WP7: Learning complete super models
This workpackage will develop machine learning methods for learning complete supermodels. We expect that supermodels will be
built in three phases. The machine learning methods will first generate diverse models, then select a set of complementary models, and
finally learn the interconnections between the constituent models of an ensemble. These three phases
naturally lead to the three tasks that constitute this WP.
JSI, which has ample expertise in machine learning, will lead this WP and the constituent tasks.
MASA will also contribute significantly, in close collaboration with JSI, providing insights
from non-linear dynamics.
WP7 is chaired by Saso Dzeroski
