282;Method for mode matching [J] microelectronics of Liu Jinling, the Feng Wanli based on Feature Dependence relationship with
Computer, 2011,28 (12): 167-170;Liu Jinling, Feng Wanli, Zhang Yahong initialize cluster class center and reconstruct scaling function
Text cluster [J] computer application research, 2011,28 (11): 4115-4117;Liu Jinling, Feng Wanli, Zhang Yahong are based on
Again Chinese short message Text Clustering Method [J] the computer engineering of scale and application, 2012,48 (21): 146-150.;Zhu Quan
Silver, Pan Lu, Liu Wenru wait .Web science and technology news classification extraction algorithm [J] Huaiyingong College journal, 2015,24 (5): 18-24;
Collaborative filtering recommending [J] the computer science and explore, 2014,8 that Li Xiang, Zhu Quan silver joint cluster and rating matrix are shared
(6):751-759;Quanyin Zhu,Sunqun Cao.A Novel Classifier-independent Feature
Selection Algorithm for Imbalanced Datasets.2009,p:77-82;Quanyin Zhu,Yunyang
Yan,Jin Ding,Jin Qian.The Case Study for Price Extracting of Mobile Phone
Sell Online.2011,p:282-285;Quanyin Zhu,Suqun Cao,Pei Zhou,Yunyang Yan,Hong
Zhou.Integrated Price Forecast based on Dichotomy Backfilling and Disturbance
Factor Algorithm.International Review on Computers and Software,2011,Vol.6
(6):1089-1093;Zhu Quanyin, Feng Wanli et al. application, openly with the related patents of authorization: Feng Wanli, Shao Heshuai, Zhuan Jun
A kind of intelligent refrigerated car state monitoring wireless network terminal installation: CN203616634U [P] .2014;Zhu Quanyin, Hu Rongjing, what
Su Qun, a kind of price forecasting of commodity method Chinese patent based on linear interpolation Yu Adaptive windowing mouth of such as week training: ZL
2011 1 0423015.5,2015.07.01;Zhu Quanyin, Cao Suqun, Yan Yunyang, Hu Rong wait quietly, and one kind is repaired based on two divided datas
Mend the price forecasting of commodity method Chinese patent with disturbing factors: 2,011 1 0422274.6,2013.01.02 of ZL;Li Xiang,
Zhu Quanyin, Hu Ronglin, a kind of all deep Cold Chain Logistics prestowage intelligent recommendation method China Patent Publication No. based on spectral clustering of:
CN105654267A,2016.06.08。
Domain knowledge map:
Domain knowledge map is constructed by various methods, obtains pertinent arts information, these domain knowledges
Comprising some common sense knowledges, potential entity relationship knowledge etc., the present invention mainly passes through association mining algorithm and extracts potentially
Relationship between expert assesses recommendation group by obtained frequent episode knowledge and confidence level knowledge.
Chebyshev's Distance conformability degree algorithm:
Similarity calculation is carried out to by project to be processed and history item, different similarity calculating methods can obtain different
Effect, and Chebyshev's distance does not need the processing that item attribute characteristic is normalized, i.e., without the concern for each
Dimension interacts between dimensional feature, treat processing item used here as Chebyshev distance algorithm and history item carry out it is another
A kind of calculating of similarity.Chebyshev's Distance conformability degree algorithm requires data characteristics dimension identical, and dimension is preferably greater than
Three, Chebyshev's Distance conformability degree algorithm just has its meaning, and corresponding exactly to integrated project characteristic dimension is five dimensions, meets the phase
Like the application requirement of degree algorithm.
Association mining algorithm:
Present invention research is imitated with for classical Apriori algorithm in the time using FP-Growth algorithm, the algorithm
The effectiveness embodied in rate is more excellent, and main cause is by way of generating candidate frequent item set, i.e., to change algorithm not be
Construct FP-tree method, and FP-Growth algorithm for different length rule can embody good adaptability and
High efficiency.
Frequent item set and confidence level collection are obtained by FP-Growth association mining algorithm, support of such as giving a definition and confidence
Degree, support indicate the frequency number that individual event subset occurs in entire set, and enabling in entire set includes n data item.
And confidence level is then expressed as the frequency values that regular X- > Y occurs in entire set entry, i.e., occurs in condition X
In the case of, the probability proportion of Y appearance.
And in practical application, confidence threshold value can be set to FP-Growth algorithm, be defined as if being more than confidence threshold value
Effective rule, on the contrary what is excavated is then invalid rule.
When recommending problem towards group, existing paper is based primarily upon the letter of the relationship between existing user and relative article
Secondly breath mainly considers the classification etc. for recommending article, all cross group's recommendation that Cluster Classification etc. carries out ware, also there is correlation
Patent proposition is recommended by association mining, but does not combine sufficiently and excavate the frequent episode obtained and confidence item information and be subject to
It is dissolved into recommendation.
