In this work, we face the problem of training sample collection for the estimation of biophysical parameters by adopting the active learning approach. In particular, we propose two active learning strategies specifically developed for Gaussian Process (GP) regression. The first one is based on adding samples that are distant from the current training samples in the kernel space while the second one exploits an intrinsic GP regression outcome to pick up the most difficult samples. Experiments on simulated and real data sets show the effectiveness of active selection of training samples for regression problems. © 2011 IEEE.
Gaussian process regression within an active learning scheme / Pasolli, E.; Melgani, F.. - (2011), pp. 3574-3577. (Intervento presentato al convegno 2011 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2011 tenutosi a Vancouver, BC, can nel 2011) [10.1109/IGARSS.2011.6049994].
Gaussian process regression within an active learning scheme
Pasolli E.;
2011
Abstract
In this work, we face the problem of training sample collection for the estimation of biophysical parameters by adopting the active learning approach. In particular, we propose two active learning strategies specifically developed for Gaussian Process (GP) regression. The first one is based on adding samples that are distant from the current training samples in the kernel space while the second one exploits an intrinsic GP regression outcome to pick up the most difficult samples. Experiments on simulated and real data sets show the effectiveness of active selection of training samples for regression problems. © 2011 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
