CN108960486A - Interactive set evolvement method based on grey support vector regression prediction adaptive value - Google Patents
Interactive set evolvement method based on grey support vector regression prediction adaptive value Download PDFInfo
- Publication number
- CN108960486A CN108960486A CN201810603480.9A CN201810603480A CN108960486A CN 108960486 A CN108960486 A CN 108960486A CN 201810603480 A CN201810603480 A CN 201810603480A CN 108960486 A CN108960486 A CN 108960486A
- Authority
- CN
- China
- Prior art keywords
- individual
- adaptive
- support vector
- evolution
- population
- 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
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 47
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000002452 interceptive effect Effects 0.000 title claims abstract description 22
- 230000035772 mutation Effects 0.000 claims abstract description 23
- 238000011156 evaluation Methods 0.000 claims abstract description 10
- 230000006978 adaptation Effects 0.000 claims description 12
- 230000006399 behavior Effects 0.000 claims description 5
- 238000012706 support-vector machine Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000013461 design Methods 0.000 abstract description 6
- 238000005457 optimization Methods 0.000 description 22
- 230000002068 genetic effect Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 239000003795 chemical substances by application Substances 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 239000003086 colorant Substances 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 210000000349 chromosome Anatomy 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 238000000585 Mann–Whitney U test Methods 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 125000004432 carbon atom Chemical group C* 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000012733 comparative method Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biophysics (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Development Economics (AREA)
- Genetics & Genomics (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of interactive set evolutionary computation methods based on grey support vector regression prediction adaptive value, the specific steps of which are as follows: system provides design environment for user, generates initial Advanced group species at random before evolution starts；By interactive interface, people carries out numerical Evaluation to Nc individual；System clusters population according to individual similarity, and estimation clusters interior non-user and evaluates individual fitness, and predicts adaptive value using grey support vector regression；Set individual of evolving is carried out by Pareto is dominant sequence according to diversity, distributivity uncertainty measure, and set individual of evolving is operated using adaptive crossover and mutation, the generation interim population of same size；Parent population and interim population are merged, select top n individual as progeny population；Finally, being equidistantly divided into Nc unit to progeny population, 1 individual is randomly selected from each unit, Nc individual is recommended into user altogether.
Description
Technical Field
The invention belongs to the field of intelligent calculation, and particularly relates to an interactive set evolution optimization method based on a grey support vector regression prediction adaptive value, which is used for type selection of a color matching scheme.
Background
The application of interactive evolutionary computation based on heuristic learning, which was proposed in the 90 s of the 20 th century, to solving the implicit index optimization problem requires 2 basic problems to be solved: (1) how to effectively extract implicit knowledge; (2) how to solve implicit performance indicators with high quality.
For the 1 st problem, there are mainly 2 research strategies: firstly, under the mode of simultaneous interaction and evolution, implicit knowledge is directly extracted through a human-computer interaction interface. The method mainly focuses on the research of the adaptive value assignment mode, for example, in the 2 nd phase of the journal "automatic journal" published in 2014, "the interval adaptive value interactive genetic algorithm-based weighted multi-output gaussian process proxy model" adopts uncertain numbers such as interval numbers to express the adaptive values, and reflects the preference characteristics of the user; a journal published in 2017, namely ' electronic journal ', No. 12 ' non-user adaptation value interactive genetic algorithm based on entropy maximization criterion ' and a journal published in 2016, namely ' Applied intelligences ', No. 3 ' Predicting users ' preferences using neural networks and psychology models ' are combined with a user browsing behavior to express personalized requirements, so that the defect that the value type adaptation value is preferentially expressed is overcome. The method has the advantages of small calculation amount for directly acquiring the implicit knowledge, simplicity, strong subjectivity, larger information uncertainty and higher noise content. And secondly, implicit knowledge is indirectly acquired, namely the implicit knowledge is mined in the evolution process, and valuable preference information is extracted. Journal published in 2017, "zheng zhou university press (engineering edition)", interactive genetic algorithm based on possibility condition preference network and application thereof "fit user preferences using preference network, estimate individual fitness value and guide search; similarly, the journal published in 2017, automatic Software Engineering, 3 rd "architecture based on interactive optimization and machine learning application, the next release protocol" learns the user preferences using a neural network to estimate