CN104281877A - Human activity area classification method based on improved genetic cluster - Google Patents
Human activity area classification method based on improved genetic cluster Download PDFInfo
- Publication number
- CN104281877A CN104281877A CN201410478559.5A CN201410478559A CN104281877A CN 104281877 A CN104281877 A CN 104281877A CN 201410478559 A CN201410478559 A CN 201410478559A CN 104281877 A CN104281877 A CN 104281877A
- Authority
- CN
- China
- Prior art keywords
- distance
- class
- fitness
- cluster
- chromosome
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000002068 genetic effect Effects 0.000 title claims abstract description 26
- 230000000694 effects Effects 0.000 title claims abstract description 22
- 210000000349 chromosome Anatomy 0.000 claims description 48
- 238000009395 breeding Methods 0.000 claims description 10
- 230000001488 breeding effect Effects 0.000 claims description 10
- 230000002759 chromosomal effect Effects 0.000 claims description 9
- 108090000623 proteins and genes Proteins 0.000 claims description 9
- 208000001613 Gambling Diseases 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 claims description 4
- 206010064571 Gene mutation Diseases 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 claims description 3
- 230000004048 modification Effects 0.000 claims description 3
- 238000012986 modification Methods 0.000 claims description 3
- 230000035772 mutation Effects 0.000 claims description 3
- 230000008929 regeneration Effects 0.000 claims description 3
- 238000011069 regeneration method Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 abstract description 4
- 230000004083 survival effect Effects 0.000 abstract description 4
- 230000003044 adaptive effect Effects 0.000 abstract 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a human activity area classification method based on an improved genetic cluster. The method includes the following steps of firstly, population initialization, secondly, population updating, thirdly, administrative subordination relationship judgment, fourthly, fitness calculation, fifthly, individual selection, sixthly, cross propagation, seventhly, variation and eighthly, result judgment. According to the method, an original genetic cluster is redefined for large-scale human movement data, the new fitness of the intra-class distance and the new fitness of the between-class distance are provided, the intra-class distance and the between-class distance which are not contained in a traditional judgment criterion are optimized, and therefore comprehensive balance optimum is acquired; the genetic evolution process of creature in the natural environment is simulated for adaptive overall structure optimization, an individual with the largest fitness is acquired in the mode of survival of the fittest and serves as the optimal cluster center, and therefore the problem of local optimum caused by random selection of the cluster center is solved.
Description
Technical field
The present invention relates to a kind of mankind's activity region sorting technique based on improving genetic cluster, belonging to mankind's Research of Mobility and Data Mining.
Background technology
Traditional mankind's activity region sorting technique utilizes mankind's Mobile data, and based on classical automatic clustering method, wherein K average algorithm (K-MEANS) is the most conventional.
The widely used clustering criteria of classical automatic clustering method is weighted error quadratic sum in class, by iteration function, makes clustering criteria reach minimal value as objective function, tries to achieve a locally optimal solution.This objective function is a non-convex function, has multiple minimal value and a minimum value.Minimal value is considered from local, and the value being less than its both sides is minimal value, and minimum value is only and considers from the overall situation.Meanwhile, clustering criteria only considers one of inter-object distance, both between class distances, makes it reach extreme value, to represent within a class that similarity is large, otherness is large between inhomogeneity.
In the classification of mankind's activity region, mankind's Mobile data is in large scale, because the efficiency of K-MEANS is high, becomes the first-selection of large-scale data being carried out to cluster.K-MEANS Stochastic choice K object, each object initially represents mean value or the center of a class; It, according to the distance at itself and all kinds of center, is assigned to nearest class by residue object; Then the mean value of each class is recalculated.This process constantly repeats, until criterion function converges to minimal value.The randomness at initial selected class center brings the consequence of local optimum.
Summary of the invention
(1) object:
For the shortcoming of prior art, the technical matters that the application will solve is: the local optimum that the randomness at initial selected class center causes, and clustering criteria does not consider the comprehensive optimal situation of inter-object distance, between class distance balance simultaneously.