The expert group proposed algorithm of traditional facing project have be by artificial selection come out, some is only according to item
Mesh is recommended with recommended, and artificial recommended method is due to consuming human resources, and there are certain one-sidedness for recommendation, and
The relationship that is unable between analysis project and recommended is simultaneously recommended, and causes the effect recommended inconsiderable, therefore be based on
The proposed algorithm of project and expert's relationship replaces, these methods can not make full use of the relationship between expert, is not achieved specially
The accuracy that group of family is recommended, so that the expert group recommended, which merges, is not able to satisfy corresponding recommended requirements.
Summary of the invention
Goal of the invention: in view of the above problems, the present invention provides a kind of special by the examination of comprehensive analysis history item
Family's collection and history integrated project record set are associated the processing building domain knowledge map of the methods of excavation, are unexamined project
The expert of the knowledge based map for the examination expert group for recommending customized scoring high combines recommended method.
Technical solution: the present invention proposes that a kind of expert of knowledge based map combines recommended method, includes the following steps:
(1) excavation is associated to participation integrated project expert combined data set, establishes expert's frequent episode set and expert
Between confidence level set;
(2) similarity calculation is carried out according to integrated project multidimensional characteristic, obtains with unexamined item similarity highest 3
Integrated project extracts the expert for participating in the integrated project chosen, and carries out permutation and combination according to the subset direction of expert's research, obtains
Alternative expert's combination of sets;
(3) frequent item set and confidence level collection are introduced, is scored alternative expert combination, by the expert's combination of first three of scoring
Recommend unexamined project.
The present invention mainly passes through association mining FP-Growth algorithm etc. and examines knowledge graph to history data set processing structure figures
Spectrum scores to expert's combination to be recommended by the knowledge that has been built up out, by the high combined application of score in recommendation, into
And expert is cooperateed with and examines that efficiency improves.
Further, in the step (1), the specific steps of confidence level set between expert's frequent episode set and expert are established
It is as follows:
(1.1) defining GP is that integrated project expert participates in collection, g1,g2,g3,g4,g5,g6It is the second level item of GP project respectively
Mesh, each secondary items are by a specialist examination;
(1.2) it defines expert and collects Person={ p1,p2,...,pN, only one examination direction of each expert, i.e. piIt examines
Looking into direction is gj, project examination expert's collection GPh={ pi1,pi2,pi3,pi4,pi5,pi6, GP={ GP1,GP2,...,GPH, wherein
I, i1, i2, i3, i4, i5, i6 ∈ [1, N], j ∈ [1,6], N > 6, h ∈ [1, H], N=Card (Person), H=Card (GP);
(1.3) is obtained by all Tu Shen expert groups and closes frequent item set for GP processing using association rules method FP-Growth
Relationt and figure examine expert and combine confidence level collection Confidence, Relationt={ { rel1:fr1},{rel2:
fr2},...,{relM:frM, Confidence={ { con1->conf1:nu1},
{con2->conf2:nu2},...,{conM->confM:nuK, wherein relX1={ r1,r2,...,rj1, conX2
={ s1,s2,...,sj2, confX2={ t1,t2,...,tj3, rj1,sj2,tj3∈ Person, 1≤j1, j2, j3≤N, M=
Card (Relationt), K=Card (Confidence), X1 ∈ [1, M], X2, X2 ∈ [1, K], frx1Indicate relX1Frequency
Number, nuX2Indicate conX2->confX2Confidence level.
