. 2022 Jul 31;12(8):1262. doi: 10.3390/jpm12081262

Table 1.

CGM-derived metrics used for the engineering of ML models.

Parameter	Formula
CV	$C V = \frac{S D}{\bar{G}} \times 100 %$ $, where \bar{G} = \frac{\sum_{i = 1}^{n} G_{i}}{n}$ $and S D = \sqrt{\frac{\sum_{i = 1}^{n} (G_{i} - \bar{G})}{n - 1}}$
LI	$L I = \sum_{i = 1}^{n - 1} {(G_{i} - G_{i + 1})}^{2} / 5$
LBGI	$L B G I = \frac{1}{n} \sum_{i = 1}^{n} r l (G_{i})$ , where $r l (G_{i}) = r (G_{i}) i f f (G_{i}) < 0 a n d 0 o t h e r w i s e,$ $r h (G_{i}) = r (G_{i}) i f f (G_{i}) > 0 a n d 0 o t h e r w i s e,$ $r (G_{i}) = 10 \times f^{2} (G_{i})$ , $f (G_{i}) = 1.509 \times [{(\log (18 \times G_{i}))}^{1.084} - 5.381]$
CONGA-1	$C O N G A (1) = \sqrt{\frac{\sum_{i = 2}^{n} (D_{i} - \bar{D})}{n - 1}}$ $where \bar{D} = \frac{\sum_{i = 2}^{n} D_{i}}{n - 1}$ $, D_{i} = G_{i} - G_{i - 1}$
Minimum value	$G_{\min} = \min (G_{1}, \dots, G_{n})$
DLV	$G_{n - 1} - G_{n}$
ALV	$(G_{n} - G_{n - 1}) - (G_{n - 1} - G_{n - 2})$
LC	$coefficient b_{1}$ $in a linear trend model G_{i} = b_{0} + b_{1} t_{i} + ε_{i}$

Abbreviations: ALV, acceleration over the last values; CONGA-1, 1-h continuous overlapping net glycemic action; CV, coefficient of variation; DLV, difference between the last two values; LBGI, low blood glucose index; LC, linear trend coefficient; LI, lability index.