We consider the problem of reconstructing a function from a finite set of\nnoise-corrupted samples. Two kernel algorithms are analyzed, namely kernel\nridge regression and $\\varepsilon$-support vector regression. By assuming the\nground-truth function belongs to the reproducing kernel Hilbert space of the\nchosen kernel, and the measurement noise affecting the dataset is bounded, we\nadopt an approximation theory viewpoint to establish \\textit{deterministic},\nfinite-sample error bounds for the two models. Finally, we discuss their\nconnection with Gaussian processes and two numerical examples are provided. In\nestablishing our inequalities, we hope to help bring the fields of\nnon-parametric kernel learning and system identification for robust control\ncloser to each other.\n