Fig. 1
Coefficient error \(E_2= \Vert \textbf{w}^\star -{{\widehat{\textbf{w}}}}\Vert _2/\Vert \textbf{w}^\star \Vert _2\) of WENDy applied to the Logistic Growth model vs test function radius \(m_t\) for noise levels \(\sigma _{NR}\in \{10^{-6},\dots ,10^{-1}\}\). For large enough radius, errors are dominated by noise and integration error is negligible. The minimum radius \({\underline{m}}_t\) computed as in Sect. 2.3.1 finds this noise-dominated region, which varies depending on \(\sigma _{NR}\)