Multidimensional scaling and k-means clustering

This chapter covers two multivariate statistical techniques called k-means clustering and multidimensional scaling (MDS). It shows how both methods can be used together as the algorithms involved are used to try to find the best solution without knowing any ground truth group or category labels. K-means clustering aims to partition multivariate data into several clusters, such that observations within a given cluster are more similar to each other than other cluster observations.The chapter further notes that the purpose of multidimensional scaling is to reduce complex multivariate data sets to a smaller number of dimensions to facilitate graphical representation. The chapter also discusses the multidimensional scaling in R.