Skip to content

Class: Splits

Are there recommended data splits (e.g., training, validation, testing)? If so, how are they defined and why?

__

URI: data_sheets_schema:Splits

classDiagram class Splits click Splits href "../Splits/" DatasetProperty <|-- Splits click DatasetProperty href "../DatasetProperty/" Splits : description Splits : id Splits : name Splits : split_details Splits : used_software Splits --> "*" Software : used_software click Software href "../Software/"

Inheritance

Slots

Name Cardinality and Range Description Inheritance
split_details *
String		 Free-text description of the recommended data splits (e direct
id 0..1
Uriorcurie		 An optional identifier for this property DatasetProperty
name 0..1
String		 A human-readable name for this property DatasetProperty
description 0..1
String		 A human-readable description for this property DatasetProperty
used_software *
Software		 What software was used as part of this dataset property? DatasetProperty

Usages

used by used in type used
Dataset splits range Splits
DataSubset splits range Splits

Identifier and Mapping Information

Schema Source

  • from schema: https://w3id.org/bridge2ai/data-sheets-schema

Mappings

Mapping Type Mapped Value
self data_sheets_schema:Splits
native data_sheets_schema:Splits

LinkML Source

Direct

name: Splits
description: 'Are there recommended data splits (e.g., training, validation, testing)?
  If so, how are they defined and why?

  '
from_schema: https://w3id.org/bridge2ai/data-sheets-schema
is_a: DatasetProperty
attributes:
  split_details:
    name: split_details
    description: 'Free-text description of the recommended data splits (e.g., 80/10/10
      train/ validation/test), how they are defined, and the rationale for the split
      strategy.

      '
    from_schema: https://w3id.org/bridge2ai/data-sheets-schema/composition
    rank: 1000
    slot_uri: dcterms:description
    domain_of:
    - Splits
    range: string
    multivalued: true

Induced

name: Splits
description: 'Are there recommended data splits (e.g., training, validation, testing)?
  If so, how are they defined and why?

  '
from_schema: https://w3id.org/bridge2ai/data-sheets-schema
is_a: DatasetProperty
attributes:
  split_details:
    name: split_details
    description: 'Free-text description of the recommended data splits (e.g., 80/10/10
      train/ validation/test), how they are defined, and the rationale for the split
      strategy.

      '
    from_schema: https://w3id.org/bridge2ai/data-sheets-schema/composition
    rank: 1000
    slot_uri: dcterms:description
    alias: split_details
    owner: Splits
    domain_of:
    - Splits
    range: string
    multivalued: true
  id:
    name: id
    annotations:
      d4d:docExample:
        tag: d4d:docExample
        value: https://example.org/dataset/property-001
    description: An optional identifier for this property.
    from_schema: https://w3id.org/bridge2ai/data-sheets-schema/base
    slot_uri: schema:identifier
    alias: id
    owner: Splits
    domain_of:
    - NamedThing
    - DatasetProperty
    range: uriorcurie
  name:
    name: name
    description: A human-readable name for this property.
    from_schema: https://w3id.org/bridge2ai/data-sheets-schema/base
    slot_uri: schema:name
    alias: name
    owner: Splits
    domain_of:
    - NamedThing
    - DatasetProperty
    range: string
  description:
    name: description
    description: A human-readable description for this property.
    from_schema: https://w3id.org/bridge2ai/data-sheets-schema/base
    slot_uri: schema:description
    alias: description
    owner: Splits
    domain_of:
    - NamedThing
    - DatasetProperty
    - DatasetRelationship
    range: string
  used_software:
    name: used_software
    description: What software was used as part of this dataset property?
    from_schema: https://w3id.org/bridge2ai/data-sheets-schema/base
    rank: 1000
    slot_uri: d4d:usedSoftware
    alias: used_software
    owner: Splits
    domain_of:
    - DatasetProperty
    range: Software
    multivalued: true
    inlined: true
    inlined_as_list: true