Data structures which change with time, a machine learning approach

  • Lee, Vincent (Primary Chief Investigator (PCI))
  • Hagenbuchner, Markus (Chief Investigator (CI))
  • Tsoi, Ah, (Chief Investigator (CI))
  • Gori, Marco (Partner Investigator (PI))
  • Scarselli, Franco (Partner Investigator (PI))

Project: Research

Project Description

This project proposes a novel parametric model capable of modelling the time evolution of, e.g. page importance of web pages in the Internet, with a time varying graph data structure which can consist of a mixture of directed acyclic graphs, un-directed graphs, cyclic graphs. Theoretical properties of the model including conditions under which solutions exist and universal approximator will be studied. The parameters can be estimated using a learning algorithm on a set of training examples. The model is validated by modeling time series obtained through selective crawling of a portion of the Internet at intervals capturing its key dynamics.
StatusFinished
Effective start/end date1/09/0728/02/12

Funding

  • Australian Research Council (ARC): AUD405,000.00
  • Australian Research Council (ARC)
  • Australian Research Council (ARC)