Search

Density-peak clustering (DPC) is a novel clustering algorithm that has received much attention from researchers. Its principle is to identify the cluster center of the data by density and distance. However, the density-peak clustering algorithm still suffers from poor clustering results due to irrational choices of cluster centers and misallocating samples. To tackle these problems, a novel tree structure-based multi-prototype clustering algorithm (MPCTS) is proposed in this paper. Firstly, a division-merger method is adopted to divide the dataset into multiple peak clusters based on multi-prototype clustering, which could minimize the impact of high-density sample points on the subsequent sample point assignments. Secondly, the KNN nearest-neighbor search strategy is improved to quickly obtain the sub-cluster center in each peak cluster to overcome the defect of irrational selection of initial cluster centers. Afterward, a tree-based decay strategy is designed to evaluate the correlation between sample points, and a tree-based association strategy is designed to achieve the merging of multiple peak clusters to tackle the sample points misallocation problem. Finally, the extensive experiments on clustering task are conducted over several benchmark synthetic and real datasets to verify the efficiency and superiority of the proposed MPCTS.

The density peak clustering algorithm (DPC) requires manual determination of cluster centers, and poor performance on complex datasets with varying densities or non-convexity. Hence, a novel density peak clustering algorithm is proposed for the automatic selection of clustering centers based on K-nearest neighbors (AKDPC). First, the AKDPC classifies samples according to their mutual K-nearest neighbor values into core and non-core points. Second, the AKDPC uses the average distance of K nearest neighbors of a sample as its density. The smaller the average distance is, the higher the density. Subsequently, it selects the highest density sample among all unclassified core points as a center of the new cluster, and the core points that satisfy the merging condition are added to the cluster until no core points satisfy the condition. Afterwards, the above steps are repeated to complete the clustering of all core points. Lastly, the AKDPC labels the unclassified non-core points similar to the nearest points that have been classified. In addition, to prove the validity of AKDPC, experiments on manual and real datasets are conducted. By comparing the AKDPC with classical clustering algorithms and excellent DPC-variants, this paper demonstrates that AKDPC presents higher accuracy.

DPC(clustering by fast search and find of density peaks) is an efficient clustering algorithm. However, DPC and its variations usually cannot detect the appropriate cluster centers for a dataset containing sparse and dense clusters simultaneously, resulting in the unique clustering within a dataset cannot being found. To remedy these limitations, we propose an Adaptive Nearest Neighbor Density Peak Clustering algorithm, referred to as ANN-DPC. It introduces the adaptive nearest neighbors for a point, so as to define the accurate local density of the point. Moreover, it partitions points into super-score, core, linked and slave points, and proposes techniques to detect appropriate cluster centers through introducing super-core point with higher local density to absorb the other super-core points sharing adaptive nearest neighbors with it and the dependency vector for finding next cluster center. Furthermore, novel assignment strategies are proposed by leveraging the adaptive nearest neighbors combing with breadth first search and fuzzy weighted adaptive nearest neighbors, so as to assign non-center points to the most appropriate clusters. Extensive experiments on real-world and synthetic datasets demonstrate the superiority of the proposed ANN-DPC algorithm over the counterparts in precisely detecting the cluster centers and the unique clustering within a dataset. [Display omitted] • Introducing adaptive nearest neighbors for a point to define its accurate local density and partition it as a super-core, core, linked, or slave point. • Introducing super-core point absorbing technique and the dependency vector to detect the appropriate cluster centers within a dataset. • New assignment strategies are present by leveraging the adaptive nearest neighbors combining with breadth first search and fuzzy weighted adaptive nearest neighbors. • ANN-DPC algorithm is introduced by leveraging these aforementioned contributions. • Extensive experiments demonstrate the superiority of the ANN-DPC over existing counterparts. [ABSTRACT FROM AUTHOR]