The problem that the technology of the present invention will solve is: the deficiency overcoming the existing clustering technique for mankind's Mobile data, a kind of mankind's activity region sorting technique based on improving genetic cluster is provided, for original genetic cluster redefines the new Judging index comprising inter-object distance and between class distance simultaneously, the biological genetic evolution process in physical environment of simulation carries out the optimization of self-adaptation global structure, obtain Optimal cluster centers with the pattern of " survival of the fittest ", solve the local optimum that Stochastic choice class center is brought.
(2) technical scheme:
The technical solution used in the present invention is a kind of mankind's activity region sorting technique based on improving genetic cluster, takes into account inter-object distance and between class distance, revises cluster centre.
A kind of mankind's activity region sorting technique based on improving genetic cluster of the present invention, the method includes the following step:
Step one: initialization of population
Input individual O to be sorted
i(i=1,2,3..., N), each individuality comprises the proper vector of d dimension, namely
O
i={a
(i,1),a
(i,2)...,a
(i,d)}
A
(i, j)(j=1,2..., d) represents one of them proper vector;
All individualities are divided into k class, then Stochastic choice k individuality represents k initial cluster center, the cluster centre C of m (1≤m≤k) individual class
m=O
m; All cluster centres weave into item chromosome Ind, namely
Ind={C
1,C
2,C
3...,C
k={a
(1,1),a
(1,2)...,a
(1,d),...,a
(k,1),a
(k,2)...,a
(k,d)}
This chromosome is as the former generation of in initial population; Above chromosome manufacture process repeats PSize time according to Population Size; Setting Population Size
namely in whole individuality, select whole not repeated combination numbers of any k, ensure that in population, all former generation are not identical; All former generation's chromosome forms initial population;
Step 2: Population Regeneration
Each former generation's chromosome Ind, comprises k cluster centre; For one of them cluster centre C
m, find all original individual middle distance C to be sorted
mn nearest individuality (comprises C
m), get the average of this n individuality, as replacement C
mnew cluster centre; Other cluster centres in former generation's chromosome Ind upgrade in the same way, and all former generation's chromosomes in final population upgrade, and become a new population;
Step 3: judge membership
For comprising k cluster centre { C
1, C
2, C
3..., C
keach former generation's chromosome Ind, all original individualities to be sorted can be divided into k group; Body O one by one
i, distance m cluster centre is nearest, then judge that it belongs to m group;
Step 4: calculate fitness
Fitness is the foundation that genetic cluster technology judges search, and the probability that the individuality that fitness is high participates in offspring's breeding is higher; The fitness definition improving genetic cluster both comprised inter-object distance, also comprised between class distance; Inter-object distance is less, between class distance is larger, then Clustering Effect is better, and corresponding fitness is larger; For each former generation's chromosome, according to k the group that the membership of step 3 draws, calculate inter-object distance S
in, between class distance S
out.Fitness f is
Step 5: individual choice
Selection is to obtain excellent former generation, and the probability raised up seed that fitness is high is high; The chromosomal fitness of all former generation arranges from big to small, gets the former generation of front 60% as the bred individuality of surviving;
Step 6: intersect and breed
Adopt the strategy of wheel disc gambling to carry out parents' selection, namely using the circumference of all chromosomal fitness sums as wheel disc, each chromosome occupies a sector according to fitness ratio, and fitness Gao Ze is higher by the probability chosen at random in wheel disc rotates; Often select two chromosomes, carry out intersection breeding, namely intercourse chromosomal half, form two other child chromosome different from parents;
Step 7: variation
Except inheriting parent information, also can there is the genetic mutation of certain probability in the filial generation that the breeding that intersects produces; In chromosome, the variable of every one dimension can represent a gene, and therefore gene is exactly the proper vector in k cluster centre; The probability of setting gene mutation is P=0.5%, a chromosome in Stochastic choice population of new generation, a gene g on this chromosome of random selecting
value, produce as lower variation
g
value=g
value±P×g
value
Step 8: result judges
Setting hereditary maximum algebraically is Y
max, above step 2 is less than Y to the iterations of step 7
max, then jump procedure two proceeds, and increases an iteration; Otherwise according to step 4, calculate the chromosome that in final filial generation, fitness is maximum, k cluster centre on it is complete modification cluster centre, then according to step 3, judge the final membership class of original individuality to be sorted.
Wherein, " the setting Population Size described in step one
account form is as follows:
Symbol description in formula: k represents the number of final class object; N represents individual amount to be sorted;
represent the combined number selecting k individuality from individuality.
Wherein, " distance " described in step 2, referring to the Euclidean distance (Euclidean Distance) between two individualities, is the line segment length in n-dimensional space between 2; For given two some p=(p
1, p
2..., p
n), q=(q
1, q
2..., q
n), their distance is calculated as follows:
Symbol description in formula: p, q are two given points; p
i, q
i(i=1,2 ..., n) represent p respectively, the coordinate vector of q in n-dimensional space; D (p, q) represents the distance of p to q; D (q, p) represents the distance of q to p.
Wherein, " distance " described in step 3, identical with the distance defined in step 2.
Wherein, " the inter-object distance S described in step 4
in, between class distance S
out", refer to that, for the group of the k in step 3, each group as a class:
Inter-object distance refers in this k class, the individuality of all classes to the distance average sum at its center, namely
Between class distance is the distance sum of this k Ge Leilei center to all centers mean value, namely
Above inter-object distance S
in, between class distance S
outdefinition in, its symbol description is as follows:: k is the number of class; N
i(i=1,2 ..., k) represent the scale of corresponding class; C
1(i=1,2 ..., k) represent corresponding Lei Lei center;
represent the individuality in the i-th class respectively; D (p, q) represents the distance of individual p to q.
Wherein, " iterations " described in step 8, refers to the execution number of times from step 2 to step 7, every circulation primary, and iterations increases by 1.
(3) advantage and effect:
The present invention's advantage compared with prior art and effect are:
(1) instant invention overcomes existing mankind's activity region cluster Local Search, shortcoming that cluster centre locates the result local optimum caused at random, utilize the global structure optimization that population biological heredity is evolved, go out Optimal cluster centers by the mode correction of the survival of the fittest, and then obtain optimum solution.
(2) lateral comparison, the global optimizing iteration of employing genetic evolution randomization of the present invention, compared with other classical automatic clustering methods, possesses the ability obtaining global optimum's classification.Longitudinal comparison, the more original genetic cluster of the present invention has considered the balance fitness definition of inter-object distance, between class distance, instead of only only considers the way of one of them in traditional cluster.
Accompanying drawing explanation
Fig. 1 is the method for the invention process flow diagram.
Embodiment
In order to understand the present invention better, first explanation is explained once to some concepts.
1. the distance between vector: this programme adopts Euclidean distance (Euclidean Distance), is the line segment length in n-dimensional space between 2.For given two some p=(p
1, p
2..., p
n), q=(q
1, q
2..., q
n), their distance is calculated as follows:
2. between class distance in class: in this programme, with Euclidean distance (Euclidean Distance) for standard, suppose total k class, the scale of corresponding class is { N
1, N
2, N
3..., N
k, center corresponds to { C
1, C
2, C
3..., C
k, the individuality in the i-th class is
Inter-object distance is in this k class, the individuality of all classes to the distance average sum at its center, namely
Between class distance is the distance sum of this k Ge Leilei center to all centers mean value, namely
3. fitness: fitness is that the basis of genetic cluster method relies on index, the possibility size representing individual survival and raise up seed.Fitness is larger, survive and the probability that raises up seed higher.Its mathematical definition is carried out based between class distance in class.
4. wheel disc gambling back-and-forth method: wheel disc gambling back-and-forth method selects the method for some members from colony, and selected probability is proportional with their selection gist numerical value, and numerical value is higher, and the sector area occupied in wheel disc is larger, and selected probability is also higher.This does not ensure that the member one that adaptability mark is the highest is selected into the next generation surely, only illustrates that it has maximum probability selected.
