KHM data clustering method and system, computer equipment and storage medium

Through the fuzzy logic white bone chicken optimization algorithm (FCOOT) combined with the K-harmonized mean function, the problem of K-mean clustering easily falls into local optimality and noise sensitivity is solved, and the stability and global convergence of data clustering are improved, and the more High search accuracy and convergence speed.

CN119004155APending Publication Date: 2024-11-22BEIJING PEONY ELECTRONICS GROUP CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410293049.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-03-14
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The K-mean clustering algorithm is easily trapped into local optimal solutions and is sensitive to noise. The intelligent optimization algorithm is also easily trapped into local optimal when processing high-dimensional data clustering, and the search accuracy and convergence speed are insufficient.

Method used

The fuzzy logic fuzzy optimization algorithm (FCOOT) is used to use multiple initial clustering centroids as individual positions of fuzzy chickens, and the fuzzy logic mechanism is used to update the individual positions of leaders and followers, and the K-harmony mean function is used to calculate the global maximum of the data set. Optimal solution, dynamically adjust the convergence factor and leader population proportion to improve cluster stability and global convergence.

Benefits of technology

The stability and global convergence of data clustering are improved, the local optimization and noise sensitivity problems of K-mean clustering are overcome, and the search accuracy and convergence speed are enhanced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119004155A_ABST
    Figure CN119004155A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of big data analysis, and discloses a KHM data clustering method and system, computer equipment and a storage medium, and the method comprises the steps: based on K random clustering centroids selected from a target data set, executing a KHM clustering iteration process, generating an initial clustering centroid, and carrying out the clustering of K random clustering centroids; taking the initial clustering centroid as an individual initial position in a fuzzy logic white top chicken optimization algorithm; determining a leader population proportion by using a fuzzy logic mechanism, and updating a convergence factor of a fuzzy logic white top chicken optimization algorithm to obtain an updated position of a leader individual and an updated position of a follower individual; calculating the individual fitness of the target data set and determining a current global optimal solution by using a KHM function according to the updated position of the leader individual and the updated position of the follower individual; and when the maximum number of iterations is reached, determining the current globally optimal solution as a target globally optimal solution. According to the method, the stability and global convergence of data clustering can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Background Art

[0002] Data clustering is an important method in the field of pattern recognition and data mining, and is particularly critical in the efficient storage of big data in the Internet of Vehicles. As an unsupervised learning method, the goal of cluster analysis is to cluster data objects in a data set into non-overlapping subsets, so that the data in the same cluster has high similarity, and the data in different clusters has low similarity. K-means clustering (KM) is a simple and efficient clustering algorithm, but it has obvious shortcomings such as high sensitivity to initial centroids, easy to converge prematurely, and fall into local optimality. In order to overcome the shortcomings of K-means clustering, K-harmonic means clustering (KHM) uses the average harmonic mean to obtain new cluster centers, which overcomes the sensitivity of initial centroids to a certain extent, but the model is still sensitive to noise and still has local optimality.

[0003] In recent years, data clustering with intelligent optimization algorithms has been widely studied. With the powerful global search performance advantage of intelligent algorithms, clustering divisions have higher similarity and the global optimal solution is found. For example, the harmony search algorithm, gravitational search algorithm, bacterial algorithm, particle swarm optimization clustering algorithm, hybrid intelligent optimization algorithm, mutant firefly algorithm and fireworks algorithm have all been applied to clustering problems. However, due to the performance differences in their own population evolution mechanisms, these algorithms are still prone to falling into local optimality when searching for cluster centroids. The white-bone chicken optimization algorithm COOT is a heuristic algorithm that simulates white-bone chickens preying in different movement modes on the water surface. It has few key parameters, a simple model and is easy to implement. In the optimization of complex functions, it shows better search accuracy and convergence speed than particle swarm, gravitational search and gray wolf algorithm. In dealing with high-dimensional optimization problems of data clustering, the search accuracy of COOT is highly related to the adjustment of convergence factors and search adaptability. Otherwise, it is still easy to have the shortcomings of poor search accuracy and slow convergence.

[0004] Therefore, it is urgent to provide a technical solution to solve the above problems. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a KHM data clustering method, system, computer device and storage medium.

[0006] In a first aspect, the present invention provides a KHM data clustering method, and the technical solution of the method is as follows:

[0007] Based on K random cluster centroids selected from the target data set, a KHM clustering iteration process is performed to generate multiple initial cluster centroids, and each initial cluster centroid is used as the initial position of the white-bone chicken individual in the fuzzy logic white-bone chicken optimization algorithm; wherein the white-bone chicken individuals include: leader individuals and follower individuals;

[0008] Using a fuzzy logic mechanism, determining the leader population proportion of the fuzzy logic coot optimization algorithm, and updating the convergence factor of the fuzzy logic coot optimization algorithm to obtain the updated position of the leader individual and the updated position of the follower individual;

[0009] Using the KHM function and according to the updated position of the leader individual and the updated position of the follower individual, the individual fitness of the target data set is calculated and the current global optimal solution of the clustering result of the target data set is determined;

[0010] When the current iteration number of the fuzzy logic coot optimization algorithm reaches the maximum iteration number, the current global optimal solution is determined as the target global optimal solution of the clustering result of the target data set.

