A Structural Approach to Coordinate-Free Statistics

Tom LaGatta, P. Richard Hahn

We consider the question of learning in general topological vector spaces. By exploiting known (or parametrized) covariance structures, our Main Theorem demonstrates that any conti…