Algorithm 1 K-margin(${\mathit{x}}^{(i)},K,\varphi$)

Input: A T-segments array ${\mathit{x}}^{(i)}=\{{\mathit{x}}_1^{(i)},{\mathit{x}}_2^{(i)},\dots ,{\mathit{x}}_T^{(i)}$} of an input health-condition monitoring record ${\mathit{x}}^{(i)}$, trained multi-view RCR-Net model $\varphi$ Parameter: An integer K Output: Top-K most confident segments ${\mathit{X}}_{1:K}^{(i)}$ and their label predictions ${\widehat{Y}}_{1:K}^{(i)}$ 1: $x^{{}^{\prime }(i)}\leftarrow {\mathit{x}}^{(i)}$ 2: ${\mathit{X}}_{1:K}^{(i)}\leftarrow \varnothing$, ${\widehat{Y}}_{1:K}^{(i)}\leftarrow \varnothing$, $k\leftarrow K$ 3: while$k>0$do 4:  ${\mathit{x}}_{\ast }^{(i)}\leftarrow \mathrm{Most}\mathrm{confident}\mathrm{segment}\mathrm{predicting}\mathrm{by}\varphi (x^{{}^{\prime }(i)})$ using Equation (3) 5:  ${\widehat{y}}_{\ast }^{(i)}\leftarrow \mathrm{Most}\mathrm{confident}\mathrm{label}\mathrm{predicting}\mathrm{by}\varphi (x^{{}^{\prime }(i)})$ using Equation (4) 6:  $x^{{}^{\prime }(i)}\leftarrow x^{{}^{\prime }(i)})\backslash {\mathit{x}}_{\ast }^{(i)}$ # Implementation: remove segment index of ${\mathit{x}}_{\ast }^{(i)}$ from segment indexes list of $x^{{}^{\prime }(i)}$ 7:  ${\mathit{X}}_{1:K}^{(i)}\leftarrow {\mathit{X}}_{1:K}^{(i)}\cup \{{\mathit{x}}_{\ast }^{(i)}\}$, ${\widehat{Y}}_{1:K}^{(i)}\leftarrow {\widehat{Y}}_{1:K}^{(i)}\cup \{{\widehat{y}}_{\ast }^{(i)}\}$ 8:  $k\leftarrow k-1$ 9: end while 10: return${\mathit{X}}_{1:K}^{(i)},{\widehat{Y}}_{1:K}^{(i)}$