@techreport{TR-IC-PFG-25-57,
   number = {IC-PFG-25-57},
   author = {Clara {Mattos Medeiros} and André {Santanchè}},
   title = {{A Value-Based Foreign Key Analysis Approach for Enriching
                   Object-Centric Data Platforms}},
   month = {December},
   year = {2025},
   institution = {Institute of Computing, University of Campinas},
   note = {In English, 17 pages.
    \par\selectlanguage{english}\textbf{Abstract}
       Large-scale  enterprise  ontologies  often suffer from critical
       fragmentation,  where logical connections between data entities
       remain  unmapped  due  to  siloed  development  and  unreliable 
       metadata.   This   paper   presents   a  scalable,  value-based 
       methodology   for   the   automatic   discovery   of   implicit  
       relationships within industrial big data platforms. To overcome
       the  limitations  of  metadata-driven  approaches, the proposed
       solution  implements  a  foreign  key  analysis  that  directly 
       examines  data  values. We introduce a novel confidence scoring
       model,  incorporating source saturation, target coverage, and a
       Primary  Key  Randomness  Score,  to  quantify the strength and
       validity  of  potential  links while mitigating false positives
       from  low-entropy  identifiers.  The  algorithm  was  validated 
       against  a control group of existing relationships, achieving a
       recall  rate of ~72\%, and subsequently identified thousands of
       previously  unmapped  connections,  significantly enhancing the
       graph's    connectivity.    The    methodology    was   further  
       operationalized  into  a  self-service  governance application,
       empowering  data  teams  to  continuously enrich the enterprise
       data  model  and strengthen the integrity of the organizational
       digital twin.
  }
}