Further, alternative expert's combination of sets is obtained in the step (2) specific step is as follows:
(2.1) defining GT is integrated project feature set, GTg={ tg1,tg2,tg3,tg4,tg5, wherein tg1,tg2,tg3,
tg4,tg5Respectively GTgOccupied area, number of floor levels, building height, accounts receivable and the minimum bill of project, wherein g ∈ [1,
H];
(2.2) defining G is project to be processed, G={ gt1, gt2, gt3, gt4, gt5 }, wherein gt1, gt2, gt3, gt4,
Gt5 is respectively occupied area, number of floor levels, building height, accounts receivable and the minimum bill of G project;
(2.3) cyclic variable m is defined, for traversing GT, simmIndicate integrated project G and integrated project GtmSimilarity,
Sim is similarity collection, wherein it is 1, id that m ∈ [1, H], m, which assign initial value,mIndicate GtmProject number, Sim assign initial value be 0;
(2.4) step (5) are jumped to if m≤H, otherwise jump to step (2.7);
(2.5)Wherein, simmIndicate integrated project G and integrated project GtmPassing through
The value that Chebyshev's Distance conformability degree algorithm is calculated, Sim=Sim ∪ { idm,simm};
(2.6) m=m+1;
(2.7) Sim={ { id is obtained1,sim1},{id2,sim2},...,{idH,simH}};
(2.8) orderly similarity collection Simi={ { id is obtainedf1,af1},{idf2,af2},...,{idfH,afH, wherein af1
≥af2≥...≥afH, { idft,aft∈ Sim, ft, f1, f2, fH ∈ [1, H], SimProject={ { idf1,af1},{idf2,
af2},{idf3,af3}};
(2.9) defining Forecast is that pre-selection is schemed to examine expert's collection, Forecast={ pm1,pm2,...,pmn, wherein pmi∈
(GPf1∪GPf2∪GPf3);
(2.10) C is definedg1,Cg2,Cg3,Cg4,Cg5,Cg6It respectively indicates participation and examines secondary items g1, g2, g3, g4, g5,
The expert of g6 collects, and is assigned to C according to specialist examination direction in Forecastg1,Cg2,Cg3,Cg4,Cg5,Cg6In;
(2.11) defining ExportCom is that all alternative figures for examining G examine expert's combination of sets, and Com is one of alternative
Examine that the figure of HP examines expert's combination of sets, ExportCom={ Com1,Com2,…,Comnumber, whereinComx={ px1,px2,px3,px4,px5,px6, pxd∈Cgd, d ∈ [1,6], x ∈ [1,
number]。
Further, frequent item set and confidence level collection are introduced in the step (3), are scored alternative expert combination,
To score first three expert's combination recommends unexamined project specific step is as follows:
(3.1) definition is schemed to examine expert's combined evaluation collection Valuation={ val1,val2,…,valnumber, wherein valx
Indicate ComxAssessed value, valx=0, x ∈ [1, number];
(3.2) q is defined, for traversing expert combination of sets ExportCom, and q initial value is equal to 1;
(3.3) step (3.4) are jumped to if q≤number, otherwise jump to step (3.18);
(3.4) r is defined, for traversing Relationt, it is 1 that r, which assigns initial value,;
(3.5) q=q+1;
(3.6) step (3.7) are jumped to if r≤M, otherwise jump to step (3.10);
(3.7) ifStep (3.8) are then jumped to, step (3.9) are otherwise jumped to;
(3.8)valr=valr+frr×Card(relr), jump procedure (3.9);
(3.9) r=r+1;
(3.10) v is defined, for traversing Confidence, it is 1 that v, which assigns initial value,;
(3.11 jump to step (3.12) if v≤K, otherwise jump to step (3.17);
(3.12) ifStep (3.14) are then jumped to, step (3.15) are otherwise jumped to;
(3.13)
(3.14)valr=valr×nuv, jump the step of expert combines assessment algorithm flow chart (3.16);
(3.15)valr=valr×(1-nuv), jump the step of expert combines assessment algorithm flow chart (3.16);
(3.16) v=v+1;
(3.17) val is obtainedr;
(3.18) Valuation={ val1,val2,…,valnumber};
(3.19) it chooses assessment and concentrates scoring highest three, valtop1,valtop2,valtop3;
(3.20)Comtop1,Comtop2,Comtop3It is combined for the recommendation of unexamined project G.
The present invention by adopting the above technical scheme, has the advantages that
The method of the present invention examines that expert collects using existing integrated project record set and history item, effectively recommends one
Expert's combination that kind scores high based on customized combined evaluation algorithm, improves and recommends combined accuracy, specific: this hair
Bright that the building for carrying out domain knowledge map is recorded using specialist examination history of project, which examines mainly for existing figure
Expert participates in data set, participates in collection to pretreated expert by association mining and carries out excavating potential expert's relationship, this
The relationship excavated is mainly the relationship in probabilistic logic, and relationship knowledge is applied to the assessment of expert's combination.In addition, this hair
Bright creatively to propose expert's combination assessment algorithm, the combined evaluation algorithm for including in expert's combined recommendation is depended in knowledge
Expert's Logic Relation that data with existing and reasoning obtain in map can be designed a set of by these data and relationship
It is suitable for the assessment models of expert group, combined appropriateness is calculated by the way that assessment algorithm science is accurate, finally by score
Excellent expert's combination of sets is applied in integrated project distribution to be processed, and a central evaluation algorithm optimization in this way pushes away
The accuracy rate recommended reduces the operating time of operation allocator significantly, preferably serves real application systems, improves operation distribution
And running efficiency of system.
Detailed description of the invention
Fig. 1 is overall flow figure of the invention;
Fig. 2 is that structure figures examine the flow chart that system regions knowledge mapping excavates potential expert's relationship knowledge method in Fig. 1;
Fig. 3 is the flow chart of project similarity calculation and expert's combined method in Fig. 1;
Fig. 4 is the flow chart of expert's combination evaluation method in Fig. 1.
