A Memory Efficient Distributed Fuzzy Joint Points Clustering Algorithm

ATILGAN C., Nasibov E.

10th IEEE International Conference on Application of Information and Communication Technologies (AICT), Baku, Azerbaycan, 12 - 14 Ekim 2016, ss.401-405 identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: Baku
  • Basıldığı Ülke: Azerbaycan
  • Sayfa Sayıları: ss.401-405
  • Anahtar Kelimeler: Clustering, Fuzzy joint points (FJP), Memory, efficiency
  • Dokuz Eylül Üniversitesi Adresli: Evet

Özet

The fuzzy joint points (FJP) is a method that uses a fuzzy neighborhood notion to deal with neighborhood parameter selection issue of classical density-based clustering and offers an unsupervised clustering tool. Recent works improved the method in terms of speed to enable the method for big data applications. However, space efficiency of the method is still a limiting factor. In this work, we discuss techniques to improve the space efficiency of the method, so that FJP is applicable regardless of the size of data and offer a distributed version of the algorithm.