[0011] The beneficial effects of a KHM data clustering method of the present invention are as follows:

[0012] Aiming at the shortcoming that KHM clustering is prone to fall into local optimum, the method of the present invention proposes a data clustering method of optimizing K harmonic means by using fuzzy logic coot algorithm FCOOT, which can improve the stability and global convergence of data clustering.

[0013] Based on the above solution, a KHM data clustering method of the present invention can also be improved as follows.

[0014] In an optional manner, the method further includes:

[0015] When the current number of iterations of the fuzzy logic coot optimization algorithm does not reach the maximum number of iterations, the step of using the fuzzy logic mechanism to determine the proportion of the leader population of the fuzzy logic coot optimization algorithm is returned to be executed.

[0016] In an optional manner, the convergence factor includes: a first convergence factor; and the step of updating the convergence factor of the fuzzy logic coot optimization algorithm to obtain an updated position of the leader individual includes:

[0017] The first convergence factor of the fuzzy logic coot optimization algorithm is updated by using a fuzzy logic mechanism, and the updated position of the leader individual is calculated according to the initial position of the leader individual and the updated first convergence factor.

[0018] In an optional manner, the convergence factor further includes: a second convergence factor; and the step of updating the convergence factor of the fuzzy logic coot optimization algorithm to obtain an updated position of the follower individual includes:

[0019] Using a fuzzy logic mechanism, updating a second convergence factor of the fuzzy logic coot optimization algorithm;

[0020] If the follower individual is actively updated, the updated position of the follower individual is calculated based on the initial position of the follower individual and the updated second convergence factor;

[0021] If the follower individual does not actively update, the update position of the follower individual is calculated based on the leader individual followed by the follower individual and the initial position of the follower individual.

[0022] In a second aspect, the present invention provides a KHM data clustering system, the technical solution of which is as follows:

[0023] It includes: processing module, updating module, computing module and clustering module;

[0024] The processing module is used to: perform a KHM clustering iteration process based on K random clustering centroids selected from the target data set, generate multiple initial clustering centroids, and use each initial clustering centroid as the initial position of the white bone chicken individual in the fuzzy logic white bone chicken optimization algorithm; wherein the white bone chicken individual includes: a leader individual and a follower individual;

[0025] The updating module is used to: determine the leader population proportion of the fuzzy logic white-bone chicken optimization algorithm by using a fuzzy logic mechanism, and update the convergence factor of the fuzzy logic white-bone chicken optimization algorithm to obtain the updated position of the leader individual and the updated position of the follower individual;

[0026] The calculation module is used to: calculate the individual fitness of the target data set and determine the current global optimal solution of the clustering result of the target data set by using the KHM function and according to the update position of the leader individual and the update position of the follower individual;

[0027] The clustering module is used to: when the current iteration number of the fuzzy logic white-bone chicken optimization algorithm reaches the maximum iteration number, determine the current global optimal solution as the target global optimal solution of the clustering result of the target data set.

[0028] The beneficial effects of a KHM data clustering system of the present invention are as follows:

[0029] Aiming at the shortcoming that KHM clustering is prone to fall into local optimum, the system of the present invention proposes a data clustering method of optimizing K harmonic mean by fuzzy logic white bone chicken algorithm FCOOT, which can improve the stability and global convergence of data clustering.

[0030] Based on the above solution, the KHM data clustering system of the present invention can also be improved as follows.

[0031] In an optional manner, the method further includes: an iteration module; the iteration module is used to:

[0032] When the current iteration number of the fuzzy logic coot optimization algorithm does not reach the maximum iteration number, return to call the update module.

[0033] In an optional manner, the convergence factor includes: a first convergence factor; the step of updating the convergence factor of the fuzzy logic white-bone chicken optimization algorithm in the updating module to obtain an updated position of the leader individual includes:

[0034] The first convergence factor of the fuzzy logic coot optimization algorithm is updated by using a fuzzy logic mechanism, and the updated position of the leader individual is calculated according to the initial position of the leader individual and the updated first convergence factor.

[0035] In an optional manner, the convergence factor further includes: a second convergence factor; the step of updating the convergence factor of the fuzzy logic white-bone chicken optimization algorithm in the updating module to obtain the updated position of the follower individual includes:

[0036] Using a fuzzy logic mechanism, updating a second convergence factor of the fuzzy logic coot optimization algorithm;

[0037] If the follower individual is actively updated, the updated position of the follower individual is calculated based on the initial position of the follower individual and the updated second convergence factor;

[0038] If the follower individual does not actively update, the update position of the follower individual is calculated based on the leader individual followed by the follower individual and the initial position of the follower individual.

