Signals System Index

Signals System Index

The Signals System is the observability and feature extraction layer of GEO.or.id. It converts raw interactions, content structures, and system behaviors into measurable indicators used across retrieval, evidence, trust, and reasoning layers.

Signals do not represent truth. They represent measurable properties that help other systems estimate truth.

1. Signals System Role

Signals operate as a cross-system measurement layer:

Raw Data → Signal Extraction → Aggregation → Scoring → System Consumption

Signals are used by Retrieval, Evidence, Ontology, Reasoning, and Trust systems.

2. Core Signal Types

• Retrieval Signals
• Semantic Signals
• Entity Signals
• Relationship Signals

3. Trust-Oriented Signals

Signals that directly influence system trust evaluation:

• Trust Signals
• Authority Signals
• Consistency Signals
• Freshness Signals
• Citation Signals

4. Semantic Structure Signals

Signals used for meaning construction and ontology alignment:

• Semantic Signals
• Grounding Signals
• Relationship Signals

5. Retrieval Control Signals

Signals that influence how retrieval systems rank and select information:

• Retrieval Signals

These directly affect: Retrieval Ranking, Source Selection, Re-ranking System

6. Signal Aggregation Logic

Signals are not independent. They are aggregated into weighted system features:

Signal Vector = Σ (feature signals × context weight × time decay factor)

Aggregated signals are consumed by: Trust System, Evidence System

7. Signal Flow in Architecture

Content / Query / Interaction
→ Signal Extraction
→ Signal Normalization
→ Signal Aggregation
→ System Layer Consumption

8. Signal Quality Constraints

• Signals must be measurable, not interpretive
• No signal without definable extraction logic
• No signal without downstream system usage
• Signals must decay if not reinforced

9. System Integration Points

• Retrieval System
• Evidence System
• Ontology System
• Trust System
• Reasoning System

10. System Principle

• Signals are observations, not truths
• All signals are context-dependent
• Signals must be aggregated before interpretation
• Weak signals decay without reinforcement
• No decision is made from signals alone