Student research opportunities

Looking for unusual documents in a collection

Project Code: CECS_923

This project is available at the following levels:
Honours, Summer Scholar, Masters, PhD

Keywords:

text analysis, data mining

Supervisor:

Dr Wray Buntine

Outline:

The Patent Lens is a collection of millions of patent with rich meta-data. Unusual patents could be of interest for many reasons: a computer hardware company exploring new techniques, a new cross-discipline, an obfuscated patent that needs to be better tagged, etc. How might we use the meta-data and content for different parts of the patent, claims and/or abstract, the background work, etc.? How might "unusual" be defined. What techniques can we use to find "unusual".

Goals of this project

Undertake research to establish what could be viewed as "unusual", explore the existing literature, and then apply some standard/pre-existing machine learning methods for the task.