Classifi cation described in the previous chapter generally determines discrete categories from given variables. In this section, as a task similar to classifi cation, prediction of continuous variables based on other variables will be explained. In prediction, the former and latter are called dependent variables and independent variables, respectively. As techniques for prediction of dependent variables as functions of independent variables, regression as a basic method and multivariate analysis as its advanced method will be briefl y surveyed. Model construction based on these techniques is indispensable so as to create and confi rm concrete quantitative hypotheses at the conceptual layer constituting big data application systems.
Classifi cation and prediction have much in common. Here relations between classifi cation and prediction will be explained fi rst.