[0039] In a third aspect, a technical solution of a computer device of the present invention is as follows:

[0040] The invention comprises a memory, a processor and a program stored in the memory and running on the processor. When the processor executes the program, the steps of the KHM data clustering method of the present invention are implemented.

[0041] In a fourth aspect, the present invention provides a computer-readable storage medium having the following technical solution:

[0042] The computer-readable storage medium stores instructions, and when the computer-readable storage medium reads the instructions, the computer-readable storage medium executes the steps of the KHM data clustering method of the present invention.

[0043] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are only used to illustrate the embodiments and are not to be considered as limiting the present invention. In addition, the same reference symbols are used to represent the same components throughout the accompanying drawings. In the accompanying drawings:

[0045] Figure 1 A schematic diagram of a flow chart of an embodiment of a KHM data clustering method of the present invention;

[0046] Figure 2 It is a schematic diagram of the membership function of the input quantity;

[0047] Figure 3 It is a schematic diagram of the output membership function;

[0048] Figure 4 It is a trend chart of the change of the output variable leader proportion;

[0049] Figure 5 It is a schematic diagram of the fitness error membership function;

[0050] Figure 6 Schematic diagram of the membership function of the convergence factor;

[0051] Figure 7 It is the trend diagram of the output variable convergence factor;

[0052] Figure 8 A schematic diagram of the structure of an embodiment of a KHM data clustering system of the present invention;

[0053] Fig. 9 The figure is a schematic diagram of the structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0055] Figure 1 FIG. 2 is a flow chart showing an embodiment of a KHM data clustering method provided by the present invention. Figure 1 As shown, the following steps are included:

[0056] S1. Based on K random cluster centroids selected from the target data set, perform the KHM clustering iteration process to generate multiple initial cluster centroids, and use each initial cluster centroid as the initial position of the white-bone chicken individual in the fuzzy logic white-bone chicken optimization algorithm.

[0057] 1) The target data set can be a data set of any data type, such as business order data, commodity transaction data, etc., without any restrictions.

[0058] 2) The value of K is a positive integer and does not exceed the population size in the target dataset.

[0059] 3) KHM clustering is K-harmonic mean clustering, with a maximum number of iterations KT max .

[0060] 4) The initial cluster centroid is: the cluster centroid in the target data set obtained by executing the KHM clustering iteration process.

[0061] 5) Fuzzy logic white-bone chicken optimization algorithm is referred to as fuzzy logic COOT algorithm, that is, FCOOT algorithm. White-bone chicken individuals include: leader individuals and follower individuals, and the number is determined according to actual conditions.

[0062] It should be noted that before executing the data clustering method of this embodiment, it is necessary to initialize the number of target data sets (population size N), the maximum number of iterations of K-harmonic mean clustering KT max and the maximum number of iterations T of the FCOOT algorithm max .

[0063] S2. Using a fuzzy logic mechanism, determine the leader population proportion of the fuzzy logic coot optimization algorithm, and update the convergence factor of the fuzzy logic coot optimization algorithm to obtain the updated position of the leader individual and the updated position of the follower individual.

[0064] 1) In S2, a leader individual can also be randomly selected, and the KHM algorithm (X, C) is used to calculate the individual fitness and determine the global optimal solution gBest.

[0065] 2) The number of convergence factors is two by default, namely: the first convergence factor and the second convergence factor. The first convergence factor is for the leader individual, and the second convergence factor is for the follower individual.

[0066] S3. Using the KHM function, and based on the updated positions of the leader individual and the follower individual, calculate the individual fitness of the target data set and determine the current global optimal solution of the clustering result of the target data set.

[0067] 1) In the process of calculating individual fitness, it is necessary to select and retain the best, that is, for a follower individual with better fitness, the role of the follower individual can be exchanged with its corresponding leader individual.

[0068] 2) The current global optimal solution is: the global optimal solution under the current number of iterations of the FCOOT algorithm.

[0069] S4. When the current iteration number of the fuzzy logic coot optimization algorithm reaches the maximum iteration number, the current global optimal solution is determined as the target global optimal solution of the clustering result of the target data set.

[0070] 1) The maximum number of iterations of the FCOOT algorithm can be set according to actual conditions and is not limited here.

[0071] 2) The target global optimal solution is the global optimal solution finally optimized in this example, that is, the data clustering result of the target data set.

[0072] In an optional manner, the method further includes:

[0073] When the current number of iterations of the fuzzy logic coot optimization algorithm does not reach the maximum number of iterations, the step of using the fuzzy logic mechanism to determine the proportion of the leader population of the fuzzy logic coot optimization algorithm is returned to be executed.

[0074] Specifically, when the current number of iterations of the fuzzy logic coot optimization algorithm does not reach the maximum number of iterations of the FCOOT algorithm, return to execute S2 until the current number of iterations of the fuzzy logic coot optimization algorithm reaches the maximum number of iterations, and the current global optimal solution is determined as the target global optimal solution of the clustering result of the target data set.

