Roadmap

Application development

• Create Cloud Formation Template for easy install on AWS
• Improve accuracy
  • Optimize baseline scores
  • Use data from CTSAsearch for journal category / department
  • Weight by co-author proportion
  • Bayesian likelihood for matching Scopus institutions
  • Bayesian likelihood for matching first name.
  • Discount likelihood when other user in Identity has similar name
• Improve computation time
  • Run all scoring strategies simultaneously
  • Look for redundancies
  • Use synonyms as a utility function, moving it to the bootstrap level.
  • Explore and implement SQS / other solutions for dealing with concurrency
• Standardize the way we clean names. Global sanization of all names rather than each attribute level.
• Better documentation
• Test and improve sameAs relationships between Scopus and PubMed
• Create integration with other source systems
  • Web of Science
  • WorldCat
  • CrossRef
  • Microsoft Academic
  • Symplectic Dimensions

User interface application

• Bulk feedback view
• Manually add publication
• User can maintain disambiguation profile (API and UI)
• Show top concepts associated with accepted papers

Testing

• Collaborate with group of other implementation sites
• Measure
  • Accuracy
  • Time saved
  • Ease of implementation