Incremental mitosis

Discovering clusters of arbitrary shapes and densities in dynamic data

Rania Ibrahim, Naglaa Ahmed, Noha Yousri, Mohamed A. Ismail

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

While finding natural clusters in high dimensional data is in itself a challenge, the dynamic nature of data adds another greater challenge. Many applications such as Data Warehouses and WWW demand the presence of efficient incremental clustering algorithms to handle their dynamic data. So far, numerous useful incremental clustering algorithms have been developed for large datasets such as incremental K-means, incremental DBSCAN, similarity histogram-based clustering (SHC) and mean shift. However, targeting clusters of different shapes and densities is yet to be efficiently tackled. In this work, an efficient incremental clustering algorithm (Incremental Mitosis) is proposed. It is based on Mitosis clustering algorithm which maximizes the relatedness of distances between patterns of the same cluster. The proposed algorithm is able to discover clusters of arbitrary shapes and densities in dynamic high dimensional data. Experimental results show that the proposed algorithm efficiently clusters the data and maintains the accuracy of Mitosis algorithm.

Original languageEnglish
Title of host publicationProceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012
Pages102-107
Number of pages6
Volume1
DOIs
Publication statusPublished - 1 Dec 2012
Externally publishedYes
Event11th IEEE International Conference on Machine Learning and Applications, ICMLA 2012 - Boca Raton, FL, United States
Duration: 12 Dec 201215 Dec 2012

Other

Other11th IEEE International Conference on Machine Learning and Applications, ICMLA 2012
CountryUnited States
CityBoca Raton, FL
Period12/12/1215/12/12

Fingerprint

Clustering algorithms
Data warehouses
World Wide Web
demand

Keywords

  • arbitrary shapes and densities
  • dynamic data
  • high dimensional data
  • incremental clustering

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Education

Cite this

Ibrahim, R., Ahmed, N., Yousri, N., & Ismail, M. A. (2012). Incremental mitosis: Discovering clusters of arbitrary shapes and densities in dynamic data. In Proceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012 (Vol. 1, pp. 102-107). [6406596] https://doi.org/10.1109/ICMLA.2012.26

Incremental mitosis : Discovering clusters of arbitrary shapes and densities in dynamic data. / Ibrahim, Rania; Ahmed, Naglaa; Yousri, Noha; Ismail, Mohamed A.

Proceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012. Vol. 1 2012. p. 102-107 6406596.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ibrahim, R, Ahmed, N, Yousri, N & Ismail, MA 2012, Incremental mitosis: Discovering clusters of arbitrary shapes and densities in dynamic data. in Proceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012. vol. 1, 6406596, pp. 102-107, 11th IEEE International Conference on Machine Learning and Applications, ICMLA 2012, Boca Raton, FL, United States, 12/12/12. https://doi.org/10.1109/ICMLA.2012.26
Ibrahim R, Ahmed N, Yousri N, Ismail MA. Incremental mitosis: Discovering clusters of arbitrary shapes and densities in dynamic data. In Proceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012. Vol. 1. 2012. p. 102-107. 6406596 https://doi.org/10.1109/ICMLA.2012.26
Ibrahim, Rania ; Ahmed, Naglaa ; Yousri, Noha ; Ismail, Mohamed A. / Incremental mitosis : Discovering clusters of arbitrary shapes and densities in dynamic data. Proceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012. Vol. 1 2012. pp. 102-107
@inproceedings{21805d33ba5943228e1bab14b0d7a66c,
title = "Incremental mitosis: Discovering clusters of arbitrary shapes and densities in dynamic data",
abstract = "While finding natural clusters in high dimensional data is in itself a challenge, the dynamic nature of data adds another greater challenge. Many applications such as Data Warehouses and WWW demand the presence of efficient incremental clustering algorithms to handle their dynamic data. So far, numerous useful incremental clustering algorithms have been developed for large datasets such as incremental K-means, incremental DBSCAN, similarity histogram-based clustering (SHC) and mean shift. However, targeting clusters of different shapes and densities is yet to be efficiently tackled. In this work, an efficient incremental clustering algorithm (Incremental Mitosis) is proposed. It is based on Mitosis clustering algorithm which maximizes the relatedness of distances between patterns of the same cluster. The proposed algorithm is able to discover clusters of arbitrary shapes and densities in dynamic high dimensional data. Experimental results show that the proposed algorithm efficiently clusters the data and maintains the accuracy of Mitosis algorithm.",
keywords = "arbitrary shapes and densities, dynamic data, high dimensional data, incremental clustering",
author = "Rania Ibrahim and Naglaa Ahmed and Noha Yousri and Ismail, {Mohamed A.}",
year = "2012",
month = "12",
day = "1",
doi = "10.1109/ICMLA.2012.26",
language = "English",
isbn = "9780769549132",
volume = "1",
pages = "102--107",
booktitle = "Proceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012",

}

TY - GEN

T1 - Incremental mitosis

T2 - Discovering clusters of arbitrary shapes and densities in dynamic data

AU - Ibrahim, Rania

AU - Ahmed, Naglaa

AU - Yousri, Noha

AU - Ismail, Mohamed A.

PY - 2012/12/1

Y1 - 2012/12/1

N2 - While finding natural clusters in high dimensional data is in itself a challenge, the dynamic nature of data adds another greater challenge. Many applications such as Data Warehouses and WWW demand the presence of efficient incremental clustering algorithms to handle their dynamic data. So far, numerous useful incremental clustering algorithms have been developed for large datasets such as incremental K-means, incremental DBSCAN, similarity histogram-based clustering (SHC) and mean shift. However, targeting clusters of different shapes and densities is yet to be efficiently tackled. In this work, an efficient incremental clustering algorithm (Incremental Mitosis) is proposed. It is based on Mitosis clustering algorithm which maximizes the relatedness of distances between patterns of the same cluster. The proposed algorithm is able to discover clusters of arbitrary shapes and densities in dynamic high dimensional data. Experimental results show that the proposed algorithm efficiently clusters the data and maintains the accuracy of Mitosis algorithm.

AB - While finding natural clusters in high dimensional data is in itself a challenge, the dynamic nature of data adds another greater challenge. Many applications such as Data Warehouses and WWW demand the presence of efficient incremental clustering algorithms to handle their dynamic data. So far, numerous useful incremental clustering algorithms have been developed for large datasets such as incremental K-means, incremental DBSCAN, similarity histogram-based clustering (SHC) and mean shift. However, targeting clusters of different shapes and densities is yet to be efficiently tackled. In this work, an efficient incremental clustering algorithm (Incremental Mitosis) is proposed. It is based on Mitosis clustering algorithm which maximizes the relatedness of distances between patterns of the same cluster. The proposed algorithm is able to discover clusters of arbitrary shapes and densities in dynamic high dimensional data. Experimental results show that the proposed algorithm efficiently clusters the data and maintains the accuracy of Mitosis algorithm.

KW - arbitrary shapes and densities

KW - dynamic data

KW - high dimensional data

KW - incremental clustering

UR - http://www.scopus.com/inward/record.url?scp=84873606313&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873606313&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2012.26

DO - 10.1109/ICMLA.2012.26

M3 - Conference contribution

SN - 9780769549132

VL - 1

SP - 102

EP - 107

BT - Proceedings - 2012 11th International Conference on Machine Learning and Applications, ICMLA 2012

ER -