-
Members: FreeCIS
IEEE Members: Free
Non-members: FreeLength: 01:00:30
16 Apr 2013
Abstract: Kernel Methods have seen an upsurge in computational intelligent because they are a mathematically principled methodology with excellent compromise between power, flexibility and computation complexity. This talk will critically survey the methodology, review the most common applications and will present recent developments in online kernel learning to attack both the problem of nonstationarity in time series analysis (streaming data) and large scale learning problems.
