CN108573264A - A kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm - Google Patents

A kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm Download PDF

Info

Publication number
CN108573264A
CN108573264A CN201710130878.0A CN201710130878A CN108573264A CN 108573264 A CN108573264 A CN 108573264A CN 201710130878 A CN201710130878 A CN 201710130878A CN 108573264 A CN108573264 A CN 108573264A
Authority
CN
China
Prior art keywords
bee
artificial
honeybee
cluster
fitness
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
Application number
CN201710130878.0A
Other languages
Chinese (zh)
Other versions
CN108573264B (en
Inventor
朱云龙
吕赐兴
张�浩
张丁
张丁一
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenyang Institute of Automation of CAS
Original Assignee
Shenyang Institute of Automation of CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenyang Institute of Automation of CAS filed Critical Shenyang Institute of Automation of CAS
Priority to CN201710130878.0A priority Critical patent/CN108573264B/en
Publication of CN108573264A publication Critical patent/CN108573264A/en
Application granted granted Critical
Publication of CN108573264B publication Critical patent/CN108573264B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23211—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with adaptive number of clusters
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Biophysics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Evolutionary Biology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

Household industry potential customers' recognition methods based on novel bee group's clustering algorithm that the present invention relates to a kind of, includes the following steps：Cluster centre is selected from client set and is encoded；The initial position at all artificial sweet peaks is given at random；It sorts to all artificial honeybees with the fitness according to artificial honeybee, HN position is as food source before therefrom choosing；Cluster operation is carried out according to artificial honeybee current location, and updates artificial honeybee position：More New food source.The present invention realizes simplicity, does not rely on the selection of parameter excessively, has the advantages that stronger ability of searching optimum, fast convergence rate, accuracy of identification is high, household industry potential customers are identified with the clustering problem of this complexity, there is obviously Statistical error effect.

