#### DMCA

## Clustering of datasets using PSO-K-Means and PCA-K-means

### Citations

3766 | Particle swarm optimization
- Kennedy, RC
- 1995
(Show Context)
Citation Context ... another, but are very dissimilar to objects in other clusters. The paper is organized as follows. The section 2 of this paper briefs the procedure and functionality of PSO while section 3 deals with PSO-K-means, a combination of PSO and K-means to perform clustering. The combination of principal component analysis and kmeans is described in Section 4. The experimental clustering results of various datasets implemented in MATLAB are produced in section 5. 2. Particle Swarm Optimization PSO is a population based stochastic optimization technique developed by Dr. Eberhart and Dr. Kennedy in 1995[1][5][7], inspired by social behaviour of bird flocking or fish schooling and has been rapidly applied to data mining tasks such as classification and clustering to optimize the results. Clustering using PSO has been applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to ... |

851 | The particle swarm-explosion, stability, and convergence in a multidimensional complex space
- Clerc, Kennedy
- 2002
(Show Context)
Citation Context ...other, but are very dissimilar to objects in other clusters. The paper is organized as follows. The section 2 of this paper briefs the procedure and functionality of PSO while section 3 deals with PSO-K-means, a combination of PSO and K-means to perform clustering. The combination of principal component analysis and kmeans is described in Section 4. The experimental clustering results of various datasets implemented in MATLAB are produced in section 5. 2. Particle Swarm Optimization PSO is a population based stochastic optimization technique developed by Dr. Eberhart and Dr. Kennedy in 1995[1][5][7], inspired by social behaviour of bird flocking or fish schooling and has been rapidly applied to data mining tasks such as classification and clustering to optimize the results. Clustering using PSO has been applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to opt... |

299 | On clustering validation techniques,"
- Halkidi, Batistakis, et al.
- 2001
(Show Context)
Citation Context ...means and PCA-KMeans improves the performance of basic K-means in terms of accuracy and computational time. Keywords: Clustering, PSO, PSO-K-means, QPSO 1. Introduction Data clustering is a technique in which data with similar characteristics are grouped together to form clusters. Clustering has been studied by many researchers for a long time and been applied in areas such as pattern recognition, gene expression analysis, customer segmentation, educational research and etc. Clustering techniques are generally categorized into hierarchical, partitional, density based and grid based clustering [10]. Hierarchical Clustering: These methods start with each point being considered a cluster and recursively combine pairs of clusters (subsequently updating the inter-cluster distances) until all points are part of one hierarchically constructed cluster. Divisive or Partitional Clustering: Partitional clustering, on the other hand, performs a partition of patterns into K number of clusters, such that patterns in a cluster are more similar to each other than to patterns in different clusters. Recent studies have shown that partitional clustering algorithms are more suitable for clustering large d... |

201 | K-means clustering via principal component analysis.
- Ding, He
- 2004
(Show Context)
Citation Context ...etween intra cluster distance and inter cluster distance as in eqn (4). The clustering which gives a minimum value for the validity measure gives the ideal value of K. ValidityCluster = Intra / Inter (4) 4. PCA-K-Means Principal component analysis is invented by Karl Pearson[11] and the general idea of using PCA is to reduce the dimensionality of data prior to some other statistical processing. Dimension reduction is closely related to unsupervised clustering. PCA based data reduction outperforms traditional noise reduction techniques and has been applied to clustering gene expression profiles[13] . PCA has also been demonstrated in the applications of face recognition, image compression, and to find patterns in high dimensional data. The main objective of using PCA is dimensionality reduction of data and to find new underlying variables. An approach to find or extract key components from original data influences the division of clusters. Principal component analysis is one such approach constructs a linear combination of a set of vectors that can best describe the variance of data. It is more common that PCA can be used to project the data into lower dimension subspace by picking up t... |

49 |
Particle swarm optimization with particles having quantum behavior,”
- Sun, Feng, et al.
- 2004
(Show Context)
Citation Context ...bust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to optimum solution but also higher accuracy. PSO-Kmeans is successfully demonstrated on Bus standards for static and transient security evaluation[12].The variants of PSO and its applications are proposed in [8] and [9]. A disadvantage of the global PSO is it tends to be trapped in a local optimum under some initialization conditions[6]. In PSO, the potential solutions, called particles, searches the whole space guided by its previous best position(pbest) and best position of the swarm(gbest). The velocity and position of the particles are updated based on its best experience. The ith particle of swarm is represented as Xi = (Xi1, Xi2, …, XiD) while the velocity for ith particle is represented as Vi = (Vi1, Vi2, …, ViD). The best previous position (the position giving the best fitness value) of the ith particle is recorded and represented as pi = (pi1, pi2, … , piD). At each step, the particles are manipulated according to t... |

44 | Document clustering using particle swarm optimization,”
- Cui, Potok, et al.
- 2005
(Show Context)
Citation Context ...of various datasets implemented in MATLAB are produced in section 5. 2. Particle Swarm Optimization PSO is a population based stochastic optimization technique developed by Dr. Eberhart and Dr. Kennedy in 1995[1][5][7], inspired by social behaviour of bird flocking or fish schooling and has been rapidly applied to data mining tasks such as classification and clustering to optimize the results. Clustering using PSO has been applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to optimum solution but also higher accuracy. PSO-Kmeans is successfully demonstrated on Bus standards for static and transient security evaluation[12].The variants of PSO and its applications are proposed in [8] and [9]. A disadvantage of the global PSO is it tends to be trapped in a local optimum under some initialization conditions[6]. In PSO, the potential solutions, called particles, se... |

