Software engineering data analysis techniques

The purpose of this tutorial is to provide a comprehensive coverage of software engineering data analysis techniques. It will briefly cover the basic product and process metrics, their description, use and interpretation. A systematic approach for analyzing and interpreting software engineering data will be introduced. It explicitly recognizes that metrics data tend to have high dimensionality, are highly correlated and suffer from redundancy. The techniques to be presented become progressively more sophisticated in terms of the underlying theory and analysis as well as in their ability to provide insights into the software project and the development process. The specific techniques to be covered are: statistical analyses; regression modeling; stochastic models; classification trees; and neural networks. Our emphasis is on the so-called data mining techniques within the KDD (knowledge discovery in databases) framework.