Many classifiers and methods are proposed to deal with letter recognition problem. Among them, clustering is a widely used method. But only one time for clustering is not adequately. Here, we adopt data preprocessing ...Many classifiers and methods are proposed to deal with letter recognition problem. Among them, clustering is a widely used method. But only one time for clustering is not adequately. Here, we adopt data preprocessing and a re kernel clustering method to tackle the letter recognition problem. In order to validate effectiveness and efficiency of proposed method, we introduce re kernel clustering into Kernel Nearest Neighbor classification(KNN), Radial Basis Function Neural Network(RBFNN), and Support Vector Machine(SVM). Furthermore, we compare the difference between re kernel clustering and one time kernel clustering which is denoted as kernel clustering for short. Experimental results validate that re kernel clustering forms fewer and more feasible kernels and attain higher classification accuracy.展开更多
Multiple kernel clustering is an unsupervised data analysis method that has been used in various scenarios where data is easy to be collected but hard to be labeled.However,multiple kernel clustering for incomplete da...Multiple kernel clustering is an unsupervised data analysis method that has been used in various scenarios where data is easy to be collected but hard to be labeled.However,multiple kernel clustering for incomplete data is a critical yet challenging task.Although the existing absent multiple kernel clustering methods have achieved remarkable performance on this task,they may fail when data has a high value-missing rate,and they may easily fall into a local optimum.To address these problems,in this paper,we propose an absent multiple kernel clustering(AMKC)method on incomplete data.The AMKC method rst clusters the initialized incomplete data.Then,it constructs a new multiple-kernel-based data space,referred to as K-space,from multiple sources to learn kernel combination coefcients.Finally,it seamlessly integrates an incomplete-kernel-imputation objective,a multiple-kernel-learning objective,and a kernel-clustering objective in order to achieve absent multiple kernel clustering.The three stages in this process are carried out simultaneously until the convergence condition is met.Experiments on six datasets with various characteristics demonstrate that the kernel imputation and clustering performance of the proposed method is signicantly better than state-of-the-art competitors.Meanwhile,the proposed method gains fast convergence speed.展开更多
Multiple kernel clustering based on local kernel alignment has achieved outstanding clustering performance by applying local kernel alignment on each sample.However,we observe that most of existing works usually assum...Multiple kernel clustering based on local kernel alignment has achieved outstanding clustering performance by applying local kernel alignment on each sample.However,we observe that most of existing works usually assume that each local kernel alignment has the equal contribution to clustering performance,while local kernel alignment on different sample actually has different contribution to clustering performance.Therefore this assumption could have a negative effective on clustering performance.To solve this issue,we design a multiple kernel clustering algorithm based on self-weighted local kernel alignment,which can learn a proper weight to clustering performance for each local kernel alignment.Specifically,we introduce a new optimization variable-weight-to denote the contribution of each local kernel alignment to clustering performance,and then,weight,kernel combination coefficients and cluster membership are alternately optimized under kernel alignment frame.In addition,we develop a three-step alternate iterative optimization algorithm to address the resultant optimization problem.Broad experiments on five benchmark data sets have been put into effect to evaluate the clustering performance of the proposed algorithm.The experimental results distinctly demonstrate that the proposed algorithm outperforms the typical multiple kernel clustering algorithms,which illustrates the effectiveness of the proposed algorithm.展开更多
A new algorithm named kernel bisecting k-means and sample removal(KBK-SR) is proposed as sampling preprocessing for support vector machine(SVM) training to improve the efficiency.The proposed algorithm tends to quickl...A new algorithm named kernel bisecting k-means and sample removal(KBK-SR) is proposed as sampling preprocessing for support vector machine(SVM) training to improve the efficiency.The proposed algorithm tends to quickly produce balanced clusters of similar sizes in the kernel feature space,which makes it efficient and effective for reducing training samples.Theoretical analysis and experimental results on three UCI real data benchmarks both show that,with very short sampling time,the proposed algorithm dramatically accelerates SVM sampling and training while maintaining high test accuracy.展开更多
The fuzzy C-means clustering algorithm(FCM) to the fuzzy kernel C-means clustering algorithm(FKCM) to effectively perform cluster analysis on the diversiform structures are extended, such as non-hyperspherical data, d...The fuzzy C-means clustering algorithm(FCM) to the fuzzy kernel C-means clustering algorithm(FKCM) to effectively perform cluster analysis on the diversiform structures are extended, such as non-hyperspherical data, data with noise, data with mixture of heterogeneous cluster prototypes, asymmetric data, etc. Based on the Mercer kernel, FKCM clustering algorithm is derived from FCM algorithm united with kernel method. The results of experiments with the synthetic and real data show that the FKCM clustering algorithm is universality and can effectively unsupervised analyze datasets with variform structures in contrast to FCM algorithm. It is can be imagined that kernel-based clustering algorithm is one of important research direction of fuzzy clustering analysis.展开更多
For the kernel K-mean cluster method is run in an implicit feature space, the initial and iterative cluster centers cannot be defined explicitly. Against the deficiency of the initial cluster centers selected in the o...For the kernel K-mean cluster method is run in an implicit feature space, the initial and iterative cluster centers cannot be defined explicitly. Against the deficiency of the initial cluster centers selected in the original space discretionarily in the existing methods, this paper proposes a new method for ensuring the clustering center that virtual clustering centers are defined in the feature space by the original classification as the initial cluster centers and the iteration clustering centers are ensured by the further virtual classification. The improved method is used for fault diagnosis of roller bearing that achieves a good cluster and diagnosis result, which demonstrates the effectiveness of the proposed method.展开更多
To deal with the nonlinear separable problem, the generalized noise clustering （GNC） algorithm is extended to a kernel generalized noise clustering （KGNC） model. Different from the fuzzy c-means （FCM） model and ...To deal with the nonlinear separable problem, the generalized noise clustering （GNC） algorithm is extended to a kernel generalized noise clustering （KGNC） model. Different from the fuzzy c-means （FCM） model and the GNC model which are based on Euclidean distance, the presented model is based on kernel-induced distance by using kernel method. By kernel method the input data are nonlinearly and implicitly mapped into a high-dimensional feature space, where the nonlinear pattern appears linear and the GNC algorithm is performed. It is unnecessary to calculate in high-dimensional feature space because the kernel function can do it just in input space. The effectiveness of the proposed algorithm is verified by experiments on three data sets. It is concluded that the KGNC algorithm has better clustering accuracy than FCM and GNC in clustering data sets containing noisy data.展开更多
Over the last fifteen years, face recognition has become a popular area of research in image analysis and one of the most successful applications of machine learning and understanding. To enhance the classification ra...Over the last fifteen years, face recognition has become a popular area of research in image analysis and one of the most successful applications of machine learning and understanding. To enhance the classification rate of the image recognition, several techniques are introduced, modified and combined. The suggested model extracts the features using Fourier-Gabor filter, selects the best features using signal to noise ratio, deletes or modifies anomalous images using fuzzy c-mean clustering, uses kernel least square and optimizes it by using wild dog pack optimization. To compare the suggested method with the previous methods, four datasets are used. The results indicate that the suggested methods without fuzzy clustering and with fuzzy clustering outperform state- of-art methods for all datasets.展开更多
A novel model of fuzzy clustering using kernel methods is proposed. This model is called kernel modified possibilistic c-means （KMPCM） model. The proposed model is an extension of the modified possibilistic c-means ...A novel model of fuzzy clustering using kernel methods is proposed. This model is called kernel modified possibilistic c-means （KMPCM） model. The proposed model is an extension of the modified possibilistic c-means （MPCM） algorithm by using kernel methods. Different from MPCM and fuzzy c-means （FCM） model which are based on Euclidean distance, the proposed model is based on kernel-induced distance. Furthermore, with kernel methods the input data can be mapped implicitly into a high-dimensional feature space where the nonlinear pattern now appears linear. It is unnecessary to do calculation in the high-dimensional feature space because the kernel function can do it. Numerical experiments show that KMPCM outperforms FCM and MPCM.展开更多
Fuzzy c-means(FCM) clustering algorithm is sensitive to noise points and outlier data, and the possibilistic fuzzy c-means(PFCM) clustering algorithm overcomes the problem well, but PFCM clustering algorithm has some ...Fuzzy c-means(FCM) clustering algorithm is sensitive to noise points and outlier data, and the possibilistic fuzzy c-means(PFCM) clustering algorithm overcomes the problem well, but PFCM clustering algorithm has some problems: it is still sensitive to initial clustering centers and the clustering results are not good when the tested datasets with noise are very unequal. An improved kernel possibilistic fuzzy c-means algorithm based on invasive weed optimization(IWO-KPFCM) is proposed in this paper. This algorithm first uses invasive weed optimization(IWO) algorithm to seek the optimal solution as the initial clustering centers, and introduces kernel method to make the input data from the sample space map into the high-dimensional feature space. Then, the sample variance is introduced in the objection function to measure the compact degree of data. Finally, the improved algorithm is used to cluster data. The simulation results of the University of California-Irvine(UCI) data sets and artificial data sets show that the proposed algorithm has stronger ability to resist noise, higher cluster accuracy and faster convergence speed than the PFCM algorithm.展开更多
道路点云数据的障碍物检测技术在智能交通系统和自动驾驶中至关重要.传统的基于密度的空间聚类(DensityBased Spatial Clustering of Applications with Noise,DBSCAN)算法在处理高维或不同密度区域数据时,由于距离度量低效、参数组合...道路点云数据的障碍物检测技术在智能交通系统和自动驾驶中至关重要.传统的基于密度的空间聚类(DensityBased Spatial Clustering of Applications with Noise,DBSCAN)算法在处理高维或不同密度区域数据时,由于距离度量低效、参数组合确定困难导致聚类效果欠佳,因此,提出了一种基于改进DBSCAN的道路障碍物点云聚类方法 .首先,在确定Eps领域时利用孤立核函数来改进传统的距离度量方式,提高了DBSCAN聚类对不同密度区域的适应性和准确性.其次,针对猎豹优化算法(Cheetah Optimizer,CO)在信息共享和迭代更新方面的不足,提出了一种基于及时更新机制与兼容度量策略的CO优化算法(Timely Updating Mechanisms and Compatible Metric Strategies for CO Algorithms,TCCO),通过实时更新操作确保每次迭代的优秀信息得到及时沟通共享,并在全局更新时基于非支配排序与拥挤距离优化淘汰机制,平衡全局搜索和局部开发能力,提高了收敛速度和收敛精度.最后,利用孤立度量改进Eps领域,并利用TCCO优化DBSCAN聚类,自适应确定参数,提高了聚类精度和效率.在八个UCI数据集上进行测试,仿真结果表明,提出的TCCO-DBSCAN算法与CO-DBSCAN,SSA-DBSCAN,DBSCAN,KMC方法相比,F-Measure,ARI,NMI指标均有明显提升,且聚类精度更优.通过激光雷达点云数据障碍物聚类的实验验证,证明TCCO-DBSCAN能够有效地适应点云数据密度变化,获得更好的道路障碍物聚类效果,为辅助驾驶中障碍物检测提供支持.展开更多
基金Supported by the National Science Foundation(No.IIS-9988642)the Multidisciplinary Research Program
文摘Many classifiers and methods are proposed to deal with letter recognition problem. Among them, clustering is a widely used method. But only one time for clustering is not adequately. Here, we adopt data preprocessing and a re kernel clustering method to tackle the letter recognition problem. In order to validate effectiveness and efficiency of proposed method, we introduce re kernel clustering into Kernel Nearest Neighbor classification(KNN), Radial Basis Function Neural Network(RBFNN), and Support Vector Machine(SVM). Furthermore, we compare the difference between re kernel clustering and one time kernel clustering which is denoted as kernel clustering for short. Experimental results validate that re kernel clustering forms fewer and more feasible kernels and attain higher classification accuracy.
基金funded by National Natural Science Foundation of China under Grant Nos.61972057 and U1836208Hunan Provincial Natural Science Foundation of China under Grant No.2019JJ50655+3 种基金Scientic Research Foundation of Hunan Provincial Education Department of China under Grant No.18B160Open Fund of Hunan Key Laboratory of Smart Roadway and Cooperative Vehicle Infrastructure Systems(Changsha University of Science and Technology)under Grant No.kfj180402the“Double First-class”International Cooperation and Development Scientic Research Project of Changsha University of Science and Technology under Grant No.2018IC25the Researchers Supporting Project No.(RSP-2020/102)King Saud University,Riyadh,Saudi Arabia.