Description

A kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm
Technical field
Household industry potential customers' recognition methods based on novel bee group's clustering algorithm that the present invention relates to a kind of, belongs to household The e-commerce field of industry, while being related to Swarm Intelligence Algorithm and clustering algorithm field.
Background technology
With the needs of progress and the enterprise market sane development of scientific idea and technology, customer capital is as enterprise One important intangible asset, importance have had received widespread attention, become weigh enterprise market value key element it One.Under the market environment of " customer-centric ", can the real demand that be well understood by customer action and market have become Determine the key of enterprise competitiveness.The success or not of enterprise is largely dependent on whether enterprise can quickly and accurately respond visitor The demand at family, and the business strategy for the opponent that copes with competition variation.Thus, it is provided how to efficiently use the client of magnanimity, dispersion Source supports the intelligent marketing decision of small/medium enterprise group to be one and have key technology to be solved.
The clustering problem of this complexity is identified for household industry potential customers, if using traditional data digging method It solves, then can not simultaneously do the trick at two aspects of accuracy of identification and efficiency of client.In recent years, simulation honeybee looks for The artificial bee colony intelligent optimization algorithm of food behavior has obtained the extensive concern of scholars, uses it to solve in conjunction with clustering algorithm such Problem yields good result, and shows the clustering algorithm based on bio-inspired computing in solving the problems, such as target identification Unique advantage.But the ant group algorithm more early proposed or genetic algorithm etc. realize complexity, stability is poor, the cluster result of optimization Randomness is big；Although the particle cluster algorithm speed of service is very fast, it is often stuck with the part of solution space, is unable to get global optimum Solution, it is more helpless especially for high-dimensional optimization；The pure clustering algorithms such as classical k-means, are influenced by parameter Larger, as a result randomness is larger, cannot achieve the accurate identification to client.These already present algorithms are relative complex in solution When optimization problem, performance can't reach satisfied precision and stability requirement.
Invention content
It is lacked for what existing calculating was exposed when household industry potential customers identify the clustering problem of this complexity It falls into, the present invention proposes a kind of family based on novel bee group's clustering algorithm using for reference multiple honeybees cooperation foraging behaviors in the Nature Occupy industry potential customers' recognition methods.
The technical solution adopted by the present invention to solve the technical problems is：A kind of household based on novel bee group's clustering algorithm Industry potential customers' recognition methods, includes the following steps：
1) from client set X={ x1,x2,...,xNIn arbitrarily select c point as cluster centre y1,y2,...,yc, and It is encoded：
Wherein,Indicate the coding of any one artificial honeybee；G indicates role of the artificial honeybee in population；S indicates people Label of the work honeybee in the subgroups role g, t indicate current iterative step；
2) initial position at all artificial sweet peaks is given at random；All artificial honeybees are arranged with the fitness according to artificial honeybee Sequence, HN position is as food source before therefrom choosing；
3) cluster operation is carried out according to artificial honeybee current location, and the suitable of each artificial honeybee is obtained according to cluster result Response；Artificial honeybee position is updated by following steps：
3-1) to employing bee or bee being followed to carry out location updating：vsq=ysq+δsq(ysq-yhq)
Wherein, s=1,2 ..., n；N indicates population scale；H be it is determining at random, it is not identical as s；vsqIt is employed after representing update It hires bee or follows the position of bee；ysqRepresentative currently employs bee or follows the position of bee；yhqRepresentative, which randomly selects, employs bee or follows The position of bee；δsqFor parameter, randomly generated in [- 1,1] range；
Location updating 3-2) is carried out to investigation bee：
Wherein, σ is the random number in [- 1,1] range；Represent the updated position of investigation bee；Represent investigation bee A minimum dimension in the vector of current location；Represent a maximum dimension in the vector of investigation bee current location；
WhenWhen, this search bee becomes employing bee；Indicate the fitness of the t times iteration,In expression The maximum fitness of an iteration intermediate value；
4) more New food source：The fitness of all artificial honeybee individual current locations is calculated, therefrom selects m a more than original M position of fitness minimum in original food source is replaced in the position of food source fitness；
5) if current iterations have reached preset maximum times TmaxOr final result is less than convergence essence ξ is spent, then stops iteration, exports current artificial honeybee position as final cluster result；Otherwise return to step 3).
The fitness：
Wherein,
Wherein, N is the quantity of client, feFor fitness, rxnFor parameter.
The artificial honeybee current location of basis carries out cluster operation and includes the following steps：
1.1) using artificial honeybee current location as v1,v2,...,vc, i.e. c cluster centre；
1.2) with v1,v2,...,vcCentered on point to client set X into row set divide：
If | | xk-vi||2≤||xk-vj||2,xk∈ X, i=1,2 ..., c, j=1,2 ..., c, i ≠ j, then by xkIt draws Assign to cluster client set AiIn, wherein
1.3) new cluster centre is calculated：
Wherein, NiIndicate i-th of set AiThe quantity of middle vector；
1.4) v is enabledi=v 'i, calculate between class distance：Between class distance is less than to the x of threshold valuek It is divided into point v 'iClass in；
1.5) when this cluster centre is unchanged compared with last time cluster centre, calculating terminates, cluster centre at this time For cluster result；Otherwise, return to step 1.2).
Classified to client according to final cluster result, completes identification.
The invention has the advantages that and advantage：
1. in order to more preferably solve these problems, honeybee foraging behavior is simulated, in conjunction with classical clustering algorithm and incorporates mould person in servitude Category degree concept, has invented household industry potential customers' recognition methods based on novel bee group's clustering algorithm, which realizes certainly The advantages that adaptation cluster, has ability of searching optimum strong, fast convergence rate, accuracy of identification is high.
2. algorithm realizes easy, the not excessively selection of dependence parameter, with stronger ability of searching optimum, convergence rate Soon, the advantages that accuracy of identification is high identifies household industry potential customers the clustering problem of this complexity, has obviously excellent Change recognition effect.
Description of the drawings
Fig. 1 is the overall construction drawing of client's identification process；
Fig. 2 is cluster result figure of the bee colony clustering algorithm to client properties collection.
Specific implementation mode
With reference to embodiment, the present invention is described in further detail.
As Fig. 1 illustrates the overall structure of client's identification process.First, from customer database data acquisition, such as client Essential information, preference behavior of client etc. form training and test sample set, obtain the height of the quality of data largely On affect the final quality for obtaining result；Then, in order to realize the cluster of client, potential customers' identification model is established, training The model gone out needs after assessment, just can be used for potential customers' identification；Finally, according to potential customers' identification model of structure, The client newly accessed is identified, finds potential customers, carries out target marketing.
Step 1：Establish client's identification model based on k-means clustering methods
1.1) cluster number c is given；
1.2) from client set X={ x1,x2,...,xNIn arbitrarily select c points v1,v2,...,vcIt is poly- respectively as c The cluster centre of class set；Wherein, xkClient is indicated to the preference of family product, including color, the information such as material, for example (,) it is certain The family product etc. of consumer preference red and glass material；K=1 ..., N；
1.3) with v1,v2,...,vcCentered on point X is divided into row set, the principle of division is：If | | xk-vi||2≤ ||xk-vj||2,xk∈ X, i=1,2 ..., c, j=1,2 ..., c, i ≠ j, then by xkIt is divided into cluster client set AiIn, In
1.4) according to cluster client set A1,A2,...,AcIn point calculate new central point：
Wherein NiIndicate set AiThe quantity of middle vector；
1.5) v is enabledi=v 'i, between class distance is calculated according to the following formula：
1.6) terminate when cluster centre no longer changes calculating, otherwise, return to step 1.3).
Step 2：Coding
The core of clustering algorithm is the determination of cluster centre, so to (client for updating the data i.e. testing data again collects Close X={ x1,x2,...,xNIn arbitrarily select c point) cluster centre y1,y2,...,ycIt is encoded.Any artificial honeybee can It is as follows to encode：
Wherein,Indicate the coding (position) of any one artificial honeybee；G indicates role of the artificial honeybee in population；s Indicate that label of the artificial honeybee in the subgroups role g, t indicate current iterative step；ylIndicate first of cluster centre, wherein l =1,2 ..., c；
Step 3：Initialize all kinds of parameters
Given cluster number c；Population scale n, the i.e. number of artificial bee colony；The initial of all artificial peak individuals is given at random Position；With sorting from big to small to all individuals according to fitness, HN position is as food source before therefrom choosing；Greatest iteration time Number Tmax, convergence precision ξ；
Step 4：Data update
Artificial bee shares three types：It employs bee, follow bee and investigation bee.Wherein, the quantity of bee is employed to account for sum Half, remaining is to follow bee and search bee.What search bee carried out is the exploration in space, and employs bee and follow bee point not same order Section executes development process in search space.
By the cluster result of step 1, the fitness (functional value) of artificial bee is calculated according to following formula, it is therefore an objective to make cluster Error rate is minimum：
Wherein,
Wherein, N is the quantity for needing to identify client, feFor error rate.Here sentenced according to historic customer grouped data Disconnected clustering result.
Current iteration number t=0 is set, after completing initial phase, following two iteration is executed to each artificial honeybee Process：
3.1) using following formula to employing bee and bee being followed to carry out location updating：
vsq=ysq+δsq(ysq-yhq)
Wherein, s=1,2 ..., n；Q=1,2 ..., c are the indexes randomly selected, and h is determining at random, but cannot be with S is identical；vsqBee is employed after representing update and follows the position of bee；ysqBee is employed before representing update and follows the position of bee；yhqGeneration Table randomly selects the position employed bee or follow bee；δsqIt is randomly generated in [- 1,1] range, this state modulator in ysqWeek The generation in the new nectar source enclosed.
3.2) location updating is carried out to investigation bee using following formula：
Wherein, σ is the random number in [- 1,1] range；Represent the updated position of investigation bee；Represent investigation bee A minimum dimension in the vector of current location；Represent a maximum dimension in the vector of investigation bee current location.WhenWhen, i.e., after search bee finds the abundanter nectar source of honey content, this search bee will become one and employ bee. Indicate the fitness of the t times iteration of the search bee,Indicate the maximum value of all artificial honeybee fitness in last iteration；
Step 5：More New food source：The fitness of all honeybee individuals current location is calculated, therefrom selection is better than original food Replace position poor in food source in the position of material resource fitness；
Step 6：If current iterations have reached preset maximum times TmaxOr final result is less than in advance Determine convergence precision ξ requirements, then stops iteration, cluster result is exported, as shown in Fig. 2, otherwise, t=t+1 goes to step 4.
The cluster centre of cluster result indicates that client to the preference of family product, divides client for cluster centre Group completes client's identification.

