labelling

Terms from Artificial Intelligence: humans at the heart of algorithms

Page numbers are for draft copy at present; they will be replaced with correct numbers when final book is formatted. Chapter numbers are correct and will not change now.

Labelling is the process of giving data items some form of label, often a classification. In supervised machine learning, human labelling is often necessary for training, but sometimes there may be an automatic labelling, for example, if there are multiple kinds of sensors, or post-hoc knowledge about outcomes. The execution phase of machine learning often involves allocating labels to unlabelled data.

Used on Chap. 9: pages 194, 195; Chap. 12: page 265; Chap. 20: page 495