[0075] In an optional manner, the step of updating the convergence factor of the fuzzy logic coot optimization algorithm to obtain an updated position of the leader individual includes:

[0076] The first convergence factor of the fuzzy logic coot optimization algorithm is updated by using a fuzzy logic mechanism, and the updated position of the leader individual is calculated according to the initial position of the leader individual and the updated first convergence factor.

[0077] Among them, according to the initial position of the leader individual and the updated first convergence factor, the formula involved in the process of calculating the updated position of the leader individual is:

[0078]

[0079] LeadPos(i) is the position of leader i, gBest is the current global optimal solution, R3, R4∈[0,1] are random values, and B is the first convergence factor, defined as: B=2-t / T max .

[0080] In an optional manner, the convergence factor further includes: a second convergence factor; and the step of updating the convergence factor of the fuzzy logic coot optimization algorithm to obtain an updated position of the follower individual includes:

[0081] The second convergence factor of the fuzzy logic coot optimization algorithm is updated using a fuzzy logic mechanism.

[0082] If the follower individual is actively updated, the updated position of the follower individual is calculated according to the initial position of the follower individual and the updated second convergence factor.

[0083] Among them, according to the initial position of the follower individual and the updated second convergence factor, the formula involved in the process of calculating the updated position of the follower individual is:

[0084] FollowPos(i)=FollowPos(i)+A×R1×(Q-FollowPos(i)); FollowPos(i) on the left side of the equation is the updated position of the ith follower, and FollowPos(i) on the right side of the equation is the initial position of the ith follower. R1∈[0,1] is a random value, and A is the second convergence factor, defined as: A=1-t / T max .

[0085] Or, FollowPos(i)=0.5×(FollowPos(i-1)+FollowPos(i)); FollowPos(i-1) is the position of the previous follower individual adjacent to the i-th follower.

[0086] If the follower individual does not actively update, the update position of the follower individual is calculated based on the leader individual followed by the follower individual and the initial position of the follower individual.

[0087] In order to better illustrate the technical solution of this embodiment, the following content is used to describe it in detail, specifically:

[0088] (1) K-Harmonic Means Clustering (KHM Algorithm)

[0089] The KHM algorithm uses the harmonic mean as a distance metric to perform clustering. The KHM algorithm clustering objective function can be defined as:

[0090]

[0091] Among them, x i For the target data set X = (x1, x2, ..., x n ), n is the total number of data objects in the target data set, C = (c1, c2, ..., c k ) is the cluster centroid set, k is the number of centroids, c j is the cluster centroid j.

[0092] The KHM clustering process is:

[0093] Step 1: Randomly select k cluster centroids in the D-dimensional search space;

[0094] Step 2: Calculate the objective function value according to formula (1);

[0095] Step 3: For each data point object x i , calculate x i In each cluster centroid c j The membership degree within is calculated as:

[0096]

[0097] At the same time, calculate the data object x i In the cluster centroid c j The weight value within:

[0098]

[0099] Step 4: For each cluster centroid, recalculate the data object x according to its membership and weight function i The cluster centroid of is:

[0100]

[0101] Step 5: Repeat steps 2 to 4 until the cluster centroid does not change or the maximum number of iterations is reached;

[0102] Step 6: Transform the data vector x i Reassign to the group with the maximum membership m(c j \x i )’s cluster centroid j.

[0103] (2) Standard COOT algorithm

[0104] There are three behavior modes in the white-bone chicken population: chain movement, leader movement, and follower movement. The entire population is divided into leaders and followers based on their fitness. Leaders have a stronger ability to approach food sources, and followers have two modes of position update: active update and passive update. During active update, followers can update their positions in random and chain movement modes to enrich the diversity of search, and are not dependent on leaders at this time. During passive update, followers are pulled by the leader and approach it.

[0105] The initial population position of white-bone chickens is generated as follows:

[0106] CootPos(i)=r(1,d)×(ub-lb)+lb (5)

[0107] Where r() is a random value in [0,1], d is the search dimension, [lb,ub] is the search range of the individual, and CootPos(i) is the position of individual i.

[0108] 1) Follower location updates

[0109] Followers can update their positions actively or passively according to probability. Active updates can be divided into two modes: random movement and chain movement. In random movement, a random position of an individual is first generated by the following formula:

[0110] Q=r(1,d)×(ub-lb)+lb (6)

[0111] Under the influence of the random position, the white-bone chicken individual will move in the direction of the random position. At this time, the position update method is defined as:

[0112] FollowPos(i)=FollowPos(i)+A×R1×(Q-FollowPos(i)) (7)

[0113] Where FollowPos(i) is the position of follower i, R1∈[0,1] is a random value, and A is the convergence factor, which is defined as:

[0114] A=1-t / T max (8)

[0115] Where, T max is the maximum number of iterations.

