PAGE TITLE
Multi-temporal information object classification and categorization software
PAGE SUMMARY
Drexel researchers have developed an integrated software system that disambiguates, classifies, enriches, and categorizes digital informational objects, such as manuscripts and authors. Indexing big data troves are computationally expensive endeavors. With the developed system, the organization and linking of granular data elements of electronic information is optimized. Computational time is reduced, as the number of iterations to reach convergence is lower with the implementation of incremental affinity propagation than with existing methods. From applying machine learning algorithms, the system can automatically adjust the timing, scope, and scale of data indexing. While this software has been applied to journal articles, additional real-world datasets are being tested and accuracy benchmarking is being performed.
APPLICATIONS
TITLE: Applications
Big data analysis
Maintain and update indexes for accurate data access
ADVANTAGES
TITLE:Advantages
System can incrementally update as new documents are added
Accelerated clustering
Faster processing times reduce computational needs
Modular system with adjustable timing, scope, and scale
Leverage automated and manual data linking operations
IP STATUS
Intellectual Property and Development Status
United States Issued Patent- 11,482,307
https://patents.google.com/patent/US11482307B2/en?oq=11%2c482%2c307
CONTACT INFORMATION
Tanvi Muni
Licensing Manager
Drexel University
tm3439@drexel.edu