Clustering

Abstract

Clustering is employed to divide a data set into an appropriate number of groups. Clustering is a form of unsupervised learning, which means a data scientist can bring labelled features of interest into the mining model. Furthermore, after dividing the data set, the data scientist can label each cluster. In business, clustering is used to analyze a customer or product segment that matches a target market. This chapter introduces clustering techniques including K-Means, Hierarchical Clustering, and DBSCAN, as well as techniques to indicate the efficiency of the clustering analysis. Data scientists can assess the efficiency of clustering analysis in two ways. Firstly, subjective measurement is where a data scientist consults a domain expert to confirm the efficiency of the cluster analysis, and secondly, data scientists can use objective measurements that test the efficiency of the cluster analysis result based on calculations. This chapter demonstrates cluster analysis adoption with RapidMiner so that readers can follow the process step-by-step.