Whole implementation procedure is as follows:
A kind of mankind's activity region sorting technique based on improving genetic cluster of the present invention, as shown in Figure 1, the method includes following implementation step:
Step one: initialization of population
Mankind's Mobile data, by (time, position, quantity) and at least one variable of comprising respectively thereof, is formed with different combinations.Input individual O to be sorted
i(i=1,2,3..., N), each individuality comprises the proper vector that d (d>=3) ties up, namely
O
i={a
(i,1),a
(i,2)...,a
(i,d)}
A
(i, j)(j=1,2..., d) represents one of them proper vector.
All individualities are divided into k class, then Stochastic choice k individuality represents k initial cluster center, the cluster centre C of m (1≤m≤k) individual class
m=O
m.All cluster centres weave into item chromosome Ind, namely
Ind={C
1,C
2,C
3...,C
k}={a
(1,1),a
(1,2)...,a
(1,d)...,a
(k,1),a
(k,2)...,a
(k,d),}
This chromosome is as the former generation of in initial population.Above chromosome manufacture process repeats PSize time according to Population Size.Setting Population Size
namely in whole individuality, select whole not repeated combination numbers of any k, ensure that in population, all former generation are not identical.All former generation's chromosome forms initial population.
Step 2: Population Regeneration
Each former generation's chromosome Ind, comprises k cluster centre.For one of them cluster centre C
m, find all original individual middle distance C to be sorted
mnearest n individual { O
1, O
2..., O
n(wherein comprise C
m) get this n individual average, as replacement C
mnew cluster centre
namely
Other cluster centres in former generation's chromosome Ind upgrade in the same way.All former generation's chromosomes in final population upgrade, and become a new population.
Step 3: judge membership
For comprising k cluster centre { C
1, C
2, C
3..., C
keach former generation's chromosome Ind, all original individualities to be sorted can be divided into k group.Body O one by one
idistance m cluster centre recently, namely and if only if D (O
i, C
m)≤D (O
i, C
n), when n=1,2..., k, judge O
ibelong to m group.
Step 4: calculate fitness
The probability that the individuality that fitness is high participates in offspring's breeding is higher.The fitness definition improving genetic cluster both comprised inter-object distance, also comprised between class distance.Inter-object distance is less, between class distance is larger, then Clustering Effect is better, and corresponding fitness is larger.For each former generation's chromosome, according to k the group that the membership of step 3 draws, calculate inter-object distance S
in, between class distance S
out.Fitness f is
Step 5: individual choice
Selection is to obtain excellent former generation, and the probability raised up seed that fitness is high is high.The chromosomal fitness of all former generation arranges from big to small, gets the former generation of front 60% as the bred individuality of surviving.
Step 6: intersect and breed
Adopt the strategy of wheel disc gambling to carry out parents' selection, namely using the circumference of all chromosomal fitness sums as wheel disc, each chromosome occupies a sector according to fitness ratio, and fitness Gao Ze is higher by the probability chosen at random in wheel disc rotates.Often select two chromosomes, carry out intersection breeding, namely intercourse chromosomal half, be formed at two other child chromosome that parents are different.
Step 7: variation
Except inheriting parent information, also can there is the genetic mutation of certain probability in the filial generation that the breeding that intersects produces.In chromosome, the variable of every one dimension can represent a gene, and therefore gene is exactly the proper vector in k cluster centre.The probability of setting gene mutation is P=0.5%, a chromosome in Stochastic choice population of new generation, a gene g on this chromosome of random selecting
value, produce as lower variation
g
value=g
value±P×g
value
Step 8: result judges
The maximum algebraically of initial setting heredity is Y
max, step 2 is less than Y to the iterations of step 7
max, then jump procedure two order performs, and increases an iteration; Otherwise according to step 4, calculate the chromosome that in final filial generation, fitness is maximum, k cluster centre on it is complete modification cluster centre, then according to membership, judge the final membership class of original individuality to be sorted.