24 | An analysis of publications on particle swarm optimisation applications.
- Poli
- 2007
(Show Context)
Citation Context ... applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to optimum solution but also higher accuracy. PSO-Kmeans is successfully demonstrated on Bus standards for static and transient security evaluation[12].The variants of PSO and its applications are proposed in [8] and [9]. A disadvantage of the global PSO is it tends to be trapped in a local optimum under some initialization conditions[6]. In PSO, the potential solutions, called particles, searches the whole space guided by its previous best position(pbest) and best position of the swarm(gbest). The velocity and position of the particles are updated based on its best experience. The ith particle of swarm is represented as Xi = (Xi1, Xi2, …, XiD) while the velocity for ith particle is represented as Vi = (Vi1, Vi2, …, ViD). The best previous position (the position giving the best fitness value) of the ith parti... |

11 |
Particle swarm optimization for clustering of wireless sensors”,
- Tillett, Sahin, et al.
- 2003
(Show Context)
Citation Context ...rm clustering. The combination of principal component analysis and kmeans is described in Section 4. The experimental clustering results of various datasets implemented in MATLAB are produced in section 5. 2. Particle Swarm Optimization PSO is a population based stochastic optimization technique developed by Dr. Eberhart and Dr. Kennedy in 1995[1][5][7], inspired by social behaviour of bird flocking or fish schooling and has been rapidly applied to data mining tasks such as classification and clustering to optimize the results. Clustering using PSO has been applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to optimum solution but also higher accuracy. PSO-Kmeans is successfully demonstrated on Bus standards for static and transient security evaluation[12].The variants of PSO and its applications are proposed in [8] and [9]. A disadvantage of the global PSO is... |

9 |
Particle swarm optimization based K-Means clustering approach for security assessment in power systems, Expert systems with applications,
- Kalyani, Swarup
- 2011
(Show Context)
Citation Context ... and clustering to optimize the results. Clustering using PSO has been applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to optimum solution but also higher accuracy. PSO-Kmeans is successfully demonstrated on Bus standards for static and transient security evaluation[12].The variants of PSO and its applications are proposed in [8] and [9]. A disadvantage of the global PSO is it tends to be trapped in a local optimum under some initialization conditions[6]. In PSO, the potential solutions, called particles, searches the whole space guided by its previous best position(pbest) and best position of the swarm(gbest). The velocity and position of the particles are updated based on its best experience. The ith particle of swarm is represented as Xi = (Xi1, Xi2, …, XiD) while the velocity for ith particle is represented as Vi = (Vi1, Vi2, …, ViD). The best previous p... |

8 |
Some issues and practices for particle swarms, in:
- Kennedy
- 2007
(Show Context)
Citation Context ...er, but are very dissimilar to objects in other clusters. The paper is organized as follows. The section 2 of this paper briefs the procedure and functionality of PSO while section 3 deals with PSO-K-means, a combination of PSO and K-means to perform clustering. The combination of principal component analysis and kmeans is described in Section 4. The experimental clustering results of various datasets implemented in MATLAB are produced in section 5. 2. Particle Swarm Optimization PSO is a population based stochastic optimization technique developed by Dr. Eberhart and Dr. Kennedy in 1995[1][5][7], inspired by social behaviour of bird flocking or fish schooling and has been rapidly applied to data mining tasks such as classification and clustering to optimize the results. Clustering using PSO has been applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to optimu... |

3 |
Ahmadyfard and Hamidreza Modares, “Combining PSO and k- means to enhance data clustering”,
- Alireza
- 2008
(Show Context)
Citation Context ...technique developed by Dr. Eberhart and Dr. Kennedy in 1995[1][5][7], inspired by social behaviour of bird flocking or fish schooling and has been rapidly applied to data mining tasks such as classification and clustering to optimize the results. Clustering using PSO has been applied in wireless sensor networks [2], tested against random search and simulated annealing, and found to be more robust. PSO has also been applied in document clustering [3] which demonstrated that the hybrid PSO algorithm generated more compact clusters in comparison to the K-means algorithm. Combining K-means and PSO[4] for data clustering achieved not only fast convergence to optimum solution but also higher accuracy. PSO-Kmeans is successfully demonstrated on Bus standards for static and transient security evaluation[12].The variants of PSO and its applications are proposed in [8] and [9]. A disadvantage of the global PSO is it tends to be trapped in a local optimum under some initialization conditions[6]. In PSO, the potential solutions, called particles, searches the whole space guided by its previous best position(pbest) and best position of the swarm(gbest). The velocity and position of the particles a... |

1 |
K-Means Clustering via Pricipal Component Analysis”,
- Ding, He
- 2004
(Show Context)
Citation Context ...uster Validity International Journal of Computational Intelligence and Informatics, Vol. 1 : No. 3, October - December 2011 182 1 2 3 4 5 0 200 400 600 Fig. 2. Histogram of PSO-K-means clusters of Breast Cancer for K=5 In order to improve the PSO-K-means as an automatic one, the validity measure is calculated as the ratio between intra cluster distance and inter cluster distance as in eqn (4). The clustering which gives a minimum value for the validity measure gives the ideal value of K. ValidityCluster = Intra / Inter (4) 4. PCA-K-Means Principal component analysis is invented by Karl Pearson[11] and the general idea of using PCA is to reduce the dimensionality of data prior to some other statistical processing. Dimension reduction is closely related to unsupervised clustering. PCA based data reduction outperforms traditional noise reduction techniques and has been applied to clustering gene expression profiles[13] . PCA has also been demonstrated in the applications of face recognition, image compression, and to find patterns in high dimensional data. The main objective of using PCA is dimensionality reduction of data and to find new underlying variables. An approach to find or extra... |