PRESENTATION ABSTRACT  Many regression and classification procedures fit a function f(x;w) of predictor variables x to data 〖{x_i,y_i}〗_1^N based on some loss criterion L(y,f(x;w)). Often, regularization is applied to improve […]