Specific embodiment
Combined with specific embodiments below, the present invention is furture elucidated, it should be understood that these embodiments are merely to illustrate the present invention
Rather than limit the scope of the invention, after the present invention has been read, those skilled in the art are to various equivalences of the invention
The modification of form falls within the application range as defined in the appended claims.
As shown in Figs 1-4, a kind of expert of knowledge based map of the present invention combines recommended method, including walks as follows
It is rapid:
Step 1: being associated excavations to integrated project expert combined data set is participated in, establish expert's frequent episode set and special
Confidence level set between family, specific as shown in Figure 2:
Step 1.1: defining GP is that integrated project expert participates in collection, g1,g2,g3,g4,g5,g6It is the second level of GP project respectively
Project, each secondary items are by a specialist examination;
Step 1.2: defining expert and collect Person={ p1,p2,...,pN, only one examination direction of each expert, i.e. pi
Examination direction is gj, project examination expert's collection GPh={ pi1,pi2,pi3,pi4,pi5,pi6, GP={ GP1,GP2,...,GPH,
In, i, i1, i2, i3, i4, i5, i6 ∈ [1, N], j ∈ [1,6], N > 6, h ∈ [1, H], N=Card (Person), H=Card
(GP);
Step 1.3: using association rules method FP-Growth to GP processing, obtaining all Tu Shen expert groups and close frequent episode
Collect Relationt and schemes to examine expert's combination confidence level collection Confidence, Relationt={ { rel1:fr1},{rel2:
fr2},...,{relM:frM, Confidence={ { con1->conf1:nu1},{con2->conf2:nu2},...,{conM-
>confM:nuK, wherein relX1={ r1,r2,...,rj1, conX2={ s1,s2,...,sj2, confX2={ t1,t2,...,
tj3, rj1,sj2,tj3∈ Person, 1≤j1, j2, j3≤N, M=Card (Relationt), K=Card (Confidence),
X1 ∈ [1, M], X2, X2 ∈ [1, K], frx1Indicate relX1Frequency, nuX2Indicate conX2->confX2Confidence level.
Step 2: similarity calculation being carried out according to integrated project multidimensional characteristic, is obtained highest with unexamined item similarity
3 integrated projects extract the expert for participating in the integrated project chosen, and carry out permutation and combination according to the subset direction of expert's research,
Alternative expert's combination of sets is obtained, specific as shown in Figure 3:
Step 2.1: definition GT is integrated project feature set, GTg={ tg1,tg2,tg3,tg4,tg5, wherein tg1,tg2,
tg3,tg4,tg5Respectively GTgOccupied area, number of floor levels, building height, accounts receivable and the minimum bill of project, wherein g ∈
[1,H];
Step 2.2: definition G is project to be processed, G={ gt1, gt2, gt3, gt4, gt5 }, wherein gt1, gt2, gt3,
Gt4, gt5 are respectively occupied area, number of floor levels, building height, accounts receivable and the minimum bill of G project;
Step 2.3: cyclic variable m is defined, for traversing GT, simmIndicate integrated project G and integrated project GtmIt is similar
Degree, Sim are similarity collection, wherein it is 1, id that m ∈ [1, H], m, which assign initial value,mIndicate GtmProject number, Sim assign initial value be 0;
Step 2.4: jumping to step 2.5 if m≤H, otherwise jump to step 2.7;
Step 2.5:Wherein, simmIndicate integrated project G and integrated project GtmLogical
Cross the value that Chebyshev's Distance conformability degree algorithm is calculated, Sim=Sim ∪ { idm,simm};
Step 2.6:m=m+1;
Step 2.7: obtaining Sim={ { id1,sim1},{id2,sim2},...,{idH,simH}};
Step 2.8: obtaining orderly similarity collection Simi={ { idf1,af1},{idf2,af2},...,{idfH,afH,
In, af1≥af2≥...≥afH, { idft,aft∈ Sim, ft, f1, f2, fH ∈ [1, H], SimProject={ { idf1,af1},
{idf2,af2},{idf3,af3}};
Step 2.9: defining Forecast is that pre-selection is schemed to examine expert's collection, Forecast={ pm1,pm2,...,pmn, wherein
pmi∈(GPf1∪GPf2∪GPf3);
Step 2.10: defining Cg1,Cg2,Cg3,Cg4,Cg5,Cg6It respectively indicates participation and examines secondary items g1, g2, g3, g4,
The expert of g5, g6 collect, and are assigned to C according to specialist examination direction in Forecastg1,Cg2,Cg3,Cg4,Cg5,Cg6In;
Step 2.11: defining ExportCom is that all alternative figures for examining G examine expert's combination of sets, and Com is one of standby
Choosing examines that the figure of HP examines expert's combination of sets, ExportCom={ Com1,Com2,…,Comnumber, whereinComx={ px1,px2,px3,px4,px5,px6, pxd∈Cgd, d ∈ [1,6], x ∈ [1, number].
