One of the key challenges utilizing supervised machine learning in the real world is that most models require not only a sample of data large enough to represent the actual reality, but also labeled data. The labeling process is a tedious and drawn-out task and often expensive. But we can improve this process and save on time and money with a technique called active learning.
In this book we discuss different active learning strategies and show how these strategies can be implemented in a KNIME workflow. The result is a web-based application that leads the subject matter expert through the active learning process of training a model.