the adaptation value. The indirect acquisition of implicit knowledge is deep excavation of preference information, so that the searching capability of an algorithm is enhanced, but the modeling of an agent model is complex, estimation errors cannot be measured, and a large amount of adaptive value noise is still brought. The adaptive value estimation/expression strategy for deeply mining the user behavior mode is expected to improve the effect of implicit knowledge extraction or extract knowledge in a more complex scene, but from the overall optimization performance, the performance improvement of the algorithm obtained only by improving the implicit knowledge extraction strength is limited. For the 2 nd problem, it is mainly realized by an efficient evolution strategy to reduce user fatigue. The algorithm searching capability can be improved by adopting a large population scale, but the method needs to solve the problem of estimating the adaptation values of a large number of unexevated individuals. The agent model is adopted to execute the evolution operation, so that the workload of the user can be reduced, and the time overhead can be saved for the user. For example, in the publication of 2014, journal "electronic journal" at stage 8, "interactive genetic algorithm for selecting and evolving individuals based on elite sets" individual elite sets are constructed, and individual categories similar to the elite sets are selected and directly used for genetic operations, so that the burden of users is reduced. The idea is combined, and the performance of the algorithm can be obviously improved by adopting the evolution agent model under large-scale population.
In the algorithm, how to design a population evolution strategy is crucial, and since the proxy model generates error accumulation, the control error is a problem which is difficult to solve. If the set evolution method is applied to large-scale population cluster evolution, the search efficiency can be improved; meanwhile, if the estimated adaptive value of the existing large-scale population is re-evaluated by adopting the agent model, the accuracy of the adaptive value can be improved, and the two are fused to make up for the deficiency, so that the performance of the current interactive evolution optimization algorithm is expected to be improved. By looking up relevant documents, an interactive set evolution optimization method for predicting adaptive values by using a gray support vector regression does not exist at present. If a related efficient design system can be developed, the color matching design can be promoted, and the method has great instructive significance for other product designs.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the interactive evolution optimization method overcomes the defects of the existing adaptive value estimation technology, reduces adaptive value estimation errors, reduces user burden, enhances algorithm searching capability and improves evolution optimization quality.
The technical scheme of the invention is as follows:
firstly, a large-scale population expansion search space is adopted, and population clustering is carried out by taking a limited number of user evaluation individuals as a clustering center. Then, estimating an adaptive value of the non-clustering center individuals in the clusters according to the individual similarity of the integrated individual browsing behaviors, and further predicting the individual adaptive value by adopting a gray support vector regression model to form a set evolution individual. And finally, realizing a set evolution algorithm under an NSGA-II paradigm by adopting a set evolution strategy and self-adaptive cross mutation operation. In order to implement the present invention, 2 main problems of individual fitness value prediction and set evolution strategy design need to be solved.
(1) Adaptive value prediction for gray support vector machine
The exact number of individuals is adapted to the value f (x)k(t))，xk(t) is ∈ x (t) and is denoted as F0＝(f0(x1(t)),f0(x2(t)),…,f0(xN(t))) constituting an original sequence of individual fitness values. Then F0The 1-AGO sequence of (A) is F1＝(f1(x1(t)),f1(x2(t)),…,f1(xN(t)))，In order to make the influence weight of the original data on the objective function consistent in each feature dimension, the original sequence is firstly normalized:
the normalized data sequence is F0'＝(f0'(x1(t)),f0'(x2(t)),…,f0'(xN(t))). Then, F is established0' 1-AGO sequence F1＝(f1'(x1(t)),f1'(x2(t)),…,f1'(xN(t)))，The gray support vector model is then noted as:
in the formula: b is a bias term; ω is the weight. 1-AGO transformation term f1'(xk(t)) represents a non-linear mapping of the individual fitness value input space to the high-dimensional feature space. The unknown parameters ω and b in the equation are estimated from the training set in the high-dimensional feature space using the ε -insensitive loss function proposed by Vapnik:
due to minimizationThe minimum deviation of epsilon can be guaranteed, and the gray support vector model can be written as the following convex optimization problem:
in the formula:controlling the fitting precision of the model; and C is a regularization constant and controls the punishment degree of the samples exceeding the error. The above equation can be solved by Lagrange multiplier method:
a dual optimization model can be obtained:
in the formula < f1'(xi(t)),f1'(xj(t)) > represents the vector inner product. The invention selects a kernel function K (f)1'(xi(t)),f1'(xj(t)))＝exp(-υ||f1'(xi(t)),f1'(xj(t)) |) upsilon is more than 0, upsilon is a kernel parameter, and the kernel function is substituted into the formula to obtain αk,And b, the gray support vector regression model is:
in the formula,is for newly inputted individual adaptation valuef1'(xnew(t)) the gray support vector machine predictor.
