Bayes Factor
Evidence Router

Peer-reviewed

Maps a Bayes factor BF (or its log-scale equivalent 2·ln BF) to a categorical evidence-strength token on the Jeffreys/Kass-Raftery scale. This is the gate primitive used in ChainGraph evidence chains.

Input
10¹⁴
NANOGrav 15-yr GWB: BF > 10¹⁴ (arXiv:2306.16213)
Evidence classification
Decisive
gate_token: decisive
Building artifact…

      

        
        
      

    

  


  

    
    

      
What this tool does

      
Given a Bayes factor BF comparing two hypotheses (e.g., H1: gravitational wave background vs. H0: noise), or its log-scale 2·ln BF, this tool classifies the evidence strength using the Jeffreys (1961) / Kass & Raftery (1995) thresholds and returns a routing token for use in ChainGraph evidence chains.

      
Scales

      
    
        
  • BF < 1 — supports null hypothesis
    • 
        
  • 1 ≤ BF < 3.2 — weak (barely worth mentioning)
    • 
        
  • 3.2 ≤ BF < 10 — substantial
    • 
        
  • 10 ≤ BF < 100 — strong
    • 
        
  • BF ≥ 100 — decisive
    • 
      

      
Worked example

      
NANOGrav 15-year data set (arXiv:2306.16213) reports BF > 10¹⁴ for the gravitational wave background, classifying as decisive evidence.

      
Citations

      
    
        
  • Jeffreys, H. 1961, Theory of Probability, 3rd ed., Oxford University Press
    • 
        
  • Kass, R.E. & Raftery, A.E. 1995, JASA 90:773 (DOI 10.2307/2291091)
    • 
        
  • NANOGrav 2023, arXiv:2306.16213
    • 
      

      
Tool v1.2.0 · mandate me.omegacentauri/bayes_factor · register: peer-reviewed · OCG artifact v0.4.0