Annotatie

Annotatie, or annotation, refers to the act of adding notes, explanations, or metadata to a primary source or data object. Annotations aim to clarify content, provide interpretive guidance, or enable machine processing. They can be attached to texts, images, audio, video, or datasets and may be visible to readers or stored as metadata or markup.

There are several types of annotations. Textual annotations include glosses and marginal notes. In linguistic research, annotations cover part-of-speech tagging, lemmatization, and syntactic structure. Semantic annotations link content to concepts in ontologies or knowledge graphs. Metadata annotations describe provenance, authorship, format, or technical properties. In data science and machine learning, labeling or annotating data (images, texts, audio) provides the ground truth for training and evaluation.

Annotation can be performed manually by experts or by automated tools, often with human validation. Inter-annotator agreement is a common measure of reliability. Annotation schemes and guidelines help ensure consistency across annotators. Markup formats vary by domain, including XML-based schemas such as TEI for text or JSON-LD for semantic data.

Applications of annotation span linguistic research, digital humanities, information retrieval, and data curation. Annotations improve searchability, support both qualitative and quantitative analyses, and enhance accessibility. Ethical considerations include privacy and bias protection when annotating sensitive material.