Data Clustering and Various Clustering Approaches

Data Clustering and Various Clustering Approaches

Shashi Mehrotra, Shruti Kohli
Copyright: © 2017 |Pages: 19
DOI: 10.4018/978-1-5225-1776-4.ch004
(Individual Chapters)
No Current Special Offers


It is needed to organize the data in different groups for various purposes, where clustering is useful. The chapter covers Data Clustering in the detail, which includes; introduction to data clustering with figures, data clustering process, basic classification of clustering and applications of clustering, describing hard partition clustering and fuzzy clustering. Some most commonly used clustering method are explained in the chapter with their features, advantages, and disadvantages. A various variant of K-Means and extension method of hierarchical clustering method, density-based clustering method and grid-based clustering method are covered.
Chapter Preview


What Is Clustering?

Clustering is a method of grouping the objects based on their characteristics. The objects which are having similar characteristics are assigned the same group called cluster while the objects having different characteristics are assigned different cluster (Das, Abraham, & Konar, 2008). A cluster contains a collection of object, which is similar in nature, while different from the objects in another cluster. Thus, clustering means the grouping of the objects based on some similarity among objects. Objects belonging to a cluster are more similar to each other compared to the objects belonging to another cluster (Jain, Murty, & Flynn, 1999). Cluster analysis is an unsupervised learning i.e., no predefined class. A typical way to use cluster analysis is to get insight into data distribution or as a preprocessing step for other algorithms. The following figure is an example for clusters of various shapes.

Figure 1 shows cluster of three different shapes. One cluster has star objects, the other cluster has circle objects, and the third cluster has triangles. Each cluster has objects that are similar in shape while they are different from another cluster.

Figure 1.

Three clusters


Applications of Clustering

Cluster analysis has a lot of applications, and being used in various fields such as used for artificial intelligence, spatial data analysis, image segmentation, document collection, multimedia data analysis or social media analysis, generating a concept summary of the data, pattern analysis, outlier detection, fraud detection, recommendation system, business, marketing, customer segmentation, and for preprocessing as an intermediate step (Das, Abraham, & Konar, 2008).

A lot of research has been conducted and is going on in the above-mentioned area. Still, a lot of scope is there for research in clustering, it can be effectively used to get insight from the data. Clustering analysis research is being widely going on.

Clustering is being used very often for the medical purpose. Medical sciences use clustering to identifying lump etc., using image segmentation. Clustering is being used in various ways for web applications. The web is being used as the largest repository. Search results contains title and snippets. Results clustered in a meaning full folder will facilitate user to search relevant result in quick manner (Kohli & Mehrotra, 2016).

Search engines cluster search results for a large number of web pages to represent them in more organized way. The Clusty is the most common example, which uses clustering for search result organization.

Clustering analysis is widely used to get insight from the data and used it for policy making and decision making in many business organizations. Clustering may be helpful in target marketing, document retrieval, grouping people in social network in our day to day life (Mehrotra & Kohli, 2016, March).

For the efficient marketing, customers are grouped into different types for targeting the prospective user segment (Jain, 2010). Users are segmented according to their income group, their life style, their interest, their past purchasing behavior, their age group etc.

For example, various income group can be presented in different clusters as given below:

Figure 2 shows three clusters of the different salary group people. One cluster is for the high-income group, and two others are for medium income group and low-income group.

Figure 2.

Clusters of a different income group


Example 1

A retailer wants to know people to whom they can sell which car. For this purpose, they can create different groups of people based on their income. Customer data can be grouped into a low-income group, medium-income group, and high- income group, and to market for the car retailer will target the specific customer segment accordingly.

Example 2

A housing company can target specified income group for the specific housing scheme, such as they will target higher income group for HIG houses or for penthouses while lower income group are to be targeted for LIG houses.

Complete Chapter List

Search this Book: