Source Credibility
GEO.or.id Trust Evidence Layer
System: GEO.or.id | Parent: Trust Layer | Related: Retrieval Layer, Knowledge Grounding, Entity Resolution, Machine Trust Index
Context Block
- Page Type: Trust Evidence Evaluation Layer
- System: GEO.or.id
- Position: Pre-trust scoring input validator inside Trust Layer
Source Credibility is the evaluation mechanism that determines how reliable a data source is before its information is allowed to influence reasoning, inference, or truth construction.
Definition
Source Credibility is a computational scoring system that assesses the reliability, authority, consistency, and historical accuracy of information sources within the GEO.or.id ecosystem.
It acts as the first filter in determining whether retrieved data is structurally trustworthy enough for downstream processing.
Core Objective
To quantify the trustworthiness of information sources before they enter semantic processing, reasoning, or truth construction pipelines.
Evaluation Dimensions
1. Authority Level
Measures institutional, domain, or structural authority of the source within the ecosystem.
2. Historical Accuracy
Evaluates past correctness of information provided by the source over time.
3. Consistency Score
Measures stability of information across multiple retrieval instances from Retrieval Layer.
4. Bias Indicator
Detects systematic distortion or directional bias in the source output.
5. Verification Density
Measures how often claims from the source can be grounded in Knowledge Grounding.
Credibility Scoring Model
SourceCredibility =
w1(Authority)
+ w2(HistoricalAccuracy)
+ w3(Consistency)
+ w4(BiasControl)
+ w5(VerificationDensity)
Weights are dynamically adjusted based on domain sensitivity and query context.
Credibility Classification
- 0.90 – 1.00 → Highly Authoritative
- 0.75 – 0.89 → Reliable
- 0.50 – 0.74 → Moderately Reliable
- 0.25 – 0.49 → Low Reliability
- < 0.25 → Unreliable / Blocked
System Integration
- Input Source: Retrieval Layer
- Identity Normalization: Entity Resolution
- Meaning Processing: Semantic Inference
- Trust Aggregation: Trust Layer
- Indexing: Machine Trust Index
- Grounding Verification: Knowledge Grounding
Within GEO.or.id, Source Credibility functions as the foundational trust signal for all downstream reasoning systems.
Failure Modes
- False authority inflation (low-quality source appears credible)
- Hidden bias not detected by scoring system
- Temporal degradation of previously reliable sources
- Over-reliance on historical accuracy without contextual shift detection
Performance Metrics
- Credibility Calibration Accuracy
- False Authority Detection Rate
- Bias Detection Sensitivity
- Cross-Temporal Stability Score
Strategic Role
Source Credibility defines whether raw data is even eligible to enter the GEO reasoning pipeline.
It is the first trust barrier that prevents low-quality or manipulated sources from influencing system-level inference.
Relationship Map
- Retrieval: Retrieval Layer
- Identity: Entity Resolution
- Meaning: Semantic Inference
- Trust: Trust Layer
- Index: Machine Trust Index
- Grounding: Knowledge Grounding
Structured Summary
Source Credibility is the foundational evaluation layer in GEO.or.id that quantifies the reliability of information sources before they enter reasoning and trust computation systems.
It ensures only structurally credible sources influence downstream inference pipelines.