Data Preparation for Analytics Using SAS

Chapter 18

Multiple Interval-Scaled Observations per Subject

 

18.1 Introduction

18.2 Static Aggregation

18.3 Correlation of Values

18.4 Concentration of Values

18.5 Course over Time: Standardization of Values

18.6 Course over Time: Derived Variables

 

18.1  Introduction

If we have multiple numerical observations per subject, there are two major ways to make them available in a one-row-per-analysis subject. We can either transpose each value per subject into a separate column, or we can aggregate the multiple observations and condense the inherent information into some descriptive values. The pure transposition of values is needed for some analyses such as a repeated measurement analysis of variance.

For many data mining analyses, however, ...