文摘Multiple kernel clustering is an unsupervised data analysis method that has been used in various scenarios where data is easy to be collected but hard to be labeled.However,multiple kernel clustering for incomplete data is a critical yet challenging task.Although the existing absent multiple kernel clustering methods have achieved remarkable performance on this task,they may fail when data has a high value-missing rate,and they may easily fall into a local optimum.To address these problems,in this paper,we propose an absent multiple kernel clustering(AMKC)method on incomplete data.The AMKC method rst clusters the initialized incomplete data.Then,it constructs a new multiple-kernel-based data space,referred to as K-space,from multiple sources to learn kernel combination coefcients.Finally,it seamlessly integrates an incomplete-kernel-imputation objective,a multiple-kernel-learning objective,and a kernel-clustering objective in order to achieve absent multiple kernel clustering.The three stages in this process are carried out simultaneously until the convergence condition is met.Experiments on six datasets with various characteristics demonstrate that the kernel imputation and clustering performance of the proposed method is signicantly better than state-of-the-art competitors.Meanwhile,the proposed method gains fast convergence speed.
基金This work was supported by the National Key R&D Program of China(No.2018YFB1003203)National Natural Science Foundation of China(Nos.61672528,61773392,61772561)+1 种基金Educational Commission of Hu Nan Province,China(No.14B193)the Key Research&Development Plan of Hunan Province(No.2018NK2012).
文摘Multiple kernel clustering based on local kernel alignment has achieved outstanding clustering performance by applying local kernel alignment on each sample.However,we observe that most of existing works usually assume that each local kernel alignment has the equal contribution to clustering performance,while local kernel alignment on different sample actually has different contribution to clustering performance.Therefore this assumption could have a negative effective on clustering performance.To solve this issue,we design a multiple kernel clustering algorithm based on self-weighted local kernel alignment,which can learn a proper weight to clustering performance for each local kernel alignment.Specifically,we introduce a new optimization variable-weight-to denote the contribution of each local kernel alignment to clustering performance,and then,weight,kernel combination coefficients and cluster membership are alternately optimized under kernel alignment frame.In addition,we develop a three-step alternate iterative optimization algorithm to address the resultant optimization problem.Broad experiments on five benchmark data sets have been put into effect to evaluate the clustering performance of the proposed algorithm.The experimental results distinctly demonstrate that the proposed algorithm outperforms the typical multiple kernel clustering algorithms,which illustrates the effectiveness of the proposed algorithm.
基金National Natural Science Foundation of China (No. 60975083)Key Grant Project,Ministry of Education,China(No. 104145)
文摘A new algorithm named kernel bisecting k-means and sample removal(KBK-SR) is proposed as sampling preprocessing for support vector machine(SVM) training to improve the efficiency.The proposed algorithm tends to quickly produce balanced clusters of similar sizes in the kernel feature space,which makes it efficient and effective for reducing training samples.Theoretical analysis and experimental results on three UCI real data benchmarks both show that,with very short sampling time,the proposed algorithm dramatically accelerates SVM sampling and training while maintaining high test accuracy.
基金Supported by National Natural Science Foundation of China(60675039)National High Technology Research and Development Program of China(863 Program)(2006AA04Z217)Hundred Talents Program of Chinese Academy of Sciences
文摘The fuzzy C-means clustering algorithm(FCM) to the fuzzy kernel C-means clustering algorithm(FKCM) to effectively perform cluster analysis on the diversiform structures are extended, such as non-hyperspherical data, data with noise, data with mixture of heterogeneous cluster prototypes, asymmetric data, etc. Based on the Mercer kernel, FKCM clustering algorithm is derived from FCM algorithm united with kernel method. The results of experiments with the synthetic and real data show that the FKCM clustering algorithm is universality and can effectively unsupervised analyze datasets with variform structures in contrast to FCM algorithm. It is can be imagined that kernel-based clustering algorithm is one of important research direction of fuzzy clustering analysis.
文摘For the kernel K-mean cluster method is run in an implicit feature space, the initial and iterative cluster centers cannot be defined explicitly. Against the deficiency of the initial cluster centers selected in the original space discretionarily in the existing methods, this paper proposes a new method for ensuring the clustering center that virtual clustering centers are defined in the feature space by the original classification as the initial cluster centers and the iteration clustering centers are ensured by the further virtual classification. The improved method is used for fault diagnosis of roller bearing that achieves a good cluster and diagnosis result, which demonstrates the effectiveness of the proposed method.
基金The 15th Plan National Defence Preven-tive Research Project (No.413030201)
文摘To deal with the nonlinear separable problem, the generalized noise clustering （GNC） algorithm is extended to a kernel generalized noise clustering （KGNC） model. Different from the fuzzy c-means （FCM） model and the GNC model which are based on Euclidean distance, the presented model is based on kernel-induced distance by using kernel method. By kernel method the input data are nonlinearly and implicitly mapped into a high-dimensional feature space, where the nonlinear pattern appears linear and the GNC algorithm is performed. It is unnecessary to calculate in high-dimensional feature space because the kernel function can do it just in input space. The effectiveness of the proposed algorithm is verified by experiments on three data sets. It is concluded that the KGNC algorithm has better clustering accuracy than FCM and GNC in clustering data sets containing noisy data.
