labelgedrag

Labelgedrag is a term used to describe the behavior and impact of labels in tagging and categorization systems. It encompasses how labels are created, chosen, and presented, as well as how users and automated processes respond to them. The concept covers how labeling affects perception, decision making, and subsequent actions, as well as how labeling practices influence data quality and workflow outcomes.

The word is a portmanteau of label and gedrag, with gedrag meaning behavior in Dutch. It appears in Dutch-language discussions of information management, human–computer interaction, and data annotation, and is occasionally used in English-language writings to describe interactions between labels and users or data.

In human–computer interaction, labelgedrag examines label clarity, wording, length, and placement, and how these factors influence task performance and error rates. It also addresses label ambiguity, label fatigue, and the effects of visual clutter on usability. Designers seek labeling conventions that reduce confusion and support consistent interpretation across contexts and users.

In data labeling and machine learning, labelgedrag relates to label quality and consistency, inter-annotator agreement, and label noise. It considers how labeling choices affect model training, evaluation, and downstream decisions, including issues such as class imbalance and concept drift. Label leakage, where labels reveal unintended information, is also discussed as a related concern.

In linguistics and psychology, labelgedrag refers to how labels influence perception, categorization, memory, and judgment, including phenomena where labels bias interpretation or choice.

Related concepts include label noise, inter-annotator agreement, active learning, and concept drift.