Skip to main content
. 2022 Sep 5;25(1):188–202. doi: 10.1093/biostatistics/kxac035

Fig. 3.

Fig. 3

Computation times (seconds) for exact and approximate Inline graphic calculations as a function of increasing number of observations Inline graphic. Computational time (Inline graphic-axis) as a function of observations (Inline graphic-axis) to calculate the individual components of (a) exact Inline graphic including (i) the dissimilarity matrix Inline graphic (purple) scaling Inline graphic, (ii) the adjacency matrix (orange), and (iii) the most expensive operation Inline graphic (pink) scaling Inline graphic. Note, Inline graphic is only shown for Inline graphic = 100 and 500 observations, but the trend is shaded in for the other observations; (b,c) have different y-axes than (a). The diagonal line between (a) and (b) connects the 20-s ticks of these two axes. (b) Approximate Inline graphic estimation (HPE) using the grid search procedure including (i) the dissimilarity matrix Inline graphic (blue) scaling Inline graphic and (ii) the HPE algorithm to estimate Inline graphic (orange) scaling Inline graphic; (c) approximate Inline graphic estimation using the bootstrap procedure (HPB) (purple), which scales similarly to HPE without the computational expense required for calculating Inline graphic. Note (b) and (c) have a different Inline graphic-axis scale than (a) for an zoomed in visualization of time.