Career choice has a pivotal role in college students' life planning. In the past, professional career appraisers used questionnaires or diagnoses to quantify the factors potentially influencing career choices. However, due to the complexity of each person's goals and ideas, it is difficult to properly forecast their career choices. Recent evidence suggests that we could use students' behavioral data to predict their career choices. Based on the simple premise that the most remarkable characteristics of classes are reflected by the main samples of a category, we propose a model called the Approach Cluster Centers Based On XGBOOST (ACCBOX) model to predict students' career choices. The experimental results of predicting students' career choices clearly demonstrate the superiority of our method compared to the existing state-of-the-art techniques by evaluating on 13 M behavioral data of over four thousand students. [ABSTRACT FROM AUTHOR]

Density-peak clustering (DPC) is a novel clustering algorithm that has received much attention from researchers. Its principle is to identify the cluster center of the data by density and distance. However, the density-peak clustering algorithm still suffers from poor clustering results due to irrational choices of cluster centers and misallocating samples. To tackle these problems, a novel tree structure-based multi-prototype clustering algorithm (MPCTS) is proposed in this paper. Firstly, a division-merger method is adopted to divide the dataset into multiple peak clusters based on multi-prototype clustering, which could minimize the impact of high-density sample points on the subsequent sample point assignments. Secondly, the KNN nearest-neighbor search strategy is improved to quickly obtain the sub-cluster center in each peak cluster to overcome the defect of irrational selection of initial cluster centers. Afterward, a tree-based decay strategy is designed to evaluate the correlation between sample points, and a tree-based association strategy is designed to achieve the merging of multiple peak clusters to tackle the sample points misallocation problem. Finally, the extensive experiments on clustering task are conducted over several benchmark synthetic and real datasets to verify the efficiency and superiority of the proposed MPCTS. [ABSTRACT FROM AUTHOR]

In this work, with the purpose of improving the performance of extension neural network (ENN), we use Fuzzy c-means (FCM) clustering algorithm to locate the initial centers of every class before the training. In traditional ENN, the initial centers are defined simply by the average values of the minimum and maximum of every characteristic. Our proposed FENN (FCM–ENN) in this paper is different from tradition ENN, and the initial centers of every class are determined by the cluster centers of FCM clustering algorism. Our proposed strategy can reflect the actual training data distribution information, thereby the performance of ENN by using this strategy is more approach to practical situation. Compared with traditional ENN, the proposed FENN has a better performance. Experimental results from three different examples, including an artificial data set, a benchmark data set and a practical application, verify the effectiveness and applicability of our proposed work. [ABSTRACT FROM AUTHOR]

Career choice has a pivotal role in college students’ life planning. In the past, professional career appraisers used questionnaires or diagnoses to quantify the factors potentially influencing career choices. However, due to the complexity of each person’s goals and ideas, it is difficult to properly forecast their career choices. Recent evidence suggests that we could use students’ behavioral data to predict their career choices. Based on the simple premise that the most remarkable characteristics of classes are reflected by the main samples of a category, we propose a model called the Approach Cluster Centers Based On XGBOOST (ACCBOX) model to predict students’ career choices. The experimental results of predicting students’ career choices clearly demonstrate the superiority of our method compared to the existing state-of-the-art techniques by evaluating on 13 M behavioral data of over four thousand students.

In recent times, graph based spectral clustering algorithms have received immense attention in many areas like, data mining, object recognition, image analysis and processing. The commonly used similarity measure in the clustering algorithms is the Gaussian kernel function which uses sensitive scaling parameter and when applied to the segmentation of noise contaminated images leads to unsatisfactory performance because of neglecting the spatial pixel information. The present work introduces a novel framework for spectral clustering which embodied local spatial information and fuzzy based similarity measure to tackle the above mentioned issues. In our approach, firstly we filter the noise components from original image by using the spatial and gray-level information. The similarity matrix is then constructed by employing a similarity measure which takes into account the fuzzy c-partition matrix and vectors of the cluster centers obtained by fuzzy c-means clustering algorithm. In the last step, spectral clustering technique is realized on derived similarity matrix to obtain the desired segmentation result. Experimental results on segmentation of synthetic and Berkeley benchmark images with noise demonstrates the effectiveness and robustness of the proposed method, giving it an edge over the clustering based segmentation method reported in the literature. [ABSTRACT FROM AUTHOR]