Claims (4)

1. a kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm, it is characterised in that including following step Suddenly：
1) from client set X={ x1,x2,...,xNIn arbitrarily select c point as cluster centre y1,y2,...,yc, and carry out Coding：
Wherein,Indicate the coding of any one artificial honeybee；G indicates role of the artificial honeybee in population；S indicates artificial honey Label of the bee in the subgroups role g, t indicate current iterative step；
2) initial position at all artificial sweet peaks is given at random；It sorts to all artificial honeybees with the fitness according to artificial honeybee, HN position is as food source before therefrom choosing；
3) cluster operation is carried out according to artificial honeybee current location, and the adaptation of each artificial honeybee is obtained according to cluster result Degree；Artificial honeybee position is updated by following steps：
3-1) to employing bee or bee being followed to carry out location updating：vsq=ysq+δsq(ysq-yhq)
Wherein, s=1,2 ..., n；N indicates population scale；H be it is determining at random, it is not identical as s；vsqBee is employed after representing update Or follow the position of bee；ysqRepresentative currently employs bee or follows the position of bee；yhqRepresentative, which randomly selects, employs bee or follows bee Position；δsqFor parameter, randomly generated in [- 1,1] range；
Location updating 3-2) is carried out to investigation bee：
Wherein, σ is the random number in [- 1,1] range；Represent the updated position of investigation bee；It is current to represent investigation bee A minimum dimension in position vector；Represent a maximum dimension in the vector of investigation bee current location；
WhenWhen, this search bee becomes employing bee；Indicate the fitness of the t times iteration,Indicate that the last time changes For the maximum fitness of intermediate value；
4) more New food source：The fitness of all artificial honeybee individual current locations is calculated, therefrom selects m to be more than original food M position of fitness minimum in original food source is replaced in the position of source fitness；
5) if current iterations have reached preset maximum times TmaxOr final result is less than convergence precision ξ, Then stop iteration, exports current artificial honeybee position as final cluster result；Otherwise return to step 3).
2. a kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm according to claim 1, It is characterized in that the fitness：
Wherein,
Wherein, N is the quantity of client, feFor fitness, rxnFor parameter.
3. a kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm according to claim 1, Include the following steps it is characterized in that the artificial honeybee current location of basis carries out cluster operation：
1.1) using artificial honeybee current location as v1,v2,...,vc, i.e. c cluster centre；
1.2) with v1,v2,...,vcCentered on point to client set X into row set divide：
If | | xk-vi||2≤||xk-vj||2,xk∈ X, i=1,2 ..., c, j=1,2 ..., c, i ≠ j, then by xkIt is divided into Cluster client set AiIn, wherein
1.3) new cluster centre is calculated：
Wherein, NiIndicate i-th of set AiThe quantity of middle vector；
1.4) v is enabledi=v 'i, calculate between class distance：Between class distance is less than to the x of threshold valuekIt is divided into a little v′iClass in；
1.5) when this cluster centre is unchanged compared with last time cluster centre, calculating terminates, and cluster centre at this time is poly- Class result；Otherwise, return to step 1.2).
4. a kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm according to claim 1, It is characterized in that classifying to client according to final cluster result, identification is completed.
CN201710130878.0A 2017-03-07 2017-03-07 Household industry potential customer identification method based on novel swarm clustering algorithm Expired - Fee Related CN108573264B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710130878.0A CN108573264B (en) 2017-03-07 2017-03-07 Household industry potential customer identification method based on novel swarm clustering algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710130878.0A CN108573264B (en) 2017-03-07 2017-03-07 Household industry potential customer identification method based on novel swarm clustering algorithm

Publications (2)

Publication Number Publication Date
CN108573264A true CN108573264A (en) 2018-09-25
CN108573264B CN108573264B (en) 2021-07-20

Family

ID=63577110

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710130878.0A Expired - Fee Related CN108573264B (en) 2017-03-07 2017-03-07 Household industry potential customer identification method based on novel swarm clustering algorithm

Country Status (1)

Country Link
CN (1) CN108573264B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110852380A (en) * 2019-11-11 2020-02-28 安徽师范大学 Quantum ant lion and k-means based clustering method and intrusion detection method
CN113034157A (en) * 2019-12-24 2021-06-25 中国移动通信集团浙江有限公司 Group member identification method and device and computing equipment

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101071437A (en) * 2007-03-28 2007-11-14 腾讯科技（深圳）有限公司 User classifying method, directional advertising launching method, device and system
US20100210025A1 (en) * 2006-08-15 2010-08-19 Victor Chang Cardiac Research Institute Limited Common Module Profiling of Genes
CN101894273A (en) * 2010-05-26 2010-11-24 北京航空航天大学 Artificial bee colony refine edge potential field function-based unmanned plane target identification method
CN102779224A (en) * 2011-05-13 2012-11-14 三商行股份有限公司 Meal ordering system and device
CN103116693A (en) * 2013-01-14 2013-05-22 天津大学 Hardware and software partitioning method based on artificial bee colony
CN105100898A (en) * 2015-08-13 2015-11-25 海信集团有限公司 Intelligent television starting method and system
CN105446159A (en) * 2016-01-08 2016-03-30 北京光年无限科技有限公司 Intelligent household system and data processing method thereof
CN106373023A (en) * 2015-07-23 2017-02-01 中国科学院沈阳自动化研究所 Batching optimization method based on new multi-objective artificial bee colony algorithm