Step 3: introducing frequent item set and confidence level collection, score alternative expert combination, by the expert of first three that scores
Combined recommendation gives unexamined project, specific as shown in Figure 4:
Step 3.1: definition is schemed to examine expert's combined evaluation collection Valuation={ val1,val2,…,valnumber, wherein
valxIndicate ComxAssessed value, valx=0, x ∈ [1, number];
Step 3.2: defining q, for traversing expert combination of sets ExportCom, and q initial value is equal to 1;
Step 3.3: jumping to step 3.4 if q≤number, otherwise jump to step 3.18;
Step 3.4: defining r, for traversing Relationt, it is 1 that r, which assigns initial value,;
Step 3.5:q=q+1;
Step 3.6: jumping to step 3.7 if r≤M, otherwise jump to step 3.10;
Step 3.7: ifStep 3.8 is then jumped to, step 3.9 is otherwise jumped to;
Step 3.8:valr=valr+frr×Card(relr), jump procedure 3.9;
Step 3.9:r=r+1;
Step 3.10: defining v, for traversing Confidence, it is 1 that v, which assigns initial value,;
Step 3.11: jumping to step 3.12 if v≤K, otherwise jump to step 3.17;
Step 3.12: ifStep 3.14 is then jumped to, step 3.15 is otherwise jumped to;
Step 3.13:
Step 3.14:valr=valr×nuv, jump the step 3.16 that expert combines assessment algorithm flow chart;
Step 3.15:valr=valr×(1-nuv), jump the step 3.16 that expert combines assessment algorithm flow chart;
Step 3.16:v=v+1;
Step 3.17: obtaining valr;
Step 3.18:Valuation={ val1,val2,…,valnumber};
Step 3.19: it chooses assessment and concentrates scoring highest three, valtop1,valtop2,valtop3;
Step 3.20:Comtop1,Comtop2,Comtop3It is combined for the recommendation of unexamined project G.
Wherein, FP-Growth method excavates potential expert's relation information, obtains frequent episode knowledge and confidence level knowledge, contract
Than the similarity that snow husband's Distance conformability degree algorithm calculates project and history item to be processed, it is high that expert's combinational algorithm extracts similarity
History item participate in expert and being combined according to rule, expert combines that assessment algorithm introduces frequent episode knowledge and confidence level is known
Know and expert's combination is assessed.
By handling 37136 history specialist examination data, probabilistic logic number is obtained using FP-Growth first
Differences are excluded according to formulation knowledge mapping, then by similarity algorithm, item close with recommendation is extracted, then extracts expert simultaneously
It is combined according to rule, a kind of excellent side assessing combination by frequent episode and confidence level is devised according to relationship knowledge
Method.Improved proposed algorithm further improves the accuracy of prediction while the freshness that ensure that recommendation, from 21
It is selected in expert in the case that wherein 6 experts are as final recommendation expert group, it is finally accurate by the test of test set
It is increased to 70%.And expert's group recommending method proposed by the present invention is generally applicable to expert's combined recommendation problem.
Claims (4)
1. a kind of expert of knowledge based map combines recommended method, which comprises the steps of:
(1) excavation is associated to participation integrated project expert combined data set, establishes between expert's frequent episode set and expert and sets
Reliability set;
(2) similarity calculation is carried out according to integrated project multidimensional characteristic, obtained and highest 3 synthesis of unexamined item similarity
Project extracts the expert for participating in the integrated project chosen, and carries out permutation and combination according to the subset direction of expert's research, obtains alternative
Expert's combination of sets;
(3) frequent item set and confidence level collection are introduced, is scored alternative expert combination, by the expert's combined recommendation of first three that scores
To unexamined project.
