The method of overlapping community detection based on fuzzy clustering is sensitive to the initial-ization of community centers, which easily traps in local optima and leads to node misclassification. This paper prop...
详细信息
The method of overlapping community detection based on fuzzy clustering is sensitive to the initial-ization of community centers, which easily traps in local optima and leads to node misclassification. This paper proposes an evolutionary multiobjective algorithm based on similarity matrix and node correction to detect overlapping communities to solve the above problems. Firstly, the algorithm determines a similarity community for each node by setting the similarity threshold. Then, the central nodes are found more accurately through the similarity distribution of the similarity communities. Sec-ondly, under the framework of the evolutionary multiobjective algorithm, the similarity communities of the central nodes are used as the initial communities to obtain the nonoverlapping communities. In addition, the algorithm proposes a correction strategy for the noncentral nodes based on the similarity communities. The correction strategy obtains the adjacent nodes of each node's similarity community. It then uses each adjacent node's community to correct the nonoverlapping community. Finally, the algorithm adjusts the noncentral nodes' correction strategy. This correction strategy corrects the overlapping nodes according to the number of each overlapping node's labels. It takes the separation operation to further correct overlapping nodes to obtain the corrected overlapping communities. This paper uses seventeen real networks and a variety of synthetic networks with different parameters to verify the proposed algorithm's effectiveness. And the proposed algorithm achieves higher accuracy of community detection in most networks than four state-of-the-art overlapping community detection algorithms. (C) 2022 Elsevier B.V. All rights reserved.
作者：
Wan, YutingZhong, YanfeiMa, AilongWuhan Univ
State Key Lab Informat Engn Surveying Mapping & R Wuhan 430079 Hubei Peoples R China Wuhan Univ
Collaborat Innovat Ctr Geospatial Technol Wuhan 430079 Hubei Peoples R China
Clustering of remote sensing imagery is a tough task due to the particular and complex structure of remote sensing images and the shortage of known information. In this paper, we propose a fully automatic spectral-spa...
详细信息
Clustering of remote sensing imagery is a tough task due to the particular and complex structure of remote sensing images and the shortage of known information. In this paper, we propose a fully automatic spectral-spatial fuzzy clustering method using an adaptive multiobjective memetic algorithm (AMOMA) for multispectral remote sensing imagery. This approach is made up of two automatic layers: an automatic determination layer and an automatic clustering layer. The first layer seeks the optimal number of clusters through a self-adaptive differential evolution algorithm. The second layer then takes advantage of the AMOMA for spectral-spatial clustering using the optimal number of clusters. The knee point from the Pareto front is then selected through the angle-based method in every generation, and we then compare the knee points between generations to output the final optimal solution. The effectiveness of the proposed method is verified by the experimental results obtained with three remote sensing data sets.
This letter puts forward a new algorithm, ensemble strategy multiobjective fuzzy clustering method (ESMOFCM). To fully combine the gray information and spatial information of neighbor pixels, a new dividing fluctuant ...
详细信息
This letter puts forward a new algorithm, ensemble strategy multiobjective fuzzy clustering method (ESMOFCM). To fully combine the gray information and spatial information of neighbor pixels, a new dividing fluctuant parameter is proposed for producing a difference image. Then, we use a frame based on multiobjective fuzzy clustering to alleviate the contradiction between removing noise and preserving details in images. Ensemble strategy is adopted to integrate all Pareto optimal solutions. The experimental results show that the proposed algorithm is superior to comparison algorithms.
Rather than a whole Pareto optimal front(POF), which demands too many points, the decision maker (DM) may only be interested in a partial region, called the region of interest(ROI). In this paper, we propose a systema...
详细信息
Rather than a whole Pareto optimal front(POF), which demands too many points, the decision maker (DM) may only be interested in a partial region, called the region of interest(ROI). In this paper, we propose a systematic method to incorporate the DM's preference information into a decomposition-based evolutionary multiobjective optimization algorithm (MOEA/D-HP). Different from most existing decomposition-based preference algorithms, MOEA/D-HP guides the population to converge to the preference region by generating hierarchical reference points in the preference region, and forms some hierarchical solutions for comparison and selection by the DM. In addition, the novel reference vectors generating method of MOEA/D-HP makes the final solutions no longer uniformly distributed in the ROI, instead the closer to the preference point, the denser the obtained solution. Extensive experiments on a variety of benchmark problems with 2 to 15 objectives fully demonstrate the effectiveness of our method in obtaining preferred solutions in the ROI.
In researching multi-objective evolutionary algorithms (MOEAs), the decision-maker (DM) may not need the entire Pareto optimal front searched and may only be interested in the region of interest (ROI). Most existing p...
详细信息
In researching multi-objective evolutionary algorithms (MOEAs), the decision-maker (DM) may not need the entire Pareto optimal front searched and may only be interested in the region of interest (ROI). Most existing preference-based research focuses on determining the location of the ROI and controlling its size. Those research typically ignores the preference information provided by the DM when solving problems. Since the convergence region and diversity of the population are determined according to the DM's preference information, so we propose a preference-based MOEA that uses a normal distribution (ND) to generate a weight vector, called MOEA/D-ND. The generation of the weight vector uses the DM's preference information to guide the solution to converge to the vicinity of the preference information. Because the randomness of the normal distribution can lead to a loss of diversity, an angle-based niche selection strategy is adopted. This strategy prevents the population from falling into a local optimum during the search process. Although the reference vector generated by MOEA/D-ND using the normal distribution will make the final solution set no longer uniformly distributed in the ROI, still, the closer region to the reference point, the more solution sets are obtained. The experimental results show that this algorithm has advantages in various benchmark problems with 2 to 15 goals.
The use of evolutionary algorithms (EAs) to solve problems with multiple objectives (known as Vector Optimization Problems (VOPs)) has attracted much attention recently. Being population based approaches, EAs offer a ...
详细信息
The use of evolutionary algorithms (EAs) to solve problems with multiple objectives (known as Vector Optimization Problems (VOPs)) has attracted much attention recently. Being population based approaches, EAs offer a means to find a group of pareto-optimal solutions in a single run. Differential Evolution (DE) is an EA that was developed to handle optimization problems over continuous domains. The objective of this paper is to introduce a novel Pareto Differential Evolution (PDE) algorithm to solve VOPs. The solutions provided by the proposed algorithm for five standard test problems, is competitive to nine known evolutionary multiobjective algorithms for solving VOPs.
暂无评论