ABSTRACT
In Chapter 6, we presented data preprocessing techniques, feature reduction methods, and statistical tests for hypothesis testing. In software engineering research, the researcher may relate the software metrics (independent variables) with quality attributes (dependent variable) such as fault proneness, maintainability, reliability, or testability. The relationship between software metrics and quality attributes can be analyzed using statistical or machine learning (ML) techniques. The models are created to predict the quality attributes using performance measures or analyzers. After obtaining the values of performance measures, the hypothesis may be applied to analyze the difference between the techniques over multiple data sets. The results are then interpreted and assessed, and final conclusions are derived.