2. a kind of expert of knowledge based map according to claim 1 combines recommended method, which is characterized in that the step
Suddenly in (1), establishing confidence level set between expert's frequent episode set and expert, specific step is as follows:
(1.1) defining GP is that integrated project expert participates in collection, g1,g2,g3,g4,g5,g6It is the secondary items of GP project respectively, often
A secondary items are by a specialist examination;
(1.2) it defines expert and collects Person={ p1,p2,...,pN, only one examination direction of each expert, i.e. piExamine direction
For gj, project examination expert's collection GPh={ pi1,pi2,pi3,pi4,pi5,pi6, GP={ GP1,GP2,...,GPH, wherein i, i1,
I2, i3, i4, i5, i6 ∈ [1, N], j ∈ [1,6], N > 6, h ∈ [1, H], N=Card (Person), H=Card (GP);
(1.3) is obtained by all Tu Shen expert groups and closes frequent item set for GP processing using association rules method FP-Growth
Relationt and figure examine expert and combine confidence level collection Confidence, Relationt={ { rel1:fr1},{rel2:
fr2},...,{relM:frM, Confidence={ { con1->conf1:nu1},{con2->conf2:nu2},...,{conM-
>confM:nuK, wherein relX1={ r1,r2,...,rj1, conX2={ s1,s2,...,sj2, confX2={ t1,t2,...,
tj3, rj1,sj2,tj3∈ Person, 1≤j1, j2, j3≤N, M=Card (Relationt), K=Card (Confidence),
X1 ∈ [1, M], X2, X2 ∈ [1, K], frx1Indicate relX1Frequency, nuX2Indicate conX2->confX2Confidence level.
3. a kind of expert of knowledge based map according to claim 1 combines recommended method, which is characterized in that the step
Suddenly frequent item set and confidence level collection are introduced in (2), are scored alternative expert combination, by the expert's combined recommendation of first three that scores
To unexamined project, specific step is as follows:
(2.1) defining GT is integrated project feature set, GTg={ tg1,tg2,tg3,tg4,tg5, wherein tg1,tg2,tg3,tg4,tg5
Respectively GTgOccupied area, number of floor levels, building height, accounts receivable and the minimum bill of project, wherein g ∈ [1, H];
(2.2) defining G is project to be processed, G={ gt1, gt2, gt3, gt4, gt5 }, wherein gt1, gt2, gt3, gt4, gt5
The respectively occupied area of G project, number of floor levels, building height, accounts receivable and minimum bill;
(2.3) cyclic variable m is defined, for traversing GT, simmIndicate integrated project G and integrated project GtmSimilarity, Sim is
Similarity collection, wherein it is 1, id that m ∈ [1, H], m, which assign initial value,mIndicate GtmProject number, Sim assign initial value be 0;
(2.4) step (5) are jumped to if m≤H, otherwise jump to step (2.7);
(2.5)Wherein, simmIndicate integrated project G and integrated project GtmPassing through Qie Bixue
The value that husband's Distance conformability degree algorithm is calculated, Sim=Sim ∪ { idm,simm};
(2.6) m=m+1;
(2.7) Sim={ { id is obtained1,sim1},{id2,sim2},...,{idH,simH}};
(2.8) orderly similarity collection Simi={ { id is obtainedf1,af1},{idf2,af2},...,{idfH,afH, wherein af1≥af2
≥...≥afH, { idft,aft∈ Sim, ft, f1, f2, fH ∈ [1, H], SimProject={ { idf1,af1},{idf2,af2},
{idf3,af3}};
(2.9) defining Forecast is that pre-selection is schemed to examine expert's collection, Forecast={ pm1,pm2,...,pmn, wherein pmi∈(GPf1
∪GPf2∪GPf3);
(2.10) C is definedg1,Cg2,Cg3,Cg4,Cg5,Cg6It respectively indicates participation and examines secondary items g1, g2, g3, g4, g5, g6's
Expert's collection, is assigned to C according to specialist examination direction in Forecastg1,Cg2,Cg3,Cg4,Cg5,Cg6In;
(2.11) defining ExportCom is that all alternative figures for examining G examine expert's combination of sets, and Com, which is that one of which is alternative, to be examined
The figure of HP examines expert's combination of sets, ExportCom={ Com1,Com2,…,Comnumber, whereinComx={ px1,px2,px3,px4,px5,px6, pxd∈Cgd, d ∈ [1,6], x ∈ [1,
number]。
4. a kind of expert of knowledge based map according to claim 1 combines recommended method, which is characterized in that the step
Suddenly alternative expert's combination of sets is obtained in (3), and specific step is as follows:
(3.1) definition is schemed to examine expert's combined evaluation collection Valuation={ val1,val2,…,valnumber, wherein valxIt indicates
ComxAssessed value, valx=0, x ∈ [1, number];
(3.2) q is defined, for traversing expert combination of sets ExportCom, and q initial value is equal to 1;
(3.3) step (3.4) are jumped to if q≤number, otherwise jump to step (3.18);
(3.4) r is defined, for traversing Relationt, it is 1 that r, which assigns initial value,;
(3.5) q=q+1;
(3.6) step (3.7) are jumped to if r≤M, otherwise jump to step (3.10);
(3.7) ifStep (3.8) are then jumped to, step (3.9) are otherwise jumped to;
(3.8)valr=valr+frr×Card(relr), jump procedure (3.9);
(3.9) r=r+1;
