An Optimized Chameleon Algorithm based on Local Features

Cao, Xiaoxiao; Su, Tianyun; Wang, Pengyu; Wang, Guoyu; Lv, Zhihan; Li, Xinfang

doi:10.1145/3195106.3195118

Cited by 7 publications

(2 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The agglomerative hierarchical approaches begin with considering the individual point in data as a cluster, then continually fusing the most similar clusters until only one remains at the end. Two algorithms, BIRCH [54] and CHAMELEON, take advantage of this concept [16]. A divisive hierarchical clustering approach treats the entire data set as a single large cluster, then splits the most relevant clusters at each stage until a user-defined threshold of clusters is reached.…”

Section: Single Machine Clustering Techniquesmentioning

confidence: 99%

“…The SPMD (Single Program Multiple Data) technique is used to achieve the parallel paradigm in this program, and the message forwarding mechanism protects communication between processors. Using a work pool analogy, a parallel variant of the CHAMELEON method was described in [16]. This approach is divided into three stages-the first concerns using the concurrent K-Nearest Neighbor approach to reduce the study's temporal complexity.…”

Section: Parallel Clusteringmentioning

confidence: 99%

See 1 more Smart Citation

Big Data Clustering Techniques Challenged and Perspectives: Review

Awad¹,

Hamad²

2023

IJCAI

View full text Add to dashboard Cite

Clustering in big data is considered a critical data mining and analysis technique. There are issues with adapting clustering algorithms to large amounts of data and new challenges brought by big data. As the size of big data is up to petabytes of data, and clustering methods have high processing costs, the challenge is how to handle this issue and utilize clustering techniques for big data efficiently. This study aims to investigate the recent advancement of clustering platforms and techniques to handle big data issues, from the early suggested techniques to today's novel solutions. The methodology and specific issues for building an effective clustering mechanism are presented and evaluated, followed by a discussion of the choices for enhancing clustering algorithms. A brief literature review of the recent advancement in clustering techniques has been presented to address each solution's main characteristics and drawbacks.Povzetek: Članek predstavlja pregled tehnik gručenja za velike podatke.

show abstract

Section: Single Machine Clustering Techniquesmentioning

confidence: 99%

Section: Parallel Clusteringmentioning

confidence: 99%