Introduction

In every line of business ranging from running a small business to creating and managing a mission critical application, there are a number of tasks that are common and need to be included as a part of almost every workflow that is required during the course of executing the functions. This is true even for building robust machine learning systems. In Spark machine learning, some of these tasks range from splitting the data for model development (train, test, validate) to normalizing input feature vector data to creating ML pipelines via the Spark API. We provide a set of recipes in this chapter to enable the reader to think about what is actually required to implement an end-to-end machine learning system.

This chapter attempts to demonstrate a number of common tasks which are present in any robust Spark machine learning system implementation. To avoid redundant references these common tasks in every recipe covered in this book, we have factored out such common tasks as short...