Efficient Algorithms for Hierarchical Agglomerative Clustering

• NA

• Author / Creator

  Anandan, Ajay

• This thesis proposes and evaluates methods to improve two algorithmic ap-
  proaches for Hierarchical Agglomerative Clustering. These new methods in-
  crease the scalability and speed of the traditional Hierarchical Agglomerative
  Clustering algorithm without using any approximations. The first method
  exploits the characteristics of modern Non-Uniform Memory Access architec-
  tures, resulting in a parallel algorithm for the stored matrix version of Hier-
  archical Agglomerative Clustering. The second method uses a data structure
  called the Cover Tree to speed up the stored data version of the Hierarchical
  Agglomerative Clustering. For the second method, the thesis proposes both se-
  quential and parallel algorithms. All methods were experimentally evaluated
  and compared against the state-of-the-art approaches for high performance
  clustering. The results demonstrate the superiority of the parallel approaches
  with respect to all baselines and previous work, and the comparison between
  the stored matrix and the stored

• Subjects / Keywords

  • Cover tree
  • Hierarchical Agglomerative Clustering

• Graduation date

  Fall 2013

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/R35Q4RX5S

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.