Machine Learning with R - Fourth Edition

13

Challenging Data – Too Much, Too Little, Too Complex

Challenging data takes many forms throughout the course of a machine learning project, and the journey of each new project represents an adventure requiring a pioneer spirit. Beginning with uncharted data that must be explored, the data must then be wrangled before it can be used with the learning algorithm. Even then, there may still be wild aspects of the data that need to be tamed for the project to be successful. Extraneous information must be culled, small-but-important details must be cultivated, and tangled webs of complexity must be cleared from the learner’s path.

Conventional wisdom in the big data era suggests that data is treasure, but as the saying goes, one can have “too much ...