文摘Over the last fifteen years, face recognition has become a popular area of research in image analysis and one of the most successful applications of machine learning and understanding. To enhance the classification rate of the image recognition, several techniques are introduced, modified and combined. The suggested model extracts the features using Fourier-Gabor filter, selects the best features using signal to noise ratio, deletes or modifies anomalous images using fuzzy c-mean clustering, uses kernel least square and optimizes it by using wild dog pack optimization. To compare the suggested method with the previous methods, four datasets are used. The results indicate that the suggested methods without fuzzy clustering and with fuzzy clustering outperform state- of-art methods for all datasets.
基金Project supported by the 15th Plan for National Defence Preventive Research Project (Grant No.413030201)
文摘A novel model of fuzzy clustering using kernel methods is proposed. This model is called kernel modified possibilistic c-means （KMPCM） model. The proposed model is an extension of the modified possibilistic c-means （MPCM） algorithm by using kernel methods. Different from MPCM and fuzzy c-means （FCM） model which are based on Euclidean distance, the proposed model is based on kernel-induced distance. Furthermore, with kernel methods the input data can be mapped implicitly into a high-dimensional feature space where the nonlinear pattern now appears linear. It is unnecessary to do calculation in the high-dimensional feature space because the kernel function can do it. Numerical experiments show that KMPCM outperforms FCM and MPCM.
文摘Fuzzy c-means(FCM) clustering algorithm is sensitive to noise points and outlier data, and the possibilistic fuzzy c-means(PFCM) clustering algorithm overcomes the problem well, but PFCM clustering algorithm has some problems: it is still sensitive to initial clustering centers and the clustering results are not good when the tested datasets with noise are very unequal. An improved kernel possibilistic fuzzy c-means algorithm based on invasive weed optimization(IWO-KPFCM) is proposed in this paper. This algorithm first uses invasive weed optimization(IWO) algorithm to seek the optimal solution as the initial clustering centers, and introduces kernel method to make the input data from the sample space map into the high-dimensional feature space. Then, the sample variance is introduced in the objection function to measure the compact degree of data. Finally, the improved algorithm is used to cluster data. The simulation results of the University of California-Irvine(UCI) data sets and artificial data sets show that the proposed algorithm has stronger ability to resist noise, higher cluster accuracy and faster convergence speed than the PFCM algorithm.
文摘道路点云数据的障碍物检测技术在智能交通系统和自动驾驶中至关重要.传统的基于密度的空间聚类(DensityBased Spatial Clustering of Applications with Noise,DBSCAN)算法在处理高维或不同密度区域数据时,由于距离度量低效、参数组合确定困难导致聚类效果欠佳,因此,提出了一种基于改进DBSCAN的道路障碍物点云聚类方法 .首先,在确定Eps领域时利用孤立核函数来改进传统的距离度量方式,提高了DBSCAN聚类对不同密度区域的适应性和准确性.其次,针对猎豹优化算法(Cheetah Optimizer,CO)在信息共享和迭代更新方面的不足,提出了一种基于及时更新机制与兼容度量策略的CO优化算法(Timely Updating Mechanisms and Compatible Metric Strategies for CO Algorithms,TCCO),通过实时更新操作确保每次迭代的优秀信息得到及时沟通共享,并在全局更新时基于非支配排序与拥挤距离优化淘汰机制,平衡全局搜索和局部开发能力,提高了收敛速度和收敛精度.最后,利用孤立度量改进Eps领域,并利用TCCO优化DBSCAN聚类,自适应确定参数,提高了聚类精度和效率.在八个UCI数据集上进行测试,仿真结果表明,提出的TCCO-DBSCAN算法与CO-DBSCAN,SSA-DBSCAN,DBSCAN,KMC方法相比,F-Measure,ARI,NMI指标均有明显提升,且聚类精度更优.通过激光雷达点云数据障碍物聚类的实验验证,证明TCCO-DBSCAN能够有效地适应点云数据密度变化,获得更好的道路障碍物聚类效果,为辅助驾驶中障碍物检测提供支持.
