glam/schemas/20251121/linkml/modules/classes/DetectionThreshold.yaml

107 lines
4.6 KiB
YAML

id: https://nde.nl/ontology/hc/class/DetectionThreshold
name: DetectionThreshold
title: Detection Threshold Class
prefixes:
linkml: https://w3id.org/linkml/
hc: https://nde.nl/ontology/hc/
dqv: http://www.w3.org/ns/dqv#
schema: http://schema.org/
prov: http://www.w3.org/ns/prov#
dcterms: http://purl.org/dc/terms/
crm: http://www.cidoc-crm.org/cidoc-crm/
skos: http://www.w3.org/2004/02/skos/core#
rdfs: http://www.w3.org/2000/01/rdf-schema#
org: http://www.w3.org/ns/org#
xsd: http://www.w3.org/2001/XMLSchema#
default_prefix: hc
imports:
- linkml:types
- ../metadata
- ../slots/has_or_had_label
- ../slots/has_or_had_description
- ../slots/has_or_had_type
classes:
DetectionThreshold:
class_uri: dqv:QualityMeasurement
description: "Configuration for detection thresholds in analysis pipelines.\n\n**DEFINITION**:\n\nDetectionThreshold represents the confidence threshold settings used to filter\ndetection results. CV models output confidence scores; thresholds determine\nwhich detections are included in results.\n\n**Threshold Levels**:\n\n| Threshold | Range | Use Case |\n|-----------|-------|----------|\n| HIGH_PRECISION | 0.9+ | Production display, high confidence |\n| BALANCED | 0.7-0.9 | General use, balance precision/recall |\n| HIGH_RECALL | 0.5-0.7 | Research, review, catch more |\n| RAW | < 0.5 | Unfiltered, needs post-processing |\n\n**Ontological Alignment**:\n- **DQV**: `dqv:QualityMeasurement` - quality metric for data assessment\n- **PROV-O**: Threshold as parameter of detection activity\n\n**Migrated From** (per slot_fixes.yaml):\n- `detection_threshold` (float) now uses:\n - `filters_or_filtered` \u2192 DetectedEntity\n - `has_or_had_treshold` \u2192 DetectionThreshold (this class)\n\
\n**Usage Pattern**:\n```\nVideoAnnotation\n \u2514\u2500\u2500 filters_or_filtered \u2192 DetectedEntity\n \u2514\u2500\u2500 has_or_had_treshold \u2192 DetectionThreshold\n \u251C\u2500\u2500 threshold_value: 0.5\n \u2514\u2500\u2500 threshold_type: MINIMUM\n```\n"
exact_mappings:
- dqv:QualityMeasurement
close_mappings:
- schema:QuantitativeValue
related_mappings:
- prov:SoftwareAgent
slots:
- has_or_had_label
- has_or_had_description
- has_or_had_type
attributes:
threshold_value:
range: float
required: true
minimum_value: 0.0
maximum_value: 1.0
description: 'The numeric threshold value (0.0-1.0).
Detections with confidence >= threshold_value are included.
'
examples:
- value: 0.5
description: Standard threshold
- value: 0.9
description: High precision threshold
threshold_type:
range: string
required: false
description: 'Type of threshold application.
- MINIMUM: Lower bound for inclusion
- MAXIMUM: Upper bound (rare)
- BAND: Range between two values
'
examples:
- value: MINIMUM
description: Minimum confidence for inclusion
slot_usage:
has_or_had_label:
description: Human-readable label for this threshold configuration
examples:
- value: High Precision Threshold
- value: Research Mode Threshold
has_or_had_type:
description: Type category for the threshold (HIGH_PRECISION, BALANCED, etc.)
examples:
- value: HIGH_PRECISION
- value: BALANCED
annotations:
custodian_types: '["D"]'
custodian_types_rationale: Detection thresholds apply to digital platforms with automated analysis
specificity_score: 0.75
specificity_rationale: Fairly specific to video/media analysis contexts
comments:
- Represents detection threshold configuration
- Migrated from detection_threshold slot per slot_fixes.yaml
- threshold_value is the numeric confidence cutoff
- threshold_type indicates how threshold is applied
examples:
- value:
threshold_value: 0.5
threshold_type: MINIMUM
has_or_had_label: Standard Detection
description: Standard detection threshold at 0.5 confidence
- value:
threshold_value: 0.9
threshold_type: MINIMUM
has_or_had_label: High Precision
has_or_had_description: For production display requiring high confidence
description: High precision threshold for production use
- value:
threshold_value: 0.3
threshold_type: MINIMUM
has_or_had_label: Research Mode
has_or_had_description: Low threshold to maximize recall for research
description: Low threshold for research/review workflows