The cluster k of traditional K-means algorithm could not determine beforehand and the initial clustering centers of K-means algorithm were randomly selected, which might result in low accurary and unstable clustering. TMs paper based on the SSE for selecting the number of clusters k, based on the principle that the clustering center of the surrounding area was relatively dense, and between the clustering center distance was relatively far, selected the initial clustering center to avoid the initial clustering center focused on a small range, prevented fall into local optimum. In the case of the number of categories k was given. This paper used the standard UCI data sets for test. Tests show that，this method can select the optimal value of k, it can choose the only center of initial clustering and have the higher accuracy and the minimum errors. [ABSTRACT FROM AUTHOR]

The density peak clustering algorithm (DPC) requires manual determination of cluster centers, and poor performance on complex datasets with varying densities or non-convexity. Hence, a novel density peak clustering algorithm is proposed for the automatic selection of clustering centers based on K-nearest neighbors (AKDPC). First, the AKDPC classifies samples according to their mutual K-nearest neighbor values into core and non-core points. Second, the AKDPC uses the average distance of K nearest neighbors of a sample as its density. The smaller the average distance is, the higher the density. Subsequently, it selects the highest density sample among all unclassified core points as a center of the new cluster, and the core points that satisfy the merging condition are added to the cluster until no core points satisfy the condition. Afterwards, the above steps are repeated to complete the clustering of all core points. Lastly, the AKDPC labels the unclassified non-core points similar to the nearest points that have been classified. In addition, to prove the validity of AKDPC, experiments on manual and real datasets are conducted. By comparing the AKDPC with classical clustering algorithms and excellent DPC-variants, this paper demonstrates that AKDPC presents higher accuracy.

Data clustering is one of the most popular techniques in data mining to group data with great similarity and high dissimilarity into each cluster. This paper presents a new clustering method based on a novel heuristic optimization algorithm proposed recently and named as multivariant optimization algorithm (MOA) to locate the optimal solution automatically through global and local alternating search implemented by a global exploration group and several local exploitation groups. In order to demonstrate the performance of MOA-clustering method, it is applied to group six real-life datasets to obtain their clustering results, which may be compared with those received by employing K-means algorithm, genetic algorithm and particle swarm optimization. The results show that the proposed clustering algorithm is an effective and feasible method to reach a high accurate rate and stability in clustering problems. [ABSTRACT FROM AUTHOR]

Aiming at the problems of the K-nearest neighbor algorithm, testing complex is linear at least, and lead to the accuracy is low when the samples are large. This paper proposed a fast KNN classification algorithm faster than the traditional KNN did. The proposed algorithm innovatively introduced the training process during the KNN method, i.e., the algorithm blocked the big data by linear complexity clustering. Then, the algorithm selected the nearest cluster as new training samples and established a classification model. This process reduced the KNN algorithm testing overhead, which made the proposed algorithm could be applied to big data. Experiments result shows that the accuracy of the proposed KNN classification is similarity than the traditional KNN, but the classification speed has been significantly improved. [ABSTRACT FROM AUTHOR]

How to find the representative bands is a key issue in band selection for hyperspectral data. Very often, unsupervised band selection is associated with data clustering, and the cluster centers (or exemplars) are considered ideal representatives. However, partitioning the bands into clusters may be very time-consuming and affected by the distribution of the data points. In this letter, we propose a new band selection method, i.e., exemplar component analysis (ECA), aiming at selecting the exemplars of bands. Interestingly, ECA does not involve actual clustering. Instead, it prioritizes the bands according to their exemplar score, which is an easy-to-compute indicator defined in this letter measuring the possibility of bands to be exemplars. As a result, ECA is of high efficiency and immune to distribution structures of the data. The experiments on real hyperspectral data set demonstrate that ECA is an effective and efficient band selection method. [ABSTRACT FROM PUBLISHER]

