Annotation

A label or mark attached (electronically or manually, by a human or a machine) to a discussion object.

The annotation of past e-discussions is crucial for Deep Loop analysis, which will use them for deriving associations between aspects of a discussion, a cluster of similar discussions/ participants, and classes of, say good or bad contributions, argumentation shapes etc. The overall aim is to recognize rules for quality e-discussions and effective interventions.

...a teacher, moderator or knowledge engineer will be able to annotate parts of a discussion.

Annotations will be exploited for the off-line training / computation of Deep Loop indicators.

Annotations will be made according to annotation schemes.

Annotation manuals will be designed respectively.

Currently, we consider annotation on 4 four levels of annotations: 1) Shape-level annotations: Annotations can be attached to each shape (or node) in the discussion graph. 2) Cluster-level annotations: Annotations can be attached to a group (or cluster of nodes). 3) Annotation of single actions: Annotations can be attached to actions logs, by taking a process view. 4) Annotation of sequences of actions: Annotations can be attached to sequences of actions (or their corresponding segments of the actions logs).

(ARGUNAUT Deliverable 1.1).