[0116] In chain motion, the white-bone chicken individuals will perform chain motion between two adjacent individuals. At this time, the position update method is defined as:

[0117] FollowPos(i)=0.5×(FollowPos(i-1)+FollowPos(i)) (9)

[0118] Where FollowPos(i-1) is the position of the adjacent previous follower individual.

[0119] During passive updates, followers are led by the leader. The selected leader is:

[0120] k=1+(i)mod(N L ) (10)

[0121] In the formula, k is the leader individual index, i is the follower index, and N L is the number of leader individuals. At this time, the passive position update method of the follower is:

[0122] FollowPos(i)=LeadPos(k)+2×R2×cos(2Rπ)×(LeadPos(k)-FollowPos(i))(11)

[0123] Where LeadPos(k) is the position of leader k, and R, R2∈[0,1] are random values.

[0124] 2) Leader position update

[0125] The leader position is updated as follows:

[0126]

[0127] Where LeadPos(i) is the position of leader i, gBest is the optimal solution of the current population, R3, R4∈[0,1] are random values, and B is the convergence factor, defined as: B=2-t / T max (13)

[0128] (3) Fuzzy Logic COOT Algorithm (FCOOT Algorithm)

[0129] 1) Fuzzy logic leader population adjustment

[0130] For the COOT algorithm, the position update of the leader population is driven by the global optimal solution and plays the role of global search. However, during the algorithm iteration process, the proportion of the leader population is fixed. This determines that the global search capability remains unchanged during the entire algorithm optimization process. However, for the swarm intelligence algorithm, the early iteration should focus on large-scale global search, that is, a large proportion of leaders should be used to search the solution space extensively to accelerate the convergence of the algorithm; the late iteration should increase the proportion of local development, that is, a small proportion of leaders (with a corresponding increase in the proportion of followers) should be used to improve the mining accuracy of local areas and improve the overall convergence accuracy of the algorithm. In order to dynamically set the population proportion of the leader of the white-bone crested chicken according to the process of algorithm iteration, this paper introduces a leader population proportion adjustment mechanism based on fuzzy logic, and establishes a fuzzy reasoning system to perform nonlinear adaptive updates on the proportion of leaders in the entire population.

[0131] The input variables of the fuzzy inference system are set as the iteration process and population diversity. The iteration process variable is defined as:

[0132] iter=t / T max (14)

[0133] Among them, t is the current iteration number of the algorithm, T max is the maximum number of iterations of the algorithm. That is, the iterative process variable iter is the ratio of the two.

[0134] The population diversity variable is defined as:

[0135]

[0136] Among them, N is the total number of individuals in the population, D is the position dimension, xi,j(t) is the position of individual i in dimension j at iteration t, and xbest,j(t) is the position of the optimal solution in dimension j at this time. Population diversity div can be regarded as the cosine value of the angle between the current white-bone chicken individual and the optimal individual, which can reflect the similarity between individuals. The larger the div, the stronger the similarity, indicating that the population diversity is low; vice versa.

[0137] The iterative process variable and the population diversity variable are used as the two inputs of the fuzzy inference system, and the population diversity variable is normalized to make its value fall between [0,1]. Figure 2 The two membership functions shown. The iterative process variables are divided into Early, Early_Medium, Medium, Medium_Late and Late by five membership functions. The population diversity variables are divided into Low, Medium and High by three membership functions.

[0138] The triangular membership function of the output variable leader population ratio of the fuzzy inference system is divided into Very_Small, Small, Medium, Big and Very_Big. The five memberships are used to represent the size of the output leader population ratio, such as Figure 3 shown.

[0139] In order to select the proportion of the leader population in combination with the two system input variables, the corresponding IF-THEN fuzzy rules need to be designed. The fuzzy rules of the fuzzy reasoning system are shown in Table 1. In the process of designing fuzzy rules, it is necessary to consider: first, in the early stage of iteration, the search range should be expanded as much as possible, and a larger proportion of the leader population should be given to enhance the leading role of the leader to achieve global search; second, as the iteration proceeds, the algorithm should gradually turn to small-scale local development. At this time, the leader size should be reduced, and the accuracy of local development should be improved with a larger follower proportion; at the same time, if the population diversity is gradually lost during the iteration process and the population tends to converge prematurely, the leader proportion should also be increased in the local development process to enrich the population diversity. When the diversity is good, the proportion of the leader population in the global search can be reduced accordingly.

[0140] Table 1:

[0141]

[0142] Combined with the IF-THEN fuzzy rules shown in Table 1, the output leader ratio factor implemented by the fuzzy inference system can be obtained after defuzzification, such as Figure 4 As shown in the figure, under the influence of the two input variables, iterative process and population diversity, the leader's population proportion can be dynamically adjusted, so that the algorithm can search for the optimal solution with stronger search and development performance in the early and late stages of the iteration.

