ABSTRACT
This chapter discusses the empirical approach for software measurements using machine learning (ML) techniques. It demonstrates the usage of ML techniques for both software quality and quantity measurements. With a basic introduction to the current trends of the field and moving through problem definition, the author reaches the experimental set-up and then draws inferences from the experiments. The chapter aims to provide the reader practical and applicable knowledge of ML and deep learning for empirical software measurements. Several software engineering tasks like various development and maintenance tasks that come under software engineering can be formulated as learning problems and can be handled as application of learning algorithms. The desired output is over a range of numbers, hence, the problem can be formulated as a regression-based learning problem. In empirical software measurement, effort estimation problem is formulated as a regression-based supervised learning problem.