And finally, restoring the predicted value to the original sequence scale to obtain:
clustering population individuals and predicting adaptive value to form individual subclass x1(t),x2(t),…,xNc(t) is a set of individuals, and it is natural to consider the subclass of individuals as evolved individuals (set evolved individuals) based on the idea of set evolution.
(2) Ensemble evolution strategy
By selecting a proper performance index, the implicit performance index optimization can be converted into the following general set decision variable optimization problem:
max F(X)＝(F1(X),F2(X),…,FId(X))
wherein,is a power set of a decision space S; x ═ X1(t),x2(t),…,xNc(t) } is a population formed by the evolved individuals; fd(X), d ═ 1,2, …, Id is the performance index for population X; id is the dimension of the post-transformation optimization problem and is much smaller than Nc. In combination with the implicit performance index characteristics, this section gives a new set individual comparison measure, as follows.
The j evolution individual of the t generation evolution population x (t) is recorded asWherein x isj(t) is mediumHeart individual, individual adaptation value ofM＝|xj(t) | is the evolved individual xj(t) number of individuals involved.
The diversity is characterized by the aid of the entropy of the similarity information of the evolved individuals:
in the above formula, the first and second carbon atoms are, wherein x isirR is 1,2, …, g is the r attributes that make up the individual,is xirThe attribute value of (2). x is the number ofj(t) center individual xj(t) Cluster center similarity μ (x) with other evolved individualsi(t),xj(t)) the larger, xj(t) the more similar the other evolved individuals, the poorer the population diversity, in this case, F2The smaller (X) is. Otherwise, F2The larger (X) is, the looser the evolution individuals of the population are, and the better the diversity of the population is.
The distribution was characterized using the following formula:
in the formula, d (x)j(t)) is xj(t) a minimum crowding distance of,d*average crowding of individuals for population evolutionThe distance between the first and second electrodes,
for the evolved individuals with the same sequence value, further adopting interval gray number gray scale to depict uncertainty:
the above formula will evolve the individual xjThe adaptive value of (t) is regarded as the interval gray number, xjThe uncertainty of the evaluation of (t) can be characterized by interval gray number gray scale, namely the smaller GF (X), the smaller the uncertainty of the evolution individual, and vice versa.
When x is1||sparx2When x is selected to be arg min { GF (x)1),GF(x2) The winning individual. By the method, the advantages and the disadvantages of any 2 set evolution individuals of the population can be compared.
(3) Adaptive set intersection and mutation probabilities
Aiming at the characteristics of set evolution, the subsection gives self-adaptive set intersection and mutation operators. For the cross probability in the same evolution individual, adopting the self-adaptive cross probability:
the above formula considers the relationship between the cross operation and the diversity and uncertainty change of the set evolution individuals in the evolution process. For enhancing the searching capability, the uncertainty of the evolved individuals is in direct proportion to the cross probability; in order to keep the dominant individual, the diversity of the evolved individual is in inverse proportion to the cross probability, the overall cross probability is gradually reduced along with the evolution, and the convergence of the algorithm is enhanced; and vice versa.
