Skip to main content

Dataset Details

caution

A visual quick reference guide of this page is coming soon.

The initial step in creating a dataset is defining its details. QuickGraph accommodates various stages of annotation project maturity by supporting two core dataset types: standard and annotated. For more information, visit theConcepts - Datasets page.

The dataset details comprises the following components:

  • Name: Assign a unique name to identify the dataset, or choose a randomized name.
  • Description: (optional)
  • Annotated Dataset: Specify the type of dataset you wish to create. Options include: No, Yes - Entities, Yes - Entities and Relations. Default No.

If Yes - Enitites is selected for Annotated Dataset:

  • Suggested Annotations: Specifies whether the dataset should default to setting annotations as suggested (annotators need to accept them) or accepted.
  • Entity Ontology: Populated with system and user-created entity ontology resources. One must be selected. This is used to validate the entities in the annotated dataset.

If Yes - Entities and Relations is selected for Annotated Dataset:

  • Suggested Annotations: Specifies whether the dataset should default to setting annotations as suggested (annotators need to accept them) or accepted.
  • Entity Ontology: Populated with system and user-created entity ontology resources. One must be selected. This is used to validate the entities in the annotated dataset.
  • Relation Ontology: Populated with system and user-created relation ontology resources. One must be selected. This is used to validate the relations in the annotated dataset.