(3.10) v is defined, for traversing Confidence, it is 1 that v, which assigns initial value,;
(3.11 jump to step (3.12) if v≤K, otherwise jump to step (3.17);
(3.12) ifStep (3.14) are then jumped to, step (3.15) are otherwise jumped to;
(3.13)
(3.14)valr=valr×nuv, jump the step of expert combines assessment algorithm flow chart (3.16);
(3.15)valr=valr×(1-nuv), jump the step of expert combines assessment algorithm flow chart (3.16);
(3.16) v=v+1;
(3.17) val is obtainedr;
(3.18) Valuation={ val1,val2,…,valnumber};
(3.19) it chooses assessment and concentrates scoring highest three, valtop1,valtop2,valtop3;
(3.20)Comtop1,Comtop2,Comtop3It is combined for the recommendation of unexamined project G.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810677224.4A CN109062961A (en) | 2018-06-27 | 2018-06-27 | A kind of expert's combination recommended method of knowledge based map |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810677224.4A CN109062961A (en) | 2018-06-27 | 2018-06-27 | A kind of expert's combination recommended method of knowledge based map |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109062961A true CN109062961A (en) | 2018-12-21 |
Family
ID=64821296
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810677224.4A Pending CN109062961A (en) | 2018-06-27 | 2018-06-27 | A kind of expert's combination recommended method of knowledge based map |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109062961A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110162638A (en) * | 2019-04-12 | 2019-08-23 | 淮阴工学院 | A kind of expert's combination proposed algorithm based on figure vector |
CN110442618A (en) * | 2019-07-25 | 2019-11-12 | 昆明理工大学 | Merge convolutional neural networks evaluation expert's recommended method of expert info incidence relation |
CN111159428A (en) * | 2019-12-30 | 2020-05-15 | 智慧神州(北京)科技有限公司 | Method and device for automatically extracting event relation of knowledge graph in economic field |
CN112100370A (en) * | 2020-08-10 | 2020-12-18 | 淮阴工学院 | Picture examination expert combined recommendation method based on text convolution and similarity algorithm |
CN112100394A (en) * | 2020-08-10 | 2020-12-18 | 淮阴工学院 | Knowledge graph construction method for recommending medical experts |
CN112100323A (en) * | 2020-08-18 | 2020-12-18 | 淮阴工学院 | Hidden association mining method based on representation learning |
CN112100395A (en) * | 2020-08-11 | 2020-12-18 | 淮阴工学院 | Expert cooperation feasibility analysis method |
CN113254642A (en) * | 2021-05-28 | 2021-08-13 | 华斌 | E-government affair project evaluation expert group recommendation method based on multi-dimensional feature balance |
CN110245259B (en) * | 2019-05-21 | 2021-09-21 | 北京百度网讯科技有限公司 | Video labeling method and device based on knowledge graph and computer readable medium |
CN113806558A (en) * | 2021-09-22 | 2021-12-17 | 湖北天天数链技术有限公司 | Question selection method, knowledge graph construction method and device and electronic equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102479202A (en) * | 2010-11-26 | 2012-05-30 | 卓望数码技术(深圳)有限公司 | Recommendation system based on domain expert |
CN106897370A (en) * | 2017-01-18 | 2017-06-27 | 淮阴工学院 | A kind of figure based on Pearson came similarity and FP Growth examines expert recommendation method |
CN107423399A (en) * | 2017-07-26 | 2017-12-01 | 北京师范大学珠海分校 | A kind of science research programs information semantic of knowledge based collection of illustrative plates reasoning recommends method |
-
2018
- 2018-06-27 CN CN201810677224.4A patent/CN109062961A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102479202A (en) * | 2010-11-26 | 2012-05-30 | 卓望数码技术(深圳)有限公司 | Recommendation system based on domain expert |
CN106897370A (en) * | 2017-01-18 | 2017-06-27 | 淮阴工学院 | A kind of figure based on Pearson came similarity and FP Growth examines expert recommendation method |
CN107423399A (en) * | 2017-07-26 | 2017-12-01 | 北京师范大学珠海分校 | A kind of science research programs information semantic of knowledge based collection of illustrative plates reasoning recommends method |
Non-Patent Citations (2)
Title |
---|
朱全银 等: "Web科技新闻分类抽取算法", 《淮阴工学院学报》 * |
雷雪 等: "关联规则在领域知识推荐中的应用研究", 《情报理论与实践》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110162638B (en) * | 2019-04-12 | 2023-06-20 | 淮阴工学院 | Expert combination recommendation method based on graph vectors |
CN110162638A (en) * | 2019-04-12 | 2019-08-23 | 淮阴工学院 | A kind of expert's combination proposed algorithm based on figure vector |
CN110245259B (en) * | 2019-05-21 | 2021-09-21 | 北京百度网讯科技有限公司 | Video labeling method and device based on knowledge graph and computer readable medium |
CN110442618A (en) * | 2019-07-25 | 2019-11-12 | 昆明理工大学 | Merge convolutional neural networks evaluation expert's recommended method of expert info incidence relation |
