, does not require the definition of a metric in the predictor space, and lends itself to graphi-cal interpretation.

to the period 1961–90, describes an extended suite of surface climate variables, explicitly incorporates elevation as a predictor variable, and contains an evaluation of regional errors associated with this and other commonly used climatologies. The climatology is already being used by researchers in the areas

We present a framework for discriminative sequence classification where linear classifiers work directly in the explicit high-dimensional predictor space of all subsequences in the training set (as opposed to kernel-induced spaces). This is made feasible by employing a gradient

We consider the least-square regression problem with regularization by a block 1-norm, i.e., a sum of Euclidean norms over spaces of dimensions larger than one. This problem, referred to as the group Lasso, extends the usual regularization by the 1-norm where all spaces have dimension one, where

Online social networks such as Friendster, MySpace, or the Facebook have experienced exponential growth in membership in recent years. These networks offer attractive means for interaction and communication, but also raise privacy and security concerns. In this study we survey a representative

Soils harbor enormously diverse bacterial populations, and soil bacterial communities can vary greatly in composition across space. However, our understanding of the specific changes in soil bacterial community structure that occur across larger spatial scales is limited because most previous work

, we propose the following modifications: an observation model arising from an adaptive appearance model, an adaptive velocity motion model with adaptive noise variance, and an adaptive number of particles. The adaptivevelocity model is derived using a first-order linear predictor based

For supervised and unsupervised learning, positive definite kernels allow to use large and potentially infinite dimensional feature spaces with a computational cost that only depends on the number of observations. This is usually done through the penalization of predictor functions by Euclidean

We present a novel and flexible approach to the problem of feature selection, called grafting.Rather than considering feature selection as separate from learning, grafting treats the selection of suitable features as an integral part of learning a predictor in a regularized learning framework

Fisher's linear discriminant analysis is a valuable tool for multigroup classification. With a large number of predictors, one can nd a reduced number of discriminant coordinate functions that are "optimal" for separating the groups. With two such functions one can produce a