Mutation exercise of collective evolution individualsAdopts a single point mutation mode, and the self-adaptive mutation probability is Pm 1And Pm 2The part 2 comprises:
the above formula reflects the relationship between the diversity and convergence of the mutation operation and the evolved individuals. In order to preserve the dominant individual, the diversity and convergence should be inversely proportional to the variation probability, and vice versa.
In the formulaIs an evolved individual xj(t) individuals to be mutatedThe adaptive value of (a). The above expression reflects the relationship between mutation operation and the adaptive value of the evolved individual. I.e. selecting the evolved individual x of the variation operation to protect the dominant individual from being damagedjIn (t), the larger the adaptation value, the smaller the probability of mutation being performed.
Self-adaptive mutation probability P of evolved individualsmComprises the following steps:
the color individual coding method comprises the following steps: the RGB color mode obtains various colors through the change of three color attributes of red (R), green (G) and blue (B) and the superposition of the three color attributes, and the value range of each color attribute is 0-255. The RGB color individual chromosome adopts binary coding, the coding length is 24 bits, wherein the first 8 bits represent red attributes, the middle 8 bits represent green attributes, the last 8 bits represent blue attributes, and the corresponding binary coding range of each color attribute is 00000000-11111111. The optimization target is a certain preset target color, and through evolutionary optimization, a user obtains a color individual matched with the target color.
The invention has the advantages and positive effects that:
1. according to the method, the grey support vector machine is adopted to predict the non-user evaluation individual adaptive value, compared with other adaptive value estimation strategies, the estimation error is greatly reduced, and meanwhile, the user fatigue is reduced;
2. the invention adopts a new set evolution strategy to realize evolution, obviously improves the search efficiency and the optimization quality, and obtains good optimization effect when being applied to the color matching problem.
Drawings
FIG. 1 is a general flow diagram of the present invention;
FIG. 2 is a system interface diagram of the present invention;
FIG. 3 is a comparison of search times according to the present invention;
FIG. 4 is a comparison diagram of evolution algebra according to the present invention.
Detailed Description
The following further detailed description of the embodiments of the present invention is made with reference to fig. 1-4, and the following examples are illustrative only, not limiting, and are not intended to limit the scope of the present invention.
An interactive set evolution calculation flow based on the prediction adaptive value of a gray support vector regression is shown in fig. 1. The method comprises the following steps:
step 1, the interactive interface is composed of 3 parts, the part 1 is an evolution module positioned on the left half part of the interface, the module is presented to 12 individuals (color blocks) of a user, and an adaptive value input text box is arranged below each individual. Meanwhile, the system records the evaluation time of the user for each individual through a slider below the individual for calculating the uncertainty of the adaptive value. For comparison, a target color block is set on the periphery of the individual and a distance value from the current individual is displayed. The part 2 is an evolution parameter and information statistics module positioned at the upper right part of the interface, and a user can input RGB color values and set target colors by dragging a scroll bar before evolution. In addition, the module also displays statistical information such as evolution starting time, current evolution algebra, user evaluation individual number, total evaluation time and the like. The 3 rd part is a command button module positioned at the right lower part of the interface, a user clicks a 'start' button, and the system initializes and generates an initial population; after the user evaluates the individual, clicking a 'next' button, and carrying out evolution operations such as clustering, adaptive value estimation, set evolution and the like on the background by the system to generate a next generation of evolved population. The process loops until a termination condition is met and clicking the "end" button terminates the evolution.
The color individual coding method comprises the following steps: the RGB color mode obtains various colors through the change of three color attributes of red (R), green (G) and blue (B) and the superposition of the three color attributes, and the value range of each color attribute is 0-255. The RGB color individual chromosome adopts binary coding, the coding length is 24 bits, wherein the first 8 bits represent red attributes, the middle 8 bits represent green attributes, the last 8 bits represent blue attributes, and the corresponding binary coding range of each color attribute is 00000000-11111111. The optimization target is a certain preset target color, and through evolutionary optimization, a user obtains a color individual matched with the target color.
And 2. step 2.
First, a partial number of individuals x are selectedi(t), i is 1,2, … Nc as the individual clustering centers, and the adaptive value f (x) is evaluated by the useri(t)). Then, other non-central individuals x in the population are calculated one by one according to the formulao(t) the similarity between the individual and each central individual, clustering the individual according to the similarity, and marking as xi(t)＝{xi(t) }, i ∈ {1,2, … Nc }. Thus, the population x(t) is finally divided into Nc subclasses, denoted x1(t),x2(t),…,xNc(t) of (d). Then, a weighted average of the similarity to the center of each cluster is obtained in each individual subclass, and the non-center individual x is estimatedo(t) an adaptation value. To reduce adaptation value estimation errors, an adaptation value f (x) is estimated for non-central individualso(t)) adopting a gray support vector regression machine to predict to obtain a non-central individual final adaptive value for subsequent evolution
And 3. step 3.
And sequentially calculating diversity, distribution and uncertainty measures for the set evolution individuals.
Measure of diversity:
measure of distribution:
in the formula, d (x)j(t)) is xj(t) a minimum crowding distance of,d*the average crowding distance of individuals in population evolution,
uncertainty measure:
the above formula will evolve the individual xjThe adaptive value of (t) is regarded as the interval gray number, xjThe uncertainty of the evaluation of (t) can be characterized by interval gray number gray scale, namely the smaller GF (X), the smaller the uncertainty of the evolution individual, and vice versa.
When x is1||sparx2When x is selected to be arg min { GF (x)1),GF(x2) The winning individual. By the method, the advantages and the disadvantages of any 2 evolved individuals of the population can be compared.
And 4. step 4.
And (4) generating the temporary population with the same scale by adopting self-adaptive crossover and mutation operations on the ensemble evolved individuals. Self-adaptive cross probability:
the above formula considers the relationship between the cross operation and the diversity and uncertainty change of the set evolution individuals in the evolution process. For enhancing the searching capability, the uncertainty of the evolved individuals is in direct proportion to the cross probability; in order to keep the dominant individual, the diversity of the evolved individual is in inverse proportion to the cross probability, the overall cross probability is gradually reduced along with the evolution, and the convergence of the algorithm is enhanced; and vice versa.
The mutation operation of the set evolution individuals adopts a single point mutation mode, and the probability of the self-adaption mutation is determined byAndthe part constitutes:
the above formula reflects the relationship between the mutation operation and the diversity and distribution of the evolved individuals. In order to preserve the dominant individual, the diversity and distribution should be inversely proportional to the probability of variation, and vice versa.
In the formulaIs an evolved individual xj(t) individuals to be mutatedThe adaptive value of (a). To protect dominant individuals from being destroyed, evolved individuals x of mutation operation are selectedjIn (t), the larger the adaptation value, the smaller the probability of mutation being performed.
Self-adaptive mutation probability P of evolved individualsmComprises the following steps:
and 5. step 5.
Merging the parent population and the temporary population, sequencing the merged population, and selecting the first N individuals as child populations; and finally, dividing the sub-generation population into Nc units at equal intervals, randomly selecting 1 individual from each unit, and recommending the Nc individuals to the user.
Comparison of the present invention with the present Algorithm
Two related Algorithms, namely 'Interactive Genetic Algorithm based on probability condition preference Network' (PCPN-IGA) 'and' Interactive Genetic Algorithm based on Elite Set selection evolution individual '(IGA-SES)' are used as comparison Algorithms to verify the effectiveness of the method in the aspects of searching efficiency, optimizing quality, reducing user fatigue and the like.
The experimental results of the three methods are shown in fig. 4. From fig. 4, it can be seen that the present invention is clearly superior to the comparison algorithm, which is embodied as:
a. in the aspect of searching time, the method consumes the least time, and improves the searching efficiency. The invention adopts a set evolution strategy and an NSGA-II algorithm engine, so that the population distribution is more uniform, the diversity is better, and the algorithm searching efficiency is improved. The convergence of the algorithm is further enhanced by adopting the self-adaptive intersection and mutation operation. Although it takes longer to set the evolving individuals, the overall search time is still shorter than for the comparative algorithm.
b. In the aspect of evolution algebra, the invention has the least evolution algebra, which not only improves the search efficiency, but also reduces the fatigue of users. The method adopts the individual similarity based on the user browsing behavior to estimate the individual adaptive value, and adopts the gray support vector regression machine to further predict the adaptive value, so that the adaptive value precision is improved, the evolution direction is more in line with the preference of people, and the algorithm convergence is accelerated.
The table below shows the search result mean and the number of completely matched users statistics for the three methods. The data in the table were subjected to a Mann-Whitney U test with a significance level of 0.05. As can be seen from the table, the number of the invention is the least in the two indexes of evaluating the number of individuals and searching the number of individuals, and the difference with the comparison method is obvious. The number of the present invention is the most significant, although not significantly different from the comparative method, in the best solution. This shows that the present invention can obtain the most satisfactory solutions with relatively least number of evaluations (including evolution algebra). This reflects the most efficient optimization of the present invention. The small number of search individuals is caused by the small number of evolutionary algebras. Because the comparison method is also large-scale population evolution with the same population scale, the more evolution generations, the more search individuals, so the more search individuals of the comparison method can not indicate that the search capability is better than the invention. The number of the users obtaining the complete matching solution is 3, the number of IGA-SES is 2, and the number of PCPN-IGA is only 1. This means that most users end up obtaining an approximate solution to the complex implicit performance indicator optimization problem of color matching. But the number of users of the invention which can obtain a complete matching solution is the largest, which reflects the best optimization performance of the invention. From the above analysis, it can be seen that the search performance of the present invention is the best.
a. IGA-SES-PCPN-IGA b. inventive IGA-SES-PCPN-IGA
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications, equivalent variations and modifications made to the above embodiment according to the technical spirit of the present invention still fall within the scope of the technical solution of the present invention.
Claims (4)
1. An interactive set evolution method for predicting adaptive values based on a gray support vector regression is characterized by comprising the following steps: adopting a large-scale population expansion search space, and carrying out population clustering by taking a limited number of user evaluation individuals as a clustering center; then, estimating adaptive values of non-clustering center individuals in the clusters according to the individual similarity of the integrated individual browsing behaviors, and further predicting the individual adaptive values by adopting a gray support vector regression model to form a set evolution individual; finally, a set evolution strategy and self-adaptive cross variation operation are adopted to realize set evolution;
(1) adaptive value prediction for gray support vector machine
Individual adaptation value f for new inputs1'(xnew(t)) the gray support vector machine predictor is:
(2) ensemble evolution strategy
Measure of diversity:
measure of distribution:
uncertainty measure:
(3) adaptive set intersection and mutation probabilities
Self-adaptive cross probability:
self-adaptive mutation probability:
2. the interactive set evolution method based on the grey support vector regression prediction adaptive value according to claim 1, characterized in that:
3. the interactive set evolution calculation method based on the prediction adaptive value of the gray support vector regression machine as claimed in claim 1, wherein: d (x)j(t)) is xj(t) a minimum crowding distance of,d*the average crowding distance of individuals in population evolution,
4. the interactive set evolution method based on the grey support vector regression prediction adaptive value according to claim 1, characterized in that: adaptive mutation probability PmByAnd2, the part is formed by the following steps,
in the formulaIs an evolved individual xj(t) individuals to be mutatedThe adaptive value of (a).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810603480.9A CN108960486B (en) | 2018-06-12 | 2018-06-12 | Interactive set evolution method for predicting adaptive value based on gray support vector regression |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810603480.9A CN108960486B (en) | 2018-06-12 | 2018-06-12 | Interactive set evolution method for predicting adaptive value based on gray support vector regression |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108960486A true CN108960486A (en) | 2018-12-07 |
| CN108960486B CN108960486B (en) | 2021-10-29 |
Family
ID=64488560
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810603480.9A Expired - Fee Related CN108960486B (en) | 2018-06-12 | 2018-06-12 | Interactive set evolution method for predicting adaptive value based on gray support vector regression |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108960486B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110162704A (en) * | 2019-05-21 | 2019-08-23 | 西安电子科技大学 | More scale key user extracting methods based on multiple-factor inheritance algorithm |
| CN110276375A (en) * | 2019-05-14 | 2019-09-24 | 嘉兴职业技术学院 | A kind of identification and processing method of crowd's dynamic clustering information |
| CN110739045A (en) * | 2019-10-14 | 2020-01-31 | 郑州航空工业管理学院 | Interactive evolution optimization method for personalized recipe design |
| CN115204064A (en) * | 2022-09-19 | 2022-10-18 | 中国石油大学(华东) | Gas injection huff-puff parameter optimization method and system in shale oil exploitation |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102270302A (en) * | 2011-07-20 | 2011-12-07 | 北京航空航天大学 | Grey support vector machine-based multi-stress accelerated life testing forecasting method |
| US20140245207A1 (en) * | 2013-02-25 | 2014-08-28 | Christian D. Poulin | Interfaces for predictive models |
| CN104596767A (en) * | 2015-01-13 | 2015-05-06 | 北京工业大学 | Method for diagnosing and predicating rolling bearing based on grey support vector machine |
| CN104749219A (en) * | 2013-07-30 | 2015-07-01 | 中国标准化研究院 | A honey detection method by support vector machine classifier parameter selection based on particle swarm optimization |
| CN106296444A (en) * | 2016-07-27 | 2017-01-04 | 国网山东省电力公司经济技术研究院 | A kind of annual whole society based on particle cluster algorithm and support vector machine power predicating method |
| CN106845136A (en) * | 2017-02-21 | 2017-06-13 | 上海大学 | A kind of needle-valve body crush and grind accuracy prediction method based on SVMs |
| CN107609681A (en) * | 2017-08-22 | 2018-01-19 | 西安建筑科技大学 | A kind of more metal multiple target ore-proportioning methods based on APSO algorithm |
| US20180082192A1 (en) * | 2016-09-21 | 2018-03-22 | Scianta Analytics, LLC | Cognitive modeling apparatus including multiple knowledge node and supervisory node devices |
| CN107993723A (en) * | 2017-12-19 | 2018-05-04 | 苏州大学 | A kind of warfarin dose prediction modeling method based on integrated evolutionary learning |
| CN108022014A (en) * | 2017-12-05 | 2018-05-11 | 国网山东省电力公司济南供电公司 | A kind of Load Prediction In Power Systems method and system |
-
2018
- 2018-06-12 CN CN201810603480.9A patent/CN108960486B/en not_active Expired - Fee Related
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102270302A (en) * | 2011-07-20 | 2011-12-07 | 北京航空航天大学 | Grey support vector machine-based multi-stress accelerated life testing forecasting method |
| US20140245207A1 (en) * | 2013-02-25 | 2014-08-28 | Christian D. Poulin | Interfaces for predictive models |
| CN104749219A (en) * | 2013-07-30 | 2015-07-01 | 中国标准化研究院 | A honey detection method by support vector machine classifier parameter selection based on particle swarm optimization |
| CN104596767A (en) * | 2015-01-13 | 2015-05-06 | 北京工业大学 | Method for diagnosing and predicating rolling bearing based on grey support vector machine |
| CN106296444A (en) * | 2016-07-27 | 2017-01-04 | 国网山东省电力公司经济技术研究院 | A kind of annual whole society based on particle cluster algorithm and support vector machine power predicating method |
| US20180082192A1 (en) * | 2016-09-21 | 2018-03-22 | Scianta Analytics, LLC | Cognitive modeling apparatus including multiple knowledge node and supervisory node devices |
| CN106845136A (en) * | 2017-02-21 | 2017-06-13 | 上海大学 | A kind of needle-valve body crush and grind accuracy prediction method based on SVMs |
| CN107609681A (en) * | 2017-08-22 | 2018-01-19 | 西安建筑科技大学 | A kind of more metal multiple target ore-proportioning methods based on APSO algorithm |
| CN108022014A (en) * | 2017-12-05 | 2018-05-11 | 国网山东省电力公司济南供电公司 | A kind of Load Prediction In Power Systems method and system |
| CN107993723A (en) * | 2017-12-19 | 2018-05-04 | 苏州大学 | A kind of warfarin dose prediction modeling method based on integrated evolutionary learning |
Non-Patent Citations (1)
| Title |
|---|
| 赵玲 等: ""基于过程支持向量回归机的油田开发指标预测模型"", 《数学的实践与认识》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110276375A (en) * | 2019-05-14 | 2019-09-24 | 嘉兴职业技术学院 | A kind of identification and processing method of crowd's dynamic clustering information |
| CN110276375B (en) * | 2019-05-14 | 2021-08-20 | 嘉兴职业技术学院 | Method for identifying and processing crowd dynamic clustering information |
| CN110162704A (en) * | 2019-05-21 | 2019-08-23 | 西安电子科技大学 | More scale key user extracting methods based on multiple-factor inheritance algorithm |
| CN110162704B (en) * | 2019-05-21 | 2022-06-10 | 西安电子科技大学 | Multi-scale key user extraction method based on multi-factor genetic algorithm |
| CN110739045A (en) * | 2019-10-14 | 2020-01-31 | 郑州航空工业管理学院 | Interactive evolution optimization method for personalized recipe design |
| CN110739045B (en) * | 2019-10-14 | 2023-05-30 | 郑州航空工业管理学院 | Interactive evolution optimization method for personalized recipe design |
| CN115204064A (en) * | 2022-09-19 | 2022-10-18 | 中国石油大学(华东) | Gas injection huff-puff parameter optimization method and system in shale oil exploitation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108960486B (en) | 2021-10-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108960486B (en) | Interactive set evolution method for predicting adaptive value based on gray support vector regression | |
| CN107465664B (en) | Intrusion detection method based on parallel multi-worker bee colony algorithm and support vector machine | |
| CN110781406B (en) | Social network user multi-attribute inference method based on variational automatic encoder | |
| CN109359140A (en) | A kind of sequence of recommendation method and device based on adaptive attention | |
| CN113792754B (en) | Converter transformer DGA online monitoring data processing method for firstly removing abnormal state and then repairing | |
| CN107203590B (en) | Personalized movie recommendation method based on improved NSGA-II | |
| Ding et al. | Product color emotional design considering color layout | |
| CN111061959B (en) | Group intelligent software task recommendation method based on developer characteristics | |
| US20120203722A1 (en) | Apparatus and method for processing information, recording medium and computer program | |
| CN108804577B (en) | Method for estimating interest degree of information tag | |
| CN108256678A (en) | A kind of method that double-deck attention network using sorting measure carries out customer relationship prediction | |
| CN110956277A (en) | Interactive iterative modeling system and method | |
| CN116628344A (en) | Knowledge graph-based intelligent recommendation method and system for owner information service | |
| CN110059251B (en) | Collaborative filtering recommendation method based on multi-relation implicit feedback confidence | |
| CN112817563A (en) | Target attribute configuration information determination method, computer device, and storage medium | |
| CN108737429B (en) | Network intrusion detection method | |
| CN112749211A (en) | Novel tea yield prediction method based on electric power big data | |
| CN111984842A (en) | Bank client data processing method and device | |
| Van Breda et al. | A feature representation learning method for temporal datasets | |
| CN115201394B (en) | Multi-component transformer oil chromatography online monitoring method and related device | |
| CN110705756A (en) | Electric power energy consumption optimization control method based on input convex neural network | |
| CN113360772B (en) | Interpretable recommendation model training method and device | |
| CN110955811B (en) | Power data classification method and system based on naive Bayes algorithm | |
| Clement et al. | Beyond explaining: XAI-based Adaptive Learning with SHAP Clustering for Energy Consumption Prediction | |
| CN113987261A (en) | Video recommendation method and system based on dynamic trust perception |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20211029 |