[0143] 2) Fuzzy logic convergence factor adjustment

[0144] In the process of random movement of followers and position update of leaders, the role of convergence factors A and B is to balance the proportion of search and development of the two population categories, and determine whether the white-bone crested chicken expands the search range or shrinks the encirclement to prey. However, from the model of the COOT algorithm, it can be seen that the two parameters are simply linearly decreased according to the number of iterations, which is obviously not in line with the predation law of the population. For the second convergence factor A, the larger the value, the greater the influence of the random individual, and the smaller the influence of the previous iteration, which is conducive to search. For the first convergence factor B, the larger the value, the greater the influence of the previous iteration, and the smaller the influence of the global optimal solution, which is conducive to development. In order to nonlinearly adjust the two convergence factors at different stages of iteration, fuzzy logic is introduced to enable the convergence factors to be adaptively adjusted according to the iterative process of the algorithm and the individual fitness error.

[0145] Calculate the convergence factor output and set the input variables of the fuzzy inference system as the iteration process and fitness error. The reason for introducing fitness error as one of the input factors of the fuzzy system is that during the iteration process, if the fitness of the individual is greatly different from that of the optimal solution, the proportion of global search and local development of the population should be optimized accordingly to control the degree of progression of the individual to the optimal solution. The iteration process calculation method is the same as formula (14). The fitness error is defined as:

[0146]

[0147] Among them, N is the total number of individuals in the population, fit(Xi) is the fitness of individual i, and fitmin is the current optimal fitness. It can be seen that the fitness error variable can be used to measure the difference between the fitness of each individual and the optimal individual.

[0148] The iteration process variable and the fitness error variable are used as two inputs of the fuzzy inference system to calculate the value of the convergence factor. Figure 2 (a) in the figure. The membership function of the fitness error variable is as follows: Figure 5 As shown in Figure 2, the three membership functions are divided into low error (Low), medium error (Medium) and high error (High).

[0149] Since the original value ranges of the convergence factors A and B are [0,1] and [0,2] respectively, when used as system output, the triangular membership function can be divided into five stages: small, small, medium, large and very large, such as Figure 6 shown.

[0150] Two convergence factors A and B are selected in combination with the two system input variables, and the IF-THEN fuzzy rules shown in Table 2 are designed.

[0151] Table 2:

[0152]

[0153] Combined with the fuzzy rules shown in Table 2, the output convergence factor of the fuzzy inference system can be obtained after defuzzification, such as Figure 7 As shown in Figure 2, under the action of the iterative process of the two input variables and the fitness error, the convergence factor can be dynamically adjusted to achieve an adaptive balance of convergence.

[0154] 3) Clustering method steps based on fuzzy logic COOT to optimize KHM

[0155] Step 1: Initialize the algorithm parameters, including the population size N, K-harmonic mean clustering maximum iteration number KT max 、FCOOT maximum iteration number T max ;

[0156] Step 2: Randomly select K cluster centroids and perform the KHM clustering iteration process to generate the initial clustering solution;

[0157] Step 3: The cluster solution generated in step 2 (selecting the cluster centroid) is used as the initial population structure of FCOOT, that is, the cluster centroid is encoded as the individual location information of the white-bone chicken;

[0158] Step 4: Use the fuzzy logic mechanism to determine the proportion of the leader population, randomly select the leader individual, use the harmonic mean function KHM(X,C) to calculate the individual fitness, and determine the global optimal solution gBest;

[0159] Step 5: Determine the roles of the individuals in the population. For the leader individual, use fuzzy logic to update the second convergence factor B and update the leader position according to formula (12);

[0160] Step 6: For the follower individual, use fuzzy logic to update the first convergence factor A. If the follower actively updates, update the position according to formula (7) or formula (9); otherwise, determine the leader individual it follows according to formula (10), and then update the position according to formula (11);

[0161] Step 7: Calculate the fitness of the individuals in the population and select the best ones to retain. Specifically, for followers with better fitness, their leader and follower roles are swapped;

[0162] Step 8: Determine whether the maximum number of iterations T has been reached max If it is not reached, return to step 4 to execute; otherwise, output the global optimal solution at this time and the algorithm terminates.

[0163] The technical solution of this embodiment aims at the deficiency that K harmonic mean data clustering is prone to fall into local optimum, and proposes a data clustering method for optimizing K harmonic mean by fuzzy logic white-bone chicken algorithm FCOOT. Combining the rapid local development of K harmonic mean clustering and the global exploration capability of white-bone chicken optimization algorithm, the initial clustering solution generated by K harmonic mean is input as the initial population structure of white-bone chickens for iterative optimization. At the same time, in order to further improve the search accuracy of COOT, fuzzy logic is designed to adaptively adjust the convergence factor and leader population proportion of COOT to balance the search and development capabilities of the algorithm. The fitness of individuals in the population is evaluated by using the clustering harmonic mean, and the clustering results are iteratively optimized in combination with the heuristic search of intelligent algorithms. The clustering performance of the technical solution of this embodiment has been effectively improved in terms of stability and global convergence.