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100210025A1 (en) * 2006-08-15 2010-08-19 Victor Chang Cardiac Research Institute Limited Common Module Profiling of Genes
CN101071437A (en) * 2007-03-28 2007-11-14 腾讯科技（深圳）有限公司 User classifying method, directional advertising launching method, device and system
CN101894273A (en) * 2010-05-26 2010-11-24 北京航空航天大学 Artificial bee colony refine edge potential field function-based unmanned plane target identification method
CN102779224A (en) * 2011-05-13 2012-11-14 三商行股份有限公司 Meal ordering system and device
CN103116693A (en) * 2013-01-14 2013-05-22 天津大学 Hardware and software partitioning method based on artificial bee colony
CN106373023A (en) * 2015-07-23 2017-02-01 中国科学院沈阳自动化研究所 Batching optimization method based on new multi-objective artificial bee colony algorithm
CN105100898A (en) * 2015-08-13 2015-11-25 海信集团有限公司 Intelligent television starting method and system
CN105446159A (en) * 2016-01-08 2016-03-30 北京光年无限科技有限公司 Intelligent household system and data processing method thereof

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
JUN ZHANG 等: "ABCluster: the artificial bee colony algorithm for cluster global optimization", 《PHYS. CHEM. CHEM. PHYS》 *
高雷阜 等: "融合改进遗传和人工蜂群的SVM参数优化算法", 《计算机工程与应用》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110852380A (en) * 2019-11-11 2020-02-28 安徽师范大学 Quantum ant lion and k-means based clustering method and intrusion detection method
CN110852380B (en) * 2019-11-11 2022-08-02 安徽师范大学 Quantum ant lion and k-means based clustering method and intrusion detection method
CN113034157A (en) * 2019-12-24 2021-06-25 中国移动通信集团浙江有限公司 Group member identification method and device and computing equipment
CN113034157B (en) * 2019-12-24 2023-12-26 中国移动通信集团浙江有限公司 Group member identification method and device and computing equipment

Also Published As

Publication number Publication date
CN108573264B (en) 2021-07-20

Similar Documents

Publication Publication Date Title
CN108363821A (en) A kind of information-pushing method, device, terminal device and storage medium
CN111667022A (en) User data processing method and device, computer equipment and storage medium
CN111179016B (en) Electricity selling package recommending method, equipment and storage medium
CN111259140B (en) False comment detection method based on LSTM multi-entity feature fusion
CN109934615B (en) Product marketing method based on deep sparse network
CN106503731A (en) A kind of based on conditional mutual information and the unsupervised feature selection approach of K means
CN110222838B (en) Document sorting method and device, electronic equipment and storage medium
CN110134803A (en) Image data method for quickly retrieving based on Hash study
CN112200667B (en) Data processing method and device and computer equipment
CN111967971A (en) Bank client data processing method and device
CN108230029A (en) Client trading behavior analysis method
CN108573264A (en) A kind of household industry potential customers' recognition methods based on novel bee group's clustering algorithm
CN117436679B (en) Meta-universe resource matching method and system
Yu et al. Representation learning based on autoencoder and deep adaptive clustering for image clustering
CN113469288A (en) High-risk personnel early warning method integrating multiple machine learning algorithms
CN109583712B (en) Data index analysis method and device and storage medium
Arockiam et al. Reclust: an efficient clustering algorithm for mixed data based on reclustering and cluster validation
CN114240539B (en) Commodity recommendation method based on Tucker decomposition and knowledge graph
CN115829683A (en) Power integration commodity recommendation method and system based on inverse reward learning optimization
CN108805199B (en) Entity business marketing method based on genetic algorithm
CN108268898A (en) A kind of electronic invoice user clustering method based on K-Means
Xiong et al. L-RBF: A customer churn prediction model based on lasso+ RBF
Waltner et al. Hibster: Hierarchical boosted deep metric learning for image retrieval
CN112488773A (en) Smart television user classification method, computer equipment and storage medium
CN105912887B (en) A kind of modified gene expression programming-fuzzy C-mean algorithm crop data sorting technique

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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210720

CF01 Termination of patent right due to non-payment of annual fee