1 / 19

A novel ant-based clustering algorithm using the kernel method

A novel ant-based clustering algorithm using the kernel method. Lei Zhang*, Qixin Cao 2011, InfSci Presented by Chien-Hao Kung 2011/10/6. Outlines. Motivation Objectives Methodology Experiments Conclusions Comments. Motivation.

ziva
Download Presentation

A novel ant-based clustering algorithm using the kernel method

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A novel ant-based clustering algorithm using the kernel method Lei Zhang*, Qixin Cao 2011, InfSci Presented by Chien-Hao Kung 2011/10/6

  2. Outlines • Motivation • Objectives • Methodology • Experiments • Conclusions • Comments

  3. Motivation • Ant-based clustering is not highly efficient because of the randomness in the algorithm. • Ant-based clustering algorithms are distance-based, if the separation boundaries between clusters are nonlinear, however, the algorithms will fail.

  4. Objectives • There are two aspects to the integration. • Kernel principal component analysis(KPCA) is applied to modify the random projection of objects. • To use kernel function calculated the distance between the objects in input space as a similarity measure.

  5. Methodology Grid Object Ant The ant-based clustering algorithm

  6. Methodology • Kernel-based clustering • Mercer kernels • Commonly used kernel functions • Kernel-based clustering

  7. Methodology Enlarging Rounding Shifting • The novel ant-based clustering with the kernel method(ACK)

  8. Methodology • Parameter setting • Size of the projection plane • Radius r • Kernel size • Scaling parameter • Threshold

  9. Experiments • Evaluation functions • The F-measure (F) • The Dunn Index (DI) • The inner cluster variance (ICV) • The error rate (ER) • Time cost (T)

  10. Experiments Experimental data

  11. Experiments

  12. Experiments

  13. Experiments

  14. Experiments

  15. Experiments

  16. Experiments

  17. Experiments

  18. Conclusions • The algorithm has some new characteristics • The algorithm can deal with some datasets with non-Gaussian distribution because of the incorporation of the kernel function. • The projection based on KPCA creates rough clusters, which reduces the running time and increases the algorithm’s efficiency. • Performing clustering in the feature space after kernel mapping can improve clustering quality.

  19. Comments • Advantages • The paper has some wrong. Application • Ant-based clustering

More Related