Development Of Algorithms For Clustering Of Large Data Sets

Abstract

The management and handling of an extensive database efficiently proceed various data-oriented tasks such as analysis, data retrieval and mapping of knowledge. newline In addition, data mining provides multiple tools such as classification, regression and clustering. newline The main objective of clustering to measure useful groups of entities and to differentiate clusters formed for a dataset. newline The conventional approach of clustering algorithms categorized into four section such as partition clustering, density-based clustering, hierarchal clustering and micro clustering. newline The most of data driven process applied the partition-based clustering approach for the formation of clusters and extraction of patterns. newline The dimension of data is always challenging tasks for the clustering algorithms. newline The various approach of clustering algorithms faces a problem of similarity, cluster validation and large number of iterations. The process of clustering requires the application of more precise definition of observation and cluster validation. The concept of groping based on attributes; it is natural employ concept of distance. newline Clustering algorithms are the backbone of data engineering and pattern analysis. newline The unknown attribute of data and large-size faces of bottleneck problem of grouping and estimating data similarity is resolved by the clustering algorithm [1, 2]. newline The diverse nature of clustering algorithm divided into several groups such as partition clustering, hierarchal clustering, density-based clustering and micro clustering [3]. newline The principle of clustering further explodes into three sections such as overlapping, partitioned and hierarchical [4, 5]. newline The efficient and straightforward clustering process is called partition clustering. The partition clustering has several variants of algorithms such as K-means, K-mode and FCM. newline The FCM clustering algorithm is also called soft clustering algorithm [6, 7, 8].