The present invention is applied to the classification of extensive mankind's Mobile data, and classifying quality is reasonable.As being applied to Beijing's metro passenger flow, utilize the passenger flow that enters the station/set off of all subway stations, according to the volume of the flow of passengers of all website different time different scales, the distribution of subway station, Beijing is marked off different regions, has comprised residential district, workspace, shopping centre, tourist district etc.
In a word, present invention utilizes the global optimization ability of biological evolution, make the territorial classification of extensive mankind's Mobile data obtain globally optimal solution.Defined by the fitness of new model simultaneously, the inter-object distance simultaneously do not comprised in traditional judgment criterion, between class distance are optimized, and then obtain overall equilbrium optimum.
The part that the present invention does not describe in detail belongs to techniques well known.
Claims (6)
1., based on the mankind's activity region sorting technique improving genetic cluster, it is characterized in that: the method includes the following step:
Step one: initialization of population
Input individual O to be sorted
i(i=1,2,3..., N), each individuality comprises the proper vector of d dimension, namely
O
i={α
(i,1),α
(i,2)...,α
(i,d)}
α
(i, j)(j=1,2 ..., d) represent one of them proper vector;
All individualities are divided into k class, then Stochastic choice k individuality represents k initial cluster center, the cluster centre C of m (1≤m≤k) individual class
m=O
m; All cluster centres weave into item chromosome Ind, namely
Ind={C
1,C
2,C
3...,C
k}={α
(1,1),α
(1,2)...,α
(1,d),...,α
(k,1),α
(k,2)...,α
(k,d)}
This chromosome is as the former generation of in initial population; Above chromosome manufacture process repeats PSize time according to Population Size; Setting Population Size
namely in whole individuality, select whole not repeated combination numbers of any k, ensure that in population, all former generation are not identical; All former generation's chromosome forms initial population;
Step 2: Population Regeneration
Each former generation's chromosome Ind, comprises k cluster centre; For one of them cluster centre C
m, find all original individual middle distance C to be sorted
mn nearest individuality (comprises C
m), get the average of this n individuality, as replacement C
mnew cluster centre; Other cluster centres in former generation's chromosome Ind upgrade in the same way, and all former generation's chromosomes in final population upgrade, and become a new population;
Step 3: judge membership
For comprising k cluster centre { C
1, C
2, C
3..., C
keach former generation's chromosome Ind, all original individualities to be sorted are all divided into k group; Body O one by one
i, distance m cluster centre is nearest, then judge that it belongs to m group;
Step 4: calculate fitness
Fitness is the foundation that genetic cluster technology judges search, and the probability that the individuality that fitness is high participates in offspring's breeding is higher; The fitness definition improving genetic cluster both comprised inter-object distance, also comprised between class distance; Inter-object distance is less, between class distance is larger, then Clustering Effect is better, and corresponding fitness is larger; For each former generation's chromosome, according to k the group that the membership of step 3 draws, calculate inter-object distance S
in, between class distance S
out, fitness f is
Step 5: individual choice
Selection is to obtain excellent former generation, and the probability raised up seed that fitness is high is high; The chromosomal fitness of all former generation arranges from big to small, gets the former generation of front 60% as the breeding individuality of surviving;
Step 6: intersect and breed
Adopt the strategy of wheel disc gambling to carry out parents' selection, namely using the circumference of all chromosomal fitness sums as wheel disc, each chromosome occupies a sector according to fitness ratio, and fitness Gao Ze is higher by the probability chosen at random in wheel disc rotates; Often select two chromosomes, carry out intersection breeding, namely intercourse chromosomal half, form two other child chromosome different from parents;
Step 7: variation
Except inheriting parent information, also can there is the genetic mutation of predetermined probability in the filial generation that the breeding that intersects produces; In chromosome, the variable of every one dimension represents a gene, and therefore gene is exactly the proper vector in k cluster centre; The probability of setting gene mutation is P=0.5%, a chromosome in Stochastic choice population of new generation, a gene g on this chromosome of random selecting
value, produce as lower variation
g
value=g
value±P×g
value;
Step 8: result judges
Setting hereditary maximum algebraically is Y
max,above step 2 is less than Y to the iterations of step 7
max, then jump procedure two proceeds, and increases an iteration; Otherwise according to step 4, calculate the chromosome that in final filial generation, fitness is maximum, k cluster centre on it is complete modification cluster centre, then according to step 3, judge the final membership class of original individuality to be sorted.