CN110442618B (en) * | 2019-07-25 | 2023-04-18 | 昆明理工大学 | Convolutional neural network review expert recommendation method fusing expert information association relation |
CN111159428A (en) * | 2019-12-30 | 2020-05-15 | 智慧神州(北京)科技有限公司 | Method and device for automatically extracting event relation of knowledge graph in economic field |
CN112100394A (en) * | 2020-08-10 | 2020-12-18 | 淮阴工学院 | Knowledge graph construction method for recommending medical experts |
CN112100370A (en) * | 2020-08-10 | 2020-12-18 | 淮阴工学院 | Picture examination expert combined recommendation method based on text convolution and similarity algorithm |
CN112100394B (en) * | 2020-08-10 | 2023-07-21 | 淮阴工学院 | Knowledge graph construction method for recommending medical expert |
CN112100370B (en) * | 2020-08-10 | 2023-07-25 | 淮阴工学院 | Picture-trial expert combination recommendation method based on text volume and similarity algorithm |
CN112100395A (en) * | 2020-08-11 | 2020-12-18 | 淮阴工学院 | Expert cooperation feasibility analysis method |
CN112100395B (en) * | 2020-08-11 | 2024-03-29 | 淮阴工学院 | Expert cooperation feasibility analysis method |
CN112100323A (en) * | 2020-08-18 | 2020-12-18 | 淮阴工学院 | Hidden association mining method based on representation learning |
CN112100323B (en) * | 2020-08-18 | 2023-11-03 | 淮阴工学院 | Hidden association mining method based on representation learning |
CN113254642A (en) * | 2021-05-28 | 2021-08-13 | 华斌 | E-government affair project evaluation expert group recommendation method based on multi-dimensional feature balance |
CN113806558A (en) * | 2021-09-22 | 2021-12-17 | 湖北天天数链技术有限公司 | Question selection method, knowledge graph construction method and device and electronic equipment |
CN113806558B (en) * | 2021-09-22 | 2024-03-26 | 湖北天天数链技术有限公司 | Question selection method, knowledge graph construction device and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109062961A (en) | A kind of expert's combination recommended method of knowledge based map | |
CN109299811B (en) | Complex network-based fraud group recognition and risk propagation prediction method | |
Kuo et al. | Integration of ART2 neural network and genetic K-means algorithm for analyzing Web browsing paths in electronic commerce | |
Yen et al. | Cluster-based under-sampling approaches for imbalanced data distributions | |
Yu et al. | Classification method for detecting coercive self-citation in journals | |
CN106897370B (en) | Picture examination expert recommendation method based on Pearson similarity and FP-Growth | |
CN111292008A (en) | Privacy protection data release risk assessment method based on knowledge graph | |
CN102572744B (en) | Recognition feature library acquisition method and device as well as short message identification method and device | |
CN111324816B (en) | Interest point recommendation method based on region division and context influence | |
WO2020135642A1 (en) | Model training method and apparatus employing generative adversarial network | |
CN110399916A (en) | A kind of cutaneum carcinoma image classification method based on image enhancement and Inception network | |
CN110162638A (en) | A kind of expert's combination proposed algorithm based on figure vector | |
Wang et al. | Improving graph-based label propagation algorithm with group partition for fraud detection | |
CN102722578A (en) | Unsupervised cluster characteristic selection method based on Laplace regularization | |
CN114782161A (en) | Method, device, storage medium and electronic device for identifying risky users | |
Bar-Yossef et al. | Cluster ranking with an application to mining mailbox networks | |
CN112837078B (en) | Method for detecting abnormal behavior of user based on clusters | |
Vo et al. | Parallel method for mining high utility itemsets from vertically partitioned distributed databases | |
WO2024001102A1 (en) | Method and apparatus for intelligently identifying family circle in communication industry, and device | |
Sharda et al. | Do attitudes affect behavioral choices or vice-versa: Uncovering latent segments within a heterogeneous population | |
Yang et al. | Fuzzy intertransaction class association rule mining using genetic network programming for stock market prediction | |
CN112150285A (en) | Abnormal financial organization hierarchy dividing system based on neighborhood topological structure and working method thereof | |
Lai et al. | An effective pruning scheme for top-k dominating query processing on uncertain data streams | |
CN116484026B (en) | Knowledge graph-based person relationship reasoning method and device and electronic equipment | |
Bielecka et al. | Taxonomy of real estate properties with the use of k-means method |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181221 |
|
RJ01 | Rejection of invention patent application after publication |