In this paper we propose a set of algorithms that combine the anisotropic smoothing using the heat kernel with the outlier rejection capability of robust statistics. The proposed ...
In this paper we present a simple hierarchical Bayesian treatment of the sparse kernel logistic regression (KLR) model based MacKay's evidence approximation. The model is re-p...
A common way to represent a time series is to divide it into shortduration blocks, each of which is then represented by a set of basis functions. A limitation of this approach, ho...
Abstract. Using the classical Parzen window (PW) estimate as the target function, the sparse kernel density estimator is constructed in a forward constrained regression manner. The...
Kernelized LASSO (Least Absolute Selection and Shrinkage Operator) has been investigated in two separate recent papers (Gao et al., 2008) and (Wang et al., 2007). This paper is co...