Abstract: Fuzzy c-means clustering with spatial constraints is considered as suitable algorithm for data clustering or data analyzing. But FCM has still lacks enough robustness to employ with noise data, because of its Euclidean distance measure objective function for finding the relationship between the objects. It can only be effective in clustering ‘spherical’ clusters, and it may not give reasonable clustering results for “non-compactly filled” spherical data such as “annular-shaped” data. This paper realized the drawbacks of the general fuzzy c-mean algorithm and it tries to introduce an extended Gaussian version of fuzzy C-means by replacing the Euclidean distance in the original object function of FCM. Firstly, this paper proposes initial kernel version of fuzzy c-means to aim at simplifying its computation and then extended it to extended Gaussian kernel version of fuzzy c-means. It derives an effective method to construct the membership matrix for objects, and it derives a robust method for updating centers from extended Gaussian version of fuzzy C-means. Furthermore, this paper proposes a new prototypes learning method and it obtains initial cluster centers using new mathematical initialization centers for the new effective objective function of fuzzy c-means, so that this paper tries to minimize the iteration of algorithms to obtain more accurate result. Initial experiment will be done with an artificially generated data to show how effectively the new proposed Gaussian version of fuzzy C-means works in obtaining clusters, and then the proposed methods can be implemented to cluster the Wisconsin breast cancer database into two clusters for the classes benign and malignant. To show the effective performance of proposed fuzzy c-means with new initialization of centers of clusters, this work compares the results with results of recent fuzzy c-means algorithm; in addition, it uses Silhouette method to validate the obtained clusters from breast cancer datasets. [Copyright &y& Elsevier]

This paper presents a new and efficient integrated neural fuzzy approach for transformer fault diagnosis using dissolved gas analysis. The proposed approach formulates the modeling problem of higher dimensions into lower dimensions by using the input feature selection based on competitive learning and neural fuzzy model. Then, the fuzzy rule base for the identification of fault is designed by applying the subtractive clustering method which is very good at handling the noisy input data. Verification of the proposed approach has been carried out by testing on standard and practical data. In comparison to the results obtained from the existing conventional and neural fuzzy techniques, the proposed method has been shown to possess superior performance in identifying the transformer fault type. [ABSTRACT FROM AUTHOR]

Generalized zero-shot classification is predicting the labels of the test images coming from seen or unseen classes. The task is difficult because of the bias problem, that is, unseen samples are easily to be misclassified to seen classes. Many methods have handled the problem by training a generative adversarial network (GAN) to generate fake samples. However, the GAN model trained with seen samples might not be appropriate for generating unseen samples. For dealing with this problem, we learn a bias alleviating generative adversarial network for generalized zero-shot classification by generating seen and unseen samples, simultaneously. We train the generator to generate more realistic unseen samples by adding semantic similarity and cluster center regularizations to alleviate the bias problem. The semantic similarity regularization is to restrict the relationships of the generated unseen visual prototypes and seen visual prototypes by their class prototypes to avoid the generated unseen samples similar to the seen samples. The cluster center regularization is to utilize the cluster property of target data to make the generated unseen visual prototypes near to the most similar cluster centers, generating realistic unseen samples. From the experiments, we can see the proposed method achieves promising results. • Train the generator by generating seen and unseen samples, simultaneously. • Make the generated unseen visual prototypes similar to the nearest real cluster centers. • Preserve the semantic relationships of the seen and the generated unseen visual prototypes. • The proposed method achieves state-of-the-art results. [ABSTRACT FROM AUTHOR]