2. a kind of mankind's activity region sorting technique based on improving genetic cluster according to claim 1, is characterized in that: " the setting Population Size described in step one
", account form is as follows:
Symbol description in formula: k represents the number of final class object; N represents individual amount to be sorted;
represent the combined number selecting k individuality from individuality.
3. a kind of mankind's activity region sorting technique based on improving genetic cluster according to claim 1, it is characterized in that: " distance " described in step 2, referring to the Euclidean distance between two individualities and Euclidean Distance, is the line segment length in n-dimensional space between 2; For given two some p=(p
1, p
2..., p
n), q=(q
1, q
2..., q
n), their distance is calculated as follows:
Symbol description in formula: p, q are two given points; p
i, q
i(i=1,2 ..., n) represent p respectively, the coordinate vector of q in n-dimensional space; D (p, q) represents the distance of p to q; D (q, p) represents the distance of q to p.
4. a kind of mankind's activity region sorting technique based on improving genetic cluster according to claim 1, is characterized in that: " distance " described in step 3, identical with the distance defined in step 2.
5. a kind of mankind's activity region sorting technique based on improving genetic cluster according to claim 1, is characterized in that: " the inter-object distance S described in step 4
in, between class distance S
out", refer to that, for the group of the k in step 3, each group as a class:
Inter-object distance refers in this k class, the individuality of all classes to the distance average sum at its center, namely
Between class distance is the distance sum of this k Ge Leilei center to all centers mean value, namely
Above inter-object distance S
in, between class distance S
outdefinition in, its symbol description is as follows:: k is the number of class; N
i(i=1,2 ..., k) represent the scale of corresponding class; C
i(i=1,2 ..., k) represent corresponding Lei Lei center;
represent the individuality in the i-th class respectively; D (p, q) represents the distance of individual p to q.
6. a kind of mankind's activity region sorting technique based on improving genetic cluster according to claim 1, it is characterized in that: " iterations " described in step 8, refer to the execution number of times from step 2 to step 7, every circulation primary, iterations increases by 1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410478559.5A CN104281877B (en) | 2014-09-18 | 2014-09-18 | A kind of mankind's activity region sorting technique based on improvement genetic cluster |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410478559.5A CN104281877B (en) | 2014-09-18 | 2014-09-18 | A kind of mankind's activity region sorting technique based on improvement genetic cluster |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104281877A true CN104281877A (en) | 2015-01-14 |
CN104281877B CN104281877B (en) | 2017-11-14 |
Family
ID=52256737
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410478559.5A Expired - Fee Related CN104281877B (en) | 2014-09-18 | 2014-09-18 | A kind of mankind's activity region sorting technique based on improvement genetic cluster |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104281877B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104700420A (en) * | 2015-03-26 | 2015-06-10 | 爱威科技股份有限公司 | Ellipse detection method and system based on Hough conversion and ovum identification method |
CN108717525A (en) * | 2018-05-09 | 2018-10-30 | 北京学之途网络科技有限公司 | A kind of information processing method, device, computer storage media and terminal |
CN112035224A (en) * | 2020-07-17 | 2020-12-04 | 中国科学院上海微系统与信息技术研究所 | Fog calculation scheduling method suitable for intelligent factory |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101923712A (en) * | 2010-08-03 | 2010-12-22 | 苏州大学 | Particle swarm optimization-based gene chip image segmenting method of K-means clustering algorithm |
CN102663230B (en) * | 2012-03-08 | 2015-06-10 | 武汉大学 | Method for land resource evaluation factor level classification based on genetic algorithm |
