457
v1v2v3 (latest)

Performance of empirical risk minimization in linear aggregation

Abstract

We study conditions under which, given a dictionary F={f1,,fM}F=\{f_1,\ldots ,f_M\} and an i.i.d. sample (Xi,Yi)i=1N(X_i,Y_i)_{i=1}^N, the empirical minimizer in span(F)\operatorname {span}(F) relative to the squared loss, satisfies that with high probability \[R\bigl(\tilde{f}^{\mathrm{ERM}}\bigr)\leq\inf_{f\in\operatorname {span}(F)}R(f)+r_N(M),\] where R()R(\cdot) is the squared risk and rN(M)r_N(M) is of the order of M/NM/N. Among other results, we prove that a uniform small-ball estimate for functions in span(F)\operatorname {span}(F) is enough to achieve that goal when the noise is independent of the design.

View on arXiv
Comments on this paper