WebAnno

WebAnno facilitates annotation of linguistic and non-linguistic corpora across morphological, syntactic, and semantic layers to support construction of training and test datasets and evaluation of natural language processing algorithms.

Key Features:

• Layered annotation: Supports annotation across morphological, syntactic, and semantic layers.
• Custom annotation layers: Allows definition of user-specified annotation layers for specialized labeling tasks.
• Schema definition: Supports explicit definition of annotation schemas to structure annotation projects.
• Corpus construction: Facilitates creation of training and test corpora for development and evaluation of NLP algorithms.
• Non-linguistic annotation: Extends annotation capabilities beyond linguistic phenomena to other domain-specific annotations.
• Comparative evaluation (PMID: 31838514): Reported score of 0.81/1.0 in a published comparative study evaluating 78 annotation tools on 26 criteria across three compliance levels.

Scientific Applications:

• NLP dataset generation: Generation of annotated training and test datasets for supervised machine learning and evaluation of NLP models.
• Linguistic research and domain annotation: Structuring and labeling corpora for linguistic studies and broader scientific projects requiring detailed annotations.

Methodology:

Evaluation described in PMID: 31838514 assessed 78 annotation tools using 26 functional and technical criteria at three levels with scores assigned, yielding WebAnno a score of 0.81/1.0.