-
2014
- 2014-09-18 CN CN201410478559.5A patent/CN104281877B/en not_active Expired - Fee Related
Non-Patent Citations (3)
Title |
---|
FU XIANGWEI,LENG BIAO,XIONG ZHANG: "Metro traffic route assignment using K-Means clustering", 《ELECTRONICS,COMMUNICATIONS AND CONTROL(ICECC)》 * |
孙秀娟: "基于遗传算法的K_means聚类算法分析研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
杨丽娜: "基于遗传算法的数据挖掘技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104700420A (en) * | 2015-03-26 | 2015-06-10 | 爱威科技股份有限公司 | Ellipse detection method and system based on Hough conversion and ovum identification method |
CN104700420B (en) * | 2015-03-26 | 2019-02-15 | 爱威科技股份有限公司 | Ellipse detection method, system and worm's ovum recognition methods based on Hough transform |
CN108717525A (en) * | 2018-05-09 | 2018-10-30 | 北京学之途网络科技有限公司 | A kind of information processing method, device, computer storage media and terminal |
CN112035224A (en) * | 2020-07-17 | 2020-12-04 | 中国科学院上海微系统与信息技术研究所 | Fog calculation scheduling method suitable for intelligent factory |
CN112035224B (en) * | 2020-07-17 | 2024-03-12 | 中国科学院上海微系统与信息技术研究所 | Fog calculation scheduling method suitable for intelligent factory |
Also Published As
Publication number | Publication date |
---|---|
CN104281877B (en) | 2017-11-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105303450A (en) | Complex network community discovery method based on spectral clustering improved intersection | |
CN103150614B (en) | A kind of Automatic configuration method for land utilization space | |
JP2005202960A (en) | Optimization method and optimization program | |
CN107122843A (en) | A kind of traveling salesman problem method for solving based on improved adaptive GA-IAGA | |
CN105187051B (en) | One kind is used to not exclusively determine Reed Muller circuit power consumptions and area-optimized method based on NSGA II | |
CN102915448B (en) | A kind of three-dimensional model automatic classification method based on AdaBoost | |
CN109815541B (en) | Method and device for dividing rail transit vehicle product parts and modules and electronic equipment | |
CN103886375A (en) | Resource scheduling optimization method based on binary space partitioning tree | |
CN106991442A (en) | The self-adaptive kernel k means method and systems of shuffled frog leaping algorithm | |
CN104281877A (en) | Human activity area classification method based on improved genetic cluster | |
CN102880754A (en) | Method for identifying action scale of land utilization fractal dimension based on genetic algorithm | |
CN117974221B (en) | Electric vehicle charging station location selection method and system based on artificial intelligence | |
CN103279796A (en) | Method for optimizing genetic algorithm evolution quality | |
CN102663230B (en) | Method for land resource evaluation factor level classification based on genetic algorithm | |
CN103714211B (en) | Integrated circuit layouts method based on Move Mode sequence Yu multi-agent particle swarm | |
Jiang et al. | ATSA: An Adaptive Tree Seed Algorithm based on double-layer framework with tree migration and seed intelligent generation | |
CN104867164A (en) | Vector quantization codebook designing method based on genetic algorithm | |
Resende et al. | A distance-based selection of parents in genetic algorithms | |
CN114742593A (en) | Logistics storage center optimal site selection method and system | |
Shuai et al. | A Self-adaptive neuroevolution approach to constructing Deep Neural Network architectures across different types | |
CN103617206A (en) | Genetic clustering based on self-adaptation ecological niches | |
CN107679326A (en) | A kind of two-value FPRM circuit areas and delay comprehensive optimization method | |
CN111382210B (en) | Classification method, device and equipment | |
CN102629255A (en) | Regional population structure analysis method based on clustering analysis of mitochondrial deoxyribonucleic acid (mtDNA) genetic information | |
Yu et al. | An ensemble learning approach for predicting phenotypes from genotypes |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into 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: 20171114 |