[0164] Figure 8 FIG. 2 is a schematic diagram showing a structure of an embodiment of a KHM data clustering system 200 provided by the present invention. Figure 8 As shown, the system 200 includes: a processing module 210, an updating module 220, a computing module 230 and a clustering module 240;

[0165] The processing module 210 is used to: perform a KHM clustering iteration process based on K random cluster centroids selected from the target data set, generate multiple initial cluster centroids, and use each initial cluster centroid as the initial position of the white bone chicken individual in the fuzzy logic white bone chicken optimization algorithm; wherein the white bone chicken individual includes: a leader individual and a follower individual;

[0166] The updating module 220 is used to: determine the leader population proportion of the fuzzy logic coot optimization algorithm by using a fuzzy logic mechanism, and update the convergence factor of the fuzzy logic coot optimization algorithm to obtain the updated position of the leader individual and the updated position of the follower individual;

[0167] The calculation module 230 is used to: calculate the individual fitness of the target data set and determine the current global optimal solution of the clustering result of the target data set by using the KHM function and according to the update position of the leader individual and the update position of the follower individual;

[0168] The clustering module 240 is used to: when the current iteration number of the fuzzy logic coot optimization algorithm reaches the maximum iteration number, determine the current global optimal solution as the target global optimal solution of the clustering result of the target data set.

[0169] In an optional manner, the method further includes: an iteration module; the iteration module is used to:

[0170] When the current iteration number of the fuzzy logic coot optimization algorithm does not reach the maximum iteration number, the update module 220 is called back.

[0171] In an optional manner, the convergence factor includes: a first convergence factor; the step of updating the convergence factor of the fuzzy logic coot optimization algorithm in the updating module 220 to obtain an updated position of the leader individual includes:

[0172] The first convergence factor of the fuzzy logic coot optimization algorithm is updated by using a fuzzy logic mechanism, and the updated position of the leader individual is calculated according to the initial position of the leader individual and the updated first convergence factor.

[0173] In an optional manner, the convergence factor further includes: a second convergence factor; the step of updating the convergence factor of the fuzzy logic white-bone chicken optimization algorithm in the updating module 220 to obtain the updated position of the follower individual includes:

[0174] Using a fuzzy logic mechanism, updating a second convergence factor of the fuzzy logic coot optimization algorithm;

[0175] If the follower individual is actively updated, the updated position of the follower individual is calculated based on the initial position of the follower individual and the updated second convergence factor;

[0176] If the follower individual does not actively update, the update position of the follower individual is calculated based on the leader individual followed by the follower individual and the initial position of the follower individual.

[0177] The technical solution of this embodiment aims at the deficiency that KHM clustering is prone to fall into local optimum, and proposes a data clustering method of optimizing K harmonic mean by fuzzy logic white bone chicken algorithm FCOOT, which can improve the stability and global convergence of data clustering.

[0178] For the above parameters and steps of each module implementing corresponding functions in the KHM data clustering system 200 of this embodiment, reference may be made to the parameters and steps in the above embodiment of the KHM data clustering method, which will not be described in detail here.

[0179] like Fig. 9As shown, a computer device 300 according to an embodiment of the present invention includes a processor 320, the processor 320 is coupled to a memory 310, and the memory 310 stores at least one computer program 330, and the at least one computer program 330 is loaded and executed by the processor 320, so that the computer device 300 implements any one of the above-mentioned KHM data clustering methods, specifically:

[0180] The computer device 300 may have relatively large differences due to different configurations or performances, and may include one or more processors 320 (Central Processing Units, CPU) and one or more memories 310, wherein at least one computer program 330 is stored in the one or more memories 310, and the at least one computer program 330 is loaded and executed by the one or more processors 320, so that the computer device 300 implements any KHM data clustering method provided in the above embodiments. Of course, the computer device 300 may also have components such as a wired or wireless network interface, a keyboard, and an input and output interface for input and output, and the computer device 300 may also include other components for implementing device functions, which will not be described in detail here.

[0181] A computer-readable storage medium according to an embodiment of the present invention stores at least one computer program, and the at least one computer program is loaded and executed by a processor to enable a computer to implement any one of the above-mentioned KHM data clustering methods.

[0182] Optionally, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, and the like.

[0183] In an exemplary embodiment, a computer program product or a computer program is also provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs any of the above-mentioned KHM data clustering methods.

[0184] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects and represent the definition of a specific order or sequence. The order of use of similar objects can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than the order shown or described.

[0185] Those skilled in the art know that the present invention can be implemented as a system, method or computer program product. Therefore, the present disclosure can be specifically implemented in the following forms, namely: it can be complete hardware, it can be complete software (including firmware, resident software, microcode, etc.), or it can be a combination of hardware and software, which is generally referred to as "circuit", "module" or "system" herein. In addition, in some embodiments, the present invention can also be implemented in the form of a computer program product in one or more computer-readable media, and the computer-readable medium contains computer-readable program code.

[0186] Any combination of one or more computer-readable media can be used. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0187] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A KHM data clustering method, It is characterized in that include: Based on K random cluster centroids selected from the target data set, a KHM clustering iteration process is performed to generate multiple initial cluster centroids, and each initial cluster centroid is used as the initial position of the white-bone chicken individual in the fuzzy logic white-bone chicken optimization algorithm; wherein the white-bone chicken individuals include: leader individuals and follower individuals; Using a fuzzy logic mechanism, determining the leader population proportion of the fuzzy logic coot optimization algorithm, and updating the convergence factor of the fuzzy logic coot optimization algorithm to obtain the updated position of the leader individual and the updated position of the follower individual; Using the KHM function and according to the updated position of the leader individual and the updated position of the follower individual, the individual fitness of the target data set is calculated and the current global optimal solution of the clustering result of the target data set is determined; When the current iteration number of the fuzzy logic coot optimization algorithm reaches the maximum iteration number, the current global optimal solution is determined as the target global optimal solution of the clustering result of the target data set.

2. The KHM data clustering method according to claim 1, It is characterized in that Also includes: When the current number of iterations of the fuzzy logic coot optimization algorithm does not reach the maximum number of iterations, the step of using the fuzzy logic mechanism to determine the proportion of the leader population of the fuzzy logic coot optimization algorithm is returned to be executed.

3. The KHM data clustering method according to claim 1, It is characterized in that The convergence factor includes: a first convergence factor; and the step of updating the convergence factor of the fuzzy logic coot optimization algorithm to obtain an updated position of the leader individual includes: The first convergence factor of the fuzzy logic coot optimization algorithm is updated by using a fuzzy logic mechanism, and the updated position of the leader individual is calculated according to the initial position of the leader individual and the updated first convergence factor.

4. The KHM data clustering method according to claim 3, It is characterized in that The convergence factor also includes: a second convergence factor; the step of updating the convergence factor of the fuzzy logic white bone chicken optimization algorithm to obtain an updated position of the follower individual includes: Using a fuzzy logic mechanism, updating a second convergence factor of the fuzzy logic coot optimization algorithm; If the follower individual is actively updated, the updated position of the follower individual is calculated based on the initial position of the follower individual and the updated second convergence factor; If the follower individual does not actively update, the update position of the follower individual is calculated based on the leader individual followed by the follower individual and the initial position of the follower individual.

5. A KHM data clustering system, It is characterized in that include: Processing module, updating module, computing module and clustering module; The processing module is used to: perform a KHM clustering iteration process based on K random clustering centroids selected from the target data set, generate multiple initial clustering centroids, and use each initial clustering centroid as the initial position of the white bone chicken individual in the fuzzy logic white bone chicken optimization algorithm; wherein the white bone chicken individual includes: a leader individual and a follower individual; The updating module is used to: determine the leader population proportion of the fuzzy logic white-bone chicken optimization algorithm by using a fuzzy logic mechanism, and update the convergence factor of the fuzzy logic white-bone chicken optimization algorithm to obtain the updated position of the leader individual and the updated position of the follower individual; The calculation module is used to: calculate the individual fitness of the target data set and determine the current global optimal solution of the clustering result of the target data set by using the KHM function and according to the update position of the leader individual and the update position of the follower individual; The clustering module is used to: when the current iteration number of the fuzzy logic white-bone chicken optimization algorithm reaches the maximum iteration number, determine the current global optimal solution as the target global optimal solution of the clustering result of the target data set.

6. The KHM data clustering system according to claim 5, It is characterized in that Also includes: Iteration module; the iteration module is used for: When the current iteration number of the fuzzy logic coot optimization algorithm does not reach the maximum iteration number, return to call the update module.

7. The KHM data clustering system according to claim 5, It is characterized in that The convergence factor includes: a first convergence factor; the step of updating the convergence factor of the fuzzy logic coot optimization algorithm in the updating module to obtain the updated position of the leader individual includes: The first convergence factor of the fuzzy logic coot optimization algorithm is updated by using a fuzzy logic mechanism, and the updated position of the leader individual is calculated according to the initial position of the leader individual and the updated first convergence factor.

8. The KHM data clustering system according to claim 7, It is characterized in that The convergence factor also includes: a second convergence factor; the step of updating the convergence factor of the fuzzy logic coot optimization algorithm in the updating module to obtain the updated position of the follower individual includes: Using a fuzzy logic mechanism, updating a second convergence factor of the fuzzy logic coot optimization algorithm; If the follower individual is actively updated, the updated position of the follower individual is calculated based on the initial position of the follower individual and the updated second convergence factor; If the follower individual does not actively update, the update position of the follower individual is calculated based on the leader individual followed by the follower individual and the initial position of the follower individual.

9. A computer device, It is characterized in that The computer device includes a processor, the processor is coupled to a memory, the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor so that the computer device implements the KHM data clustering method according to any one of claims 1 to 4.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores at least one computer program, and the at least one computer program is loaded and executed by the processor, so that the computer-readable storage medium implements the KHM data clustering method according to any one of claims 1 to 4.