Discriminant and Cluster Analysis
Maria Cristina Mariani, Osei Kofi Tweneboah, Maria Pia Beccar-Varela · 2021
This chapter presents the discriminant analysis technique. Discriminant analysis finds a set of prediction equations based on explanatory variables that are used to classify individuals into groups. There are two possible goals in a discriminant analysis: finding a predictive equation for classifying new individuals and interpreting the predictive equation to better understand the relationships that may exist among the variables. The chapter discusses cluster analysis. Cluster analysis is different from discriminant analysis in the sense that there are no predefined classes. The chapter introduces the concept of distance. Most multivariate techniques are based upon the concept of distance. The Kullback-Leibler divergence is used to measure the distance between two probability distributions. It helps to measure how one probability distribution is different from other. Chernoff distance or Chernoff-a coefficient is also useful in discriminating the statistical samples or populations. The chapter presents some background infomation of the seismic time series.