Table 1.

Algorithm-specific Proportion of Explained Variation R²

Algorithm	R2
Gaussian	$1-\left\{\frac{{\sum }_{i,j}{(V_{\mathit{ij}}-{(WH)}_{\mathit{ij}})}^2}{{\sum }_{i,j}{(V_{\mathit{ij}}-\overline{V})}^2}\right\}$
$K{L}_G^H$, $K{L}_G^{\mathit{MM}}$	$1-\left\{\frac{\widehat{{\mathit{KL}}_G}(V,WH)}{{\sum }_{i,j}\{-\log (\frac{V_{\mathit{ij}}}{\overline{V}})+\frac{V_{\mathit{ij}}}{\overline{V}}-1\}}\right\}$
$K{L}_G^d$	$1-\left\{\frac{\widehat{{\mathit{KL}}_G^d}(V,WH)}{{\sum }_{i,j}\{\log (\frac{V_{\mathit{ij}}}{\overline{V}})+\frac{\overline{V}}{V_{\mathit{ij}}}-1\}}\right\}$
JG	$1-\left\{\frac{\widehat{J_G}(V,WH)}{{\sum }_{i,j}\left\{\frac{{(V_{\mathit{ij}}-\overline{V})}^2}{V_{\mathit{ij}}\overline{V}}\right\}}\right\}$
$K{L}_{\mathit{IG}}^H$, $K{L}_{\mathit{IG}}^{\mathit{MM}}$	$1-\left\{\frac{\widehat{K{L}_{\mathit{IG}}}(V,WH)}{{\sum }_{i,j}\frac{{\left\{V_{\mathit{ij}}-\overline{V}\right\}}^2}{V_{\mathit{ij}}{\overline{V}}^2}}\right\}$
$K{L}_{\mathit{IG}}^d$	$1-\left\{\frac{\widehat{K{L}_{\mathit{IG}}^d}(V,WH)}{{\sum }_{i,j}\frac{{\left\{V_{\mathit{ij}}-\overline{V}\right\}}^2}{{V_{\mathit{ij}}}^2\overline{V}}}\right\}$