Ensemble learning algorithms such as boosting can achieve better performance by averaging over the predictions of some base hypotheses. Nevertheless, most existing algorithms are ...
In this paper, we consider the topic of iterative, one dimensional, signal reconstruction (specifically speech signals) from the magnitude spectrum and the phase spectrum. While ...
: Feature selection methods are often applied in the context of document classification. They are particularly important for processing large data sets that may contain millions of...
Janez Brank, Dunja Mladenic, Marko Grobelnik, Nata...
Vector field segmentation methods usually belong to either of three classes: methods which segment regions homogeneous in direction and/or norm, methods which detect discontinuiti...
Tristan Roy, Eric Debreuve, Michel Barlaud, Gilles...
Most game programs have a large number of parameters that are crucial for their performance. While tuning these parameters by